US 8,554,719 B2Grant
Resolving database entity information
Issue Date:2013-10-08
•38 Claims
•14 Drawing Sheets
Abstract
Entity resolution in a database comprises receiving imported data comprising imported data entities each having properties each having values; receiving first user input that selects the imported data entities for resolution to existing data entities in a database; receiving second user input that specifies matching criteria for matching the imported data entities to the existing data entities, wherein each of the matching criteria comprises a matching technique; matching the imported data entities to the existing data entities using the matching criteria, resulting in creating and storing matched entity information, wherein the matched entity information is organized in matched entity data sets associated with subsets of the matching criteria that were matched; consolidating the imported data entities into the existing data entities; storing the first user input and second user input as a named criteria set for use in subsequent entity resolution operations.
Metadata
Assignee
- Palantir Technologies, Inc.
Inventors
- Robert J. McGrew
- Stephen Cohen
Application Information
Application Number:US 11/874,644
Filing Date:2007-10-18
Priority Date:2007-10-18
Art Unit:2161
Classifications
IPC:
G06F7/00
Patent Drawings (14 sheets)
Description
Technical Field
[0001] The present disclosure generally relates to creating and storing information in databases.
Background
[0002] The approaches described in this section could be pursued, but are not necessarily approaches that have been previously conceived or pursued. Therefore, unless otherwise indicated herein, the approaches described in this section are not prior art to the claims in this application and are not admitted to be prior art by inclusion in this section.
[0003] Intelligence analysts and other computer users often interact with computer-based databases. When these users work with a large database containing information on millions of individuals, many individuals represented in the database may possess the same name, which can make it a challenge to determine what information refers to each unique real-world individual. Moreover, factors such as misspellings, alternate spellings, and missing information can cause further complications. The same issues are encountered when new data is imported into the database from external data sources. For databases that operate with terabytes of data, these problems are acute. Users and researchers in the database field desire to have systems that can ensure that there is only one database entry for each real person, such that one record relates to one unique entity in the database.
Brief Description of the Drawings
[0004] In the drawings:
[0005] FIG. 1 illustrates a revisioning database system.
[0006] FIG. 2 illustrates a method for entity resolution in a database system.
[0007] FIG. 3 illustrates an example column identification screen.
[0008] FIG. 4 illustrates setting the Role, Link Type, and Reference for a payments table.
[0009] FIG. 5 illustrates an example graphical user interface when a user is verifying table type using the data importer.
[0010] FIG. 6 illustrates a display of statistical information and errors for imported data.
[0011] FIG. 7 illustrates an example error display.
[0012] FIG. 8 illustrates an example Entity Resolver display.
[0013] FIG. 9 illustrates an example Entity Resolver display associated with choosing matching criteria.
[0014] FIG. 10 illustrates an example Entity Resolver screen showing results of matching.
[0015] FIG. 11 illustrates an example screen of the Entity Resolver relating to checking and finalizing.
[0016] FIG. 12 illustrates an example Histogram window.
[0017] FIG. 13 illustrates a computer system upon which an embodiment may be implemented.
Detailed Description
[0018] In an embodiment, entity resolution in a database comprises receiving imported data comprising imported data entities each having properties each having values; receiving first user input that selects the imported data entities for resolution to existing data entities in a database; receiving second user input that specifies matching criteria for matching the imported data entities to the existing data entities, wherein each of the matching criteria comprises a matching technique; matching the imported data entities to the existing data entities using the matching criteria, resulting in creating and storing matched entity information, wherein the matched entity information is organized in matched entity data sets associated with subsets of the matching criteria that were matched; consolidating the imported data entities into the existing data entities; storing the first user input and second user input as a named criteria set for use in subsequent entity resolution operations.
[0019] In the following description, for the purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced without these specific details. In other instances, well-known structures and devices are shown in block diagram form in order to avoid unnecessarily obscuring the present invention.
[0020] Embodiments are described herein according to the following outline:
-
- [0021] 1.0 General Overview
- [0022] 2.0 Structural and Functional Overview
- [0023] 2.1 Example Operational Context—Database System
- [0024] 2.2 Importing Data
- [0025] 2.3 Entity Resolution
- [0026] 2.3.1 Overview of Example Implementation
- [0027] 2.3.2 Saved Criteria Sets
- [0028] 2.3.3 Alternate Matching Techniques
- [0029] 2.3.4 “One Click” Approach
- [0030] 3.0 Implementation Mechanisms—Hardware Overview
- [0031] 4.0 Extensions and Alternatives
[0032] 1.0 General Overview
[0033] In an embodiment, a computer-readable data storage medium, apparatus or machine-implemented method are operable for receiving imported data comprising one or more imported data entities each having one or more properties each having one or more values; receiving first user input that selects one or more of the imported data entities for resolution to one or more existing data entities in a database; receiving second user input that specifies one or more matching criteria for matching the imported data entities to the existing data entities, wherein each of the matching criteria comprises a matching technique; matching the imported data entities to the existing data entities using the matching criteria, resulting in creating and storing matched entity information, wherein the matched entity information is organized in one or more matched entity data sets associated with one or more subsets of the matching criteria that were matched; consolidating the imported data entities into the existing data entities; storing the first user input and second user input as a named criteria set for use in subsequent entity resolution operations.
[0034] In one feature, the matching technique specifies no conflicts between values of properties of the imported data entities and the existing data entities.
[0035] Another feature further comprises receiving additional imported data for resolution to the existing data entities; receiving third user input that specifies a name of a stored criteria set; loading the stored criteria set; automatically performing the matching based on the stored criteria set and without receiving the second user input
[0036] Yet another feature further comprises receiving the imported data; receiving the first user input as a selection of one of the imported data entities and one of the existing data entities in a database; receiving third user input requesting an entity resolution of the selected imported data entity and the selected existing data entity; loading the stored criteria set; automatically performing the matching based on the stored criteria set and performing the consolidating without receiving the second user input.
[0037] A further feature further comprises receiving third user input that specifies one or more target filters for types of the existing data entities, and performing the matching only for corresponding ones of the existing data entities in the database. Yet another feature further comprises displaying the imported data entities in a histogram display that identifies the imported data entities and corresponding counts of the values of the properties of the imported data entities; receiving the first user input as selections of one or more of the imported data entities in the histogram display.
[0038] Another feature comprises displaying, in a first panel of a graphical user interface, the subsets of the matching criteria that were matched, and counts of the one or more matched entity data sets associated with the one or more subsets of the matching criteria that were matched; receiving third user input that selects one of the subsets; displaying, in response to the third user input, in a second panel of the graphical user interface, each matching entity represented in the count of the selected subset.
[0039] In one feature, the consolidating comprises merging the imported data entities into the existing data entities.
[0040] Yet another feature further comprises displaying, in a graphical user interface, results of consolidating the imported data entities into the existing data entities, the results comprising a first list of one or more associations of resolved entities each comprising one of the imported data entities, one of the existing data entities, and an associated subset of the matching criteria that was matched for that association; a second list of one or more new entities to create in the database based on one or more of the imported data entities that did not match any of the existing data entities.
[0041] Still another feature comprises displaying, in a graphical user interface, four visually expandable and collapsible graphical regions associated with steps in an entity resolution process; [FIG. 8 ] wherein a first graphical region when expanded comprises a first list of the one or more imported data entities and one or more of the properties and one or more of the values; [FIG. 9 ] wherein a second graphical region when expanded comprises a first panel comprising the one or more matching criteria for matching the imported data entities to the existing data entities and the matching technique for each of the matching criteria, and a second panel comprising one or more target filters for types of the existing data entities; [FIG. 10 ] wherein a third graphical region when expanded comprises a third panel comprising the subsets of the matching criteria that were matched, and counts of the one or more matched entity data sets associated with the one or more subsets of the matching criteria that were matched, and a fourth panel comprising each matching entity represented in the count of a selected subset; [FIG. 11 ] wherein a fourth graphical region when expanded comprises a fifth panel comprising a second list of one or more associations of resolved entities each comprising one of the imported data entities, one of the existing data entities, and an associated subset of the matching criteria that was matched for that association; and a sixth panel comprising a third list of one or more new entities to create in the database based on one or more of the imported data entities that did not match any of the existing data entities.
[0042] 2.0 Structural and Functional Overview
[0043] In one embodiment, a process for performing entity resolution for a database comprises:
[0044] 1. Receiving a selection, by a user, of a subset of imported data records for resolution.
[0045] 2. Receiving a user selection of a desired resolution approach including matching criteria.
[0046] 3. Finding matches of the subset of records in a database based on selection properties that are useful for resolution, and storing the records in further subsets based on the number of matches, receiving selections of further subsets and consolidating the subsets, and displaying a view of a list of matches, based on sets of matched properties.
[0047] 4. Receiving confirmation of the matches and finalizing the resolution of entities.
[0048] In one embodiment, the process further provides for storing a recording of user interaction with a graphical user interface that supports the preceding steps as a set of rules that can be executed automatically on other input data having the same format. Example user interaction steps that can be saved include selection of persons in the database, selection of matching types, etc. The rules may be saved as a criteria set, and a user may later load a criteria set and apply the criteria set to newly received data, rapidly importing the data into the database in a consistent and structured manner.
[0049] In an embodiment, matching techniques may include exact matches, fuzzy matches, and matches in which imported data has no conflicts with data of existing entities.
[0050] In an embodiment, a “one-click” process is provided for performing multiple entity resolution on data items represented in a graphical display, by selecting nodes of the graphical display and selecting a Consolidate function or the equivalent; in response, a data processing system automatically performs entity resolution.
[0051] Embodiments may operate with data from databases, spreadsheets, documents, XML files, and other data formats or file formats. In one embodiment, a visual data import process allows data analysts to connect with or import data from SQL, XML, email formats, semi-structured and unstructured text, and spreadsheet formats.
[0052] 2.1 Example Operational Context—Database System
[0053] The techniques described herein may be used in many different database applications. In an embodiment, a revisioning database system stores data collected in intelligence operations by intelligence analysts or other members of an intelligence community. “Intelligence,” in this context, refers to information about people, places, and events that is useful to a government or military organization in relation to issues of security or defense, and may include information obtained in covert operations or open operations. Other embodiments may be used in fields other than intelligence analysis. In an embodiment, the revisioning database system records metadata about changes to stored data, facilitates UNDO and REDO operations, can receive requests to subscribe to particular data and publish updates to such data for delivery to subscribers, and performs other functions.
[0054] FIG. 1 illustrates a revisioning database system. A revisioning database 108 is coupled to application logic 112, which implements the functions described herein for interacting with and using the revisioning database 108 in association with data such as intelligence community data. A user terminal 106 is coupled to front-end logic 130, which is coupled to application logic 112. The user terminal 106 of FIG. 1 broadly represents any host computer of an intelligence analyst, database administrator, or other user who interacts with revisioning database 108 through application logic 112. Database 108 may use a dynamic ontology of the type described in co-pending U.S. patent application Ser. No. 11/602,626, filed Nov. 20, 2006.
[0055] One or more other intelligence analysts 102, 104 can access application logic 112 through front-end logic 130 to obtain information from revisioning database 108 and to obtain services of the application logic. The intelligence analysts 102, 104 may be located in discrete security domains and different geographical locations, or may be co-located and/or within the same security domain. Each of the intelligence analysts 102, 104 may access application logic 112 through a host computer that is coupled directly or indirectly through one or more networks, network links, or internetworks to the application logic. User terminal 106 and the host computers of the intelligence analysts 102, 104 may comprise a personal computer, workstation, or other data processing system.
[0056] Front-end logic 130 comprises a memory 132 that stores copies of objects from revisioning database 108 for use in front-end operations, and a GUI engine 134 that generates graphical user interface displays for presenting data objects on user terminal 106 or displays of the intelligence analysts 102, 104.
[0057] In an embodiment, application logic 112 comprises history window logic 110, publish-subscribe module 118, rules wizard 120, rules engine 122, Data Importer 150 and Entity Resolver 152. History window logic 110 is configured to generate views and displays of historical states of objects in revisioning database 108 and historical states of graphs, comprising nodes representing the objects, which have been displayed by GUI engine 134. Publish-subscribe module 118 comprises logic configured to implement receiving and storing requests of analysts to subscribe to particular data views, detecting updates in data, and generating and sending publications of updated data to subscribers. Rules wizard 120 is configured to permit users to define rules that specify how objects in revisioning database 108 can be accessed and exported, as further described below. Rules engine 122 facilitates controlled publication and viewing of objects from revisioning database 108 based on realms and rules defined using the rules wizard 120.
[0058] Data importer 150 provides processes and graphical user interface logic for assisting a user in importing data into the revisioning database 108. Source data for importing may come from spreadsheets or other data sources at user terminal 106 or associated with the intelligence analysts. Entity resolution logic 152 provides processes and graphical user interface logic for resolving associations of entities represented in imported data to entities that are represented in revisioning database 108. Entity resolution logic 152 provides automatic processes for resolving which entities in imported data match existing entities in the database, and adding the imported data to the existing entities.
[0059] Data Importer 150 and Entity Resolver 152 are described in further detail in separate sections herein.
[0060] In an embodiment, export module 114, publish-subscribe module 118, rules wizard 120, rules engine 122, Data Importer 150 and Entity Resolver 152 may be implemented in a single computer program or module or may be integrated into revisioning database 108. Functions of rules wizard 120 may be implemented in whole or in part in front-end logic 130. Thus, the structural separation of functional modules as shown in FIG. 1 is not required and the same functions described herein for such functional modules may be implemented structurally in any other convenient manner.
[0061] FIG. 2 illustrates a method for entity resolution in a database system. In step 202, one or more unresolved entities are selected from among imported data. For example, a set of data is gathered and rows of the data are displayed. One or more rows are selected for resolution with an existing database, such as database 1080. In one embodiment, the selection of entities for resolution is performed by receiving user input in a user interface that selects particular entities, as indicated by step 202A. For example, a user might operate a pointing device to select one or more rows representing entities in a graphical user interface. Alternatively, as shown in step 202B, the selection of entities may comprise receiving input of a semantic criterion referencing properties of the entities. For example, a user might provide input such as “Select all entities whose citizenship property has the value ‘US’.” Natural language input may be used in conjunction with a parser, or a structured input mechanism such as Structured Query Language (SQL) may be used. Using input that references semantic criteria, a user can specify different matching criteria for different entities.
[0062] In step 204, one or more criteria are chosen for matching the unresolved entities to existing entities in the database. For example, matching criteria specify which properties or data fields of a row must match an entity in the database for the row to be considered the same entity.
[0063] In step 206, one or more successful matches in the database are found based on the matching criteria. For example, properties of the data rows that were selected in step 202, corresponding to the matching criteria selected in step 204, are matched against properties of entities that are already represented in the database.
[0064] In step 208, the matches are checked and finalized. In one embodiment, matches found in step 206 are displayed and a user confirms the matches for resolution. In response, the imported data for a row is associated with a matching entity in the database.
[0065] In an embodiment, an entity resolution process may be implemented as a multi-pass process in which different matching criteria are applied to entities in different passes. In such an embodiment, in step 210 a test is performed to determine if all entities were resolved. Incomplete resolution of entities may occur if entities were not selected in step 202 if no successful matches were found in the database in step 206 for the entities. If so, then control passes to step 212 in which the user may optionally save the set of criteria used in the process for future use on other data sets. If not all entities were resolved, then control passes to step 214 in which the matching criteria optionally may be modified for use in second or subsequent passes of the process; modifying the matching criteria may comprise receiving input in any of the ways described above for steps 202, 202A, 202B. Modified matching criteria may be provided in many different forms. For example, one matching approach may be used for entities representing citizens of a first country and another matching approach may be used for entities representing citizens of a second country. Control then transfers to step 202 and another pass of the process is applied to entities that were not resolved in the first pass or other prior passes.
[0066] As a result, imported data becomes associated with existing entities in the database, ensuring consistency and preventing duplication of records or entities in the database. For example, imported data for a particular individual can be associated with an existing entity in the database even when the imported data uses a different spelling for a name of the individual, provided that other defined criteria match between the imported data and the existing entity. Thus, imported data that semantically matches data in an existing entity of the database, even if the data does not syntactically or lexically match the data in the existing entity, becomes associated with the existing entity.
[0067] As another example, overlapping or duplicate data can be imported and joined to existing entities in the database based on specified criteria, such as matches in address, birth date, etc. Thus, if a user imports a data set and a particular imported row matches an existing entity on seven specified criteria, there can be strong confidence that the imported row is referring to the same individual as the existing entity in the database, and therefore all other fields of the row can be added to properties of the existing entity in the database; the imported row does not represent an entirely new data entity.
[0068] Further, if a different imported row matches an existing entity based on only two of seven properties, then an analyst might choose not to consolidate the imported row into an existing entity of the database, but create a new entity instead. As an example, if an imported row of information about an individual named “Jane Smith” matches an existing entity in the database only based on First Name and Last Name, but Address, Phone, and Birth Date are different, then consolidation is probably incorrect because different individuals are involved.
[0069] 2.2 Importing Data
[0070] In various embodiments, processes and graphical user interfaces are provided to facilitate importing data from spreadsheets or other sources into the database. Embodiments provide an adaptable data importer that can greatly reduce the amount of effort expended organizing data from data sources. In one embodiment, a spreadsheet importation feature provides a comprehensive data import mechanism that can work with any delimited data file, or spreadsheet. Thus, whenever a user needs to bring information concerning an entity, event, or relationship that is not available in existing databases into an investigation, the user can quickly create a spreadsheet containing that data for importation.
[0071] In an embodiment, any structured, delimited data file representing entities, relationships, or events can be imported and each entity, event, or relationship is treated as a composite of its different properties and sub-properties. Therefore, embodiments can handle various organizations of information successfully and efficiently, allowing users to spend less time conforming to strict data input methodologies and more time focusing on an investigation or the use of the data. A user confirms that the system has deduced column and table types correctly, and then proceeds to entity resolution, described in the next section. If the system has deduced column or table types incorrectly, then the system adapts its ontology to the new format of column naming or spreadsheet typing, thus improving itself with every use.
[0072] The description herein assumes that a user or another system has created a data set in an electronic spreadsheet file. An example is a Microsoft Excel file. In general, a user activates the Data Importer 150 and selects a file for importation. The Data Importer 150 verifies the structure, character, and content of the information. The data is entered in one or more tables in revisioning database 108. The data then can pass through an entity resolution process for integration with other entities that are already represented in the database 108.
[0073] In one embodiment, a user activates Data Importer 150 by selecting a Data Import function from a graphical user interface that the application logic 112 generates in cooperation with GUI engine 134. In response, the Data Importer 150 generates data for a Data Importer screen, which the GUI engine 134 renders and provides to user terminal 106. The Data Importer screen provides options to import data from files and to import data from a database. As an example, importing data from a file is described herein.
[0074] A section of an “Add file” button is received. In response, Data Importer 150 generates and displays a “File Open” dialog box on user terminal 106. The user navigates to the file containing the information that the user wants to import and selects “Open”. The user can import more than one file at a time, and alternatively can drag and drop the desired files onto the Data Importer screen.
[0075] In an embodiment, a database may be imported by selecting an “add database button.” In an embodiment, a large or sensitive database may be imported using a “Link as Federated” selection, which causes the Data Importer 150 to import only the structure of the selected database so that information in the database is not available for query.
[0076] In response to receiving a selection of a file to import, the Data Importer 150 generates and displays a column identification screen on the user terminal. FIG. 3 illustrates an example column identification screen of the Data Importer 150. User interaction with the column identification screen 300 ensures that the system understands the character of the information that the user is importing. In the embodiment of FIG. 3 , column identification screen 300 comprises a help panel 302, data sources panel 304, source file panel 306, column identification panel 310, details window 318, and navigation buttons 320.
[0077] The help panel 302 displays prompts and other help information to assist a user. The data sources panel 304 identifies sources of data that are undergoing importation; the currently imported source is highlighted. The source file panel 306 identifies the name and database table type 308 for the current file that is undergoing importation. The details window 318 provides a tree display of details about data that is being imported and allows the user to preview information from the contents of the data sources that the user is importing. The navigation buttons 320 enable a user to cancel importation and proceed forward or backward through functional steps in importation.
[0078] The column identification panel 310 displays the names of each of the columns in the imported file listed in a “Column Name” area 312. A “Column Type” area 314, “Role” area 316, “Link Type” area, and “Reference” area display information that the Data Importer 150 generated based on the imported file and the dynamic ontology of the system to guess what each of the columns in the imported file represents based on names in the imported file.
[0079] Key symbols 322 in FIG. 3 indicate primary keys. Primary keys comprise information that will uniquely identify each individual suspect, flight, phone call, etc., in database 108. In the example of FIG. 3 , a list of suspects is being imported, and therefore first and last names are used to identify unique individuals. However, if there are two suspects with the same name in the list to be imported, then another primary key is used, such as a social security number or address. To add or remove an identifying attribute, the user selects a key symbol or selects a location at which a key should be placed. In response, the Data Importer 150 internally stores information designating a primary key column.
[0080] FIG. 4 illustrates setting the Role, Link Type, and Reference for a payments table. In an embodiment, when a user is importing data about events such as phone calls, emails, or payments, a Role chooser is used to define which column represents the recipient(s) and which the sender(s). For example, in FIG. 4 a user selects a Role combo box in Role area 316 and selects an available role for a payment such as “From” one entity to another.
[0081] In an embodiment, a Link Type menu 406 in Link Type area 402 allows a user to define a relationship between what the column represents and the primary event or entity of that row. For instance, if in a table of suspects there is a column that lists a coworker for each of them, the coworker link type could be selected. In the example of FIG. 4 , a link type of “Participates In” has been selected for the payments represented in the imported data.
[0082] In an embodiment, a Ref. menu in Reference area 404 allows a user to set the type of entity referenced by that column, such as a person, business, or group. In the example of FIG. 4 , the payor is a person and the payee is a business, as indicated by icons in Reference area 404.
[0083] If any data is incorrect in column type area 314, role area 316, or link type 402, a user can select a text entry box that contains erroneous data and navigate menus to find an appropriate label. In response, Data Importer 150 adds the new column type format to the database so that future import steps will take the new column type format into account. In the example of FIG. 4 no changes need to be made.
[0084] In an embodiment, if a column name in the imported file does not match the data in that column based on the expected ontology, Data Importer 150 highlights the column in an alternate color. In response, to fix the problem, a user can navigate the column's pull-down menu and select the appropriate column type.
[0085] FIG. 5 illustrates an example graphical user interface when a user is verifying table type using the data importer. In an embodiment, a user can check whether the system correctly determined what general category the information in the table falls into by reviewing values shown in Table Type area 306. If the system's initial guess is incorrect, then the user navigates a pull-down menu 308. The user can select a changed table type value from list 315. In response, data importer stores the changed table type value and adapts to use the changed table type value in future importation operations.
[0086] Selecting the NEXT button causes Data Importer 150 to import the data and place the data in a holding area of database 108 for evaluation, correction or entity resolution. After importation, statistics and errors are displayed.
[0087] FIG. 6 illustrates a display of statistical information and errors for imported data. In an embodiment, Data Importer 150 generates a selection of statistics on the data that is being imported and provides a user with the opportunity to fix any errors that the system detected. In the example of FIG. 6 , a statistics window 602 displays statistics for structured data sources that include “Person” table types, as indicated by type label 608. Statistics include an object count 604 that identifies the total number of objects that were imported and a property count 606 that identifies the total number of object properties that were imported.
[0088] Error messages 610 specify errors that were found during data importing. If errors are present, the user can select a “fix errors” button 616 to activate error resolution logic. In response, Data Importer 150 generates an error display. FIG. 7 illustrates an example error display. An error display 702 may comprise an error listing 701 comprising one row for each error and comprising an Error Type column 706, Error Data column 708, Error Property column 710, and Error Data Source column 712. Thus, error listing 701 summarizes the errors and their nature.
[0089] In an embodiment, a repair panel 720 displays options for fixing the errors. For example, a user can specify replacing error data, removing a property, removing a row, or fixing the error later. A plurality of buttons 722 enable a user to apply the selected option to all errors or only a selected error.
[0090] In an embodiment, selecting an error from list 701 causes the Data Importer 150 to display the data row in which the error occurred at the bottom of the screen, with the erroneous data highlighted. In an embodiment, to fix erroneous data by replacing it with something else, a user can select the “replace with” option and type in the desired entry into the white box. Selecting “remove property” prevents the contents of the erroneous cell from being imported. Selecting “remove row” remove the entire row of data from the importation process.
[0091] Each time the user fixes data, the user has three options for applying it. To only apply the fix to the selected error, the user selects the “apply” button. If the error is a recurring error and the user wants to apply the same fix to each occurrence of the error, “apply same” is selected and applies the fix to any other errors of the same type, in the same column, and with the same field value. For instance, to remove a property on all entries of “xxxx” for a phone number, the user could select “apply same”. The “apply all” option allows the user to apply a fix to all other errors of the same type and in the same column, for example, if various terms such as “n/a”, “none”, and “xxxx” were entered for lacking phone numbers.
[0092] Selecting the “ok” button 724 causes the data importer to re-display the previous screen of FIG. 6 .
[0093] 2.3 Entity Resolution
[0094] 2.3.1 Overview of Example Implementation
[0095] FIG. 6 also illustrates information for choosing criteria for entity resolution at the end of data importation as indicated in entity resolution panel 612. In an embodiment, a user can choose whether to use new or established criteria for entity resolution.
[0096] Generally, entity resolution is a process of searching existing data stores for any matches to the information that the user is currently importing and then resolving matches into one object. When dealing with a large database containing information on millions of individuals, it is a common occurrence for many of them to posses the same name, which can make it a challenge to determine what information refers to each unique real-world individual. Moreover, factors such as misspellings, alternate spellings, and missing information can further complicate the issue. In a database that operates with tremendous amounts of data, the problem becomes even more difficult to deal with. To effectively deal with this dilemma and have only one database entry for each real person, embodiments provide a process of entity resolution. By performing algorithmic searches through all available information, entity resolution logic can identify if there are one or more records pertaining to a specific real-world entity and then merge those records together, thus “resolving” the entity. The end result is a single set of information, with each record relating to one unique entity.
[0097] Criteria refer to the search parameters used to look for matches to existing entities, which types of matches are consolidated, and the order in which different types of matches are consolidated. The criteria used for entity resolution include details such as which parameters are used to look for matches among existing entities, which types of matches are consolidated, and the order in which different types of matches are consolidated. Establishing new criteria allows a user to define a new set of these parameters and control the specific details of the entity resolution process. Selecting “Use saved criteria” enables a user to use a previously-existing set of criteria, or criteria. Using saved criteria allows the user to bypass the specifics of the entity resolution process, which is appropriate, for example, if the contents of the files to be imported will resolve easily.
[0098] With Entity Resolver 152, a user can effectively resolve newly imported information with existing records, allowing the user to immediately resume an investigation armed with the newly integrated information. Entity resolution utilizes a dynamic ontology of the database system and algorithms to find the best existing match for any new entity, event, or relationship that is imported. Every imported object is broken down into its constituent properties and sub-properties, analyzed, and compared to the existing data for matches. The properties used to define what makes an entity unique can be user-defined and adjusted to suit the situation. The Entity Resolver then seamlessly integrates the information from the two records.
[0099] For example, assume that analyst 102 (FIG. 1 ) has just received the names and some personal information of eight suspects, entered the information into a spreadsheet, and imported the data using Data Importer 150. Now the analyst 102 wants to see if any of the individuals are matches for entities already present in the database 108, which contains information on millions of individuals. However, within such a large group of people there are bound to be several potential matches, especially if any of the new suspects has a common name. With the Entity Resolver 152, the analyst 102 can scour the current information for a match using several of the entities' various characteristic properties. After finding the correct match, the Entity Resolver 152 resolves the two matching entities, combining their information into one unique individual. As a result, multiple disparate information sources can contribute to a shared data set.
[0100] Referring again to FIG. 6 , in an embodiment, an “Establish new criteria” radio button 620 is selected and then the “Import” button 614 is selected. In response, an Entity Resolver display opens.
[0101] FIG. 8 illustrates an example Entity Resolver display. In an embodiment, the Entity Resolver display 802 comprises a plurality of step selection regions 801, 816, 818, 820. Each of the regions is visually expandable or collapsible in response to user input or in response to user selection of a graphical NEXT button or a BACK button. Region 801 when expanded is associated with a first step in entity resolution comprising selecting which entities are to be resolved. Region 816 when expanded is associated with a second step in entity resolution comprising choosing matching criteria. Region 818 when expanded is associated with a third step in entity resolution comprising finding successful matches and performing resolution. Region 820 when expanded is associated with a fourth step in entity resolution comprising checking and finalizing resolution. A user selection of a region 801, 816, 818, 820 causes the Entity Resolver display to provide features and functions relating to the selected step.
[0102] FIG. 8 shows an example Entity Resolver display when initially launched or when region 801 has been selected for the purpose of selecting which entities are to be resolved. In an embodiment, Entity Resolver display 802 comprises a help panel 804 and entity listing 806. The entity listing 806 comprises one or more rows of data for entities. Thus, each row represents data relating to an entity that has been imported from a data source. Each row comprises a selection checkbox 808. By default, all entities are selected, as seen in FIG. 8 . To select only certain entities, the user de-selects or “un-checks” entities that should not be resolved, or highlights those that the user does wish to resolve selects the “select highlighted” button 810. The “select all” button 812 and “deselect all” buttons 814 enable a user to automatically cause selecting all entities or automatically cause un-selecting all entities.
[0103] In an embodiment, launching the Entity Resolver 152 also causes displaying a Histogram window. FIG. 12 illustrates an example Histogram window. In an embodiment, a Histogram window 1200 comprises an entity properties panel 1202 that lists all properties of all entities of the imported data, values of the properties, and graphical indications of the number of such values represented in the imported data. For example, entity properties panel 1202 identifies an Address property 1203 and indicates that a value of “ALAMEDA, Calif.” occurs two times in the imported data, as seen by the bar and numeral at 1206. The graphical bar represents the relative magnitude of the value “2” as compared to all other counts of values shown in the histogram. Links 1208 enable a user to show more, less, all, or no values for the associated entity property.
[0104] At any time during entity resolution, the user can use the Histogram to highlight various data cohorts on the screen. For example, selecting the value “ALAMEDA, Calif.” in the Histogram display 1200 causes selecting in FIG. 8 only those entities that include the selected value-that is, only those individuals who have an address including “ALAMEDA, Calif.” . Thus, the Histogram display 1200 interacts with screen display 802 of FIG. 8 and associated data transfers are performed under control of Entity Resolver 152. By then selecting a “select highlighted” button 810, the user can select just those entities to resolve.
[0105] Further, by interacting with the Histogram display 1200 and selecting data from that display, users can rapidly select different groups of entities to be associated with different matching criteria in subsequent stages of entity resolution. For example, the user might wish to apply one set of matching criteria to all imported data associated with “ALAMEDA, Calif.” in the Histogram display 1200 and a different set of matching criteria to imported data associated with a different property or value.
[0106] When the user is satisfied with the selection of entities, the user may select NEXT button 822 to advance to the second step in entity resolution. FIG. 9 illustrates an example Entity Resolver display associated with choosing matching criteria. In an embodiment, the Entity Resolver display includes a help panel 902, criteria panel 904, a results panel 906, and a properties menu 916.
[0107] In an embodiment, a user first chooses one or more properties of suspects to use to search for existing matches to the entities we are importing. The user navigates the properties menu 916 to select linking terms. As linking terms are selected, the linking terms are displayed in criteria panel 904. As an example, four (4) linking terms are shown in FIG. 9 , indicating that the user is attempting to find a match using the suspects' names, phone numbers, addresses, and email addresses. Linking term 908 provides a specific example of a linking term that will match telephone number values.
[0108] After establishing one or more linking terms, the user can choose types of targets to match to by adding or removing target filters. In the example of FIG. 9 , the type of target is “Person,” meaning that the matching criteria of criteria panel 904 will be matched against “Person” entities in database 108. A user can add other target filters by selecting an “edit” link 920 in results panel 906. For example, the user might want to search on “Cargo Ships” as well as “Person” if the user thought that a person had been mislabeled as a cargo ship in the imported data.
[0109] For both the linking terms and targets, filters can be added to each item by clicking on an “edit” link 912. A linking term may be deleted by selecting a delete button 914.
[0110] A user may select NEXT button to advance. In response, the Entity Resolver performs a search of all entities in the database 108 and attempts to match all the imported data to existing entities in database 108 based on the criteria, and creates and stores data representing all matching entities. The data is organized according to sets of matching criteria.
[0111] In an embodiment, an Entity Resolver screen then displays results of matching. FIG. 10 illustrates an example Entity Resolver screen showing results of matching. The Entity Resolver screen display comprises a help panel 1002, a matching criteria panel 1004, and a matching entity listing 1010. The matching criteria panel 1004 lists which sets of criteria 1006 the data has been linked to existing entities on, and a count 1008 of how many of the imported entities fit in each of the sets of criteria. One or more sets of criteria may be displayed; for example, some of the imported data might match existing entities in database 108 based on “Email” and “Address,” whereas other existing entities might match based on “Email,” “Address,” “Phone Number,” and “Name.” In that case, FIG. 10 would display two rows in matching criteria panel 1004—one row as “Email, Address” and the other row as “Email, Address, Phone Number, Name.”
[0112] Selecting one of the sets of criteria 1006 causes the Entity Resolver 152 to display the matching entity listing 1010, showing the imported entities 1014 that belong to the associated set, and which existing entities 1016 they were matched to, based on those criteria. Thus, if matching criteria panel 1004 includes multiple rows of criteria 1006, selecting a different row causes matching entity listing 1010 to change to show only the entities that matched the selected criteria. If imported data does not match any entities in the database 108, then matching criteria panel 1004 displays a “No Matches” indicator as a row of criteria 1006 and a corresponding count 1008 of the number of imported data entities that matched no existing entity.
[0113] Each row or association of an imported entity 1014 and a matched entity 1016 includes a selection checkbox 1012. In an embodiment, a user selects one or more entity associations for resolution. To select only those entities for which there is one match, a user selects a “select unique matches” button 1018. To resolve those matches that the user has highlighted using selection checkboxes 1012, the user selects a “select highlighted” button 1020. To select all matches, the user selects a “select all” button 1022, and to deselect all matches, the user selects a “deselect all” checkbox 1024.
[0114] If there are incorrect matches for an entity, a user can select a correction icon 1028 adjacent to the match that is incorrect. For example, if an imported entity incorrectly matched multiple existing entities in the database because a name is common, or matched an entity that is clearly incorrect based on name, then the user can remove the match by selecting correction icon 1028.
[0115] When the user is satisfied with the selections, the user selects a “consolidate” button 1026 and moves on to the next set of properties shown in criteria panel 1004. Selecting “consolidate” causes Entity Resolver 152 to perform a merge operation in which the imported data is merged into the matched entities existing in the database 108. The merge operation adds any new values for properties or fields of the imported database into the matched existing entity, so that the matched existing entity is updated with the imported values. In an embodiment, consolidation can be organized as a two-phase commit operation in which a first phase occurs in response to selecting “consolidate” button 1026 and a second phase or commit of the consolidation occurs after the finalization step described in connection with FIG. 11 .
[0116] When the user has completed review of all selections for all properties, the third step in entity resolution is complete and the user selects NEXT button 822 to proceed. In response, the Entity Resolver transitions to the fourth step in entity resolution, relating to checking and finalizing resolution data.
[0117] FIG. 11 illustrates an example screen of the Entity Resolver relating to checking and finalizing. In an embodiment, an Entity Resolver screen comprises a help panel 1102, resolved entities listing 1104, and new entities listing 1106.
[0118] The resolved entities listing 1104 lists the entities that the user is importing that have been matched to existing data, what they have been matched to, the group number (which round of resolution this is), and the criteria that each entity was matched upon. For example, resolved entities listing 1104 comprises an entity column 1110 that lists a name of each entity in the imported data. Resolved entities listing comprises a match column 1112 that lists a name of each existing entity in the database 108 that matches the corresponding entity in column 1110. For example, note that the imported entity “Samir Hamnett” in column 1110 matches the existing entity “Sami Hamnett” in column 1112, having slightly different spelling. Column 1114 lists a group number for a group of data undergoing resolution. Match criteria column 1116 lists all criteria of the imported entity that match the corresponding existing entity.
[0119] If there are any mismatched or unwanted entities, the user selects the corresponding “undo” boxes 1108 and then selects the “undo” button 1120. In response, Entity Resolver 152 removes those entities from this round of resolution. If the user wants to re-resolve those entities, the user can select a “continue resolution” button 1122.
[0120] In an embodiment, new entities listing 1106 displays the new entities that will be created in database 108 because no matches were found in the database during the entity resolution process.
[0121] When the user is satisfied with the displayed information, the user can select a “finish” button 1124 to complete the entity resolution process.
[0122] 2.3.2 Saved Criteria Sets
[0123] In an embodiment, when all desired linking terms are established in criteria panel 904, and the user selects the NEXT button, Entity Resolver 150 prompts the user to enter a name for the linking criteria and saves the linking criteria as a named set of criteria. Alternatively, linking criteria may be saved in response to user selection of the FINISH button 1124 of FIG. 11 . After saving the linking criteria, the user can perform future data importation operations by selecting the saved criteria, as described above for FIG. 6 , reference numerals 612, 620, 614. When saved criteria is used, the user performs entity resolution by interacting with the screen displays of FIG. 8 , FIG. 10 , and FIG. 11 , but choosing matching criteria as seen in FIG. 9 is not necessary and Entity Resolver 150 skips over the functions associated with FIG. 9 .
[0124] In an embodiment, creating and storing matching criteria also includes monitoring and recording user interactions with Histogram display 1200 (FIG. 12 ) and selections of matching criteria in panel 1004 (FIG. 10 ). In this embodiment, the Entity Resolver monitors user interaction of data selections and stores, as saved criteria, the properties that a user selected from the Histogram display 1200, the linking criteria from panel 904 of FIG. 9 , and the matching criteria sets 1006 that the user selected from panel 1004. All such data is associated and stored as a named saved criteria set. Storing may occur in response to user selection of the FINISH button 1124 (FIG. 11 ).
[0125] Thereafter, the user can specify a data source for importation, select the saved criteria at FIG. 6 , and perform an expedited form of entity resolution. In expedited entity resolution, the Entity Resolver 152 automatically performs the first three steps of entity resolution that have been described in reference to FIG. 8 , FIG. 9 , and FIG. 10 and displays a summary screen indicating the results of automatic entity resolution. Alternatively, after the first three steps are performed, the screen display of FIG. 11 is displayed and the user can check and finalize the resolutions. Any entities in imported data that do not match entities of the database 108 are created as new entities in the database. As a result, the system becomes adaptive and entity resolution proceeds far more rapidly for subsequent data sets.
[0126] 2.3.3 Alternate Matching Techniques
[0127] Referring again to FIG. 9 , in an embodiment, the form of matching to perform using the linking criteria of panel 904 is specified by selecting a value from a pull-down menu 910 in each linking term 908. In one embodiment, matching techniques in menu 910 include “exact match” and various kinds of fuzzy matching approaches such as “contains.” In an embodiment, the criteria sets 1006 of FIG. 10 reflect the matching technique that was selected. For example, panel 1004 of FIG. 10 may display “Exact Match: Email, Address” to indicate that an associated count 1008 reflects an exact match of imported data to email and address properties of an existing entity in the database 108.
[0128] In another embodiment, pull-down menu 910 includes a “No Conflicts” selection and matching occurs only if the imported data has no conflicts with entities in the database 108. An example may illustrate the meaning of “no conflicts.” Assume that entities Rosa Klebb and Ernst Blofeld are the subjects of entity resolution and each of Klebb and Blofeld has two addresses associated with it. In database 108, every property can have multiple associated values. Klebb and Blofeld would have “no conflicts” if either (1) Klebb and Blofeld have at least one address that matches or (2) Klebb has address data and Blofeld has no address data. Conversely, Klebb and Blofeld do not have “no conflicts” if any address of Klebb or Blofeld does not match any address of the other. This approach allows an analyst to disambiguate entities based on the lack of data for a particular property as compared to a conflict in known data for the property.
[0129] 2.3.4 “One Click” Approach
[0130] In an embodiment, a one-click approach supports expedited entity resolution. In the one-click approach, a user first selects imported data for two or more individuals that are believed to be the same. For example, referring to FIG. 8 , a user may select two individuals shown in rows of screen display 802 that are believed to be the same. The user then activates a special-purpose menu of options for working with the selected individuals. In an embodiment, the user selects a right side button of a mouse or other pointing device (“right clicks”), and in response, Entity Resolver 150 displays a pop-up menu of functions that can be applied to the selected individuals. In an embodiment, one function in the pop-up menu is “Resolve.” When the user selects “Resolve,” the Entity Resolver 150 automatically performs entity resolution on the selected individuals and displays results. Activation of the “one-click” approach in this manner is equivalent to selecting the “consolidate” button 1026 in the process of FIG. 10 .
[0131] 3.0 Implementation Mechanisms—Hardware Overview
[0132] FIG. 13 is a block diagram that illustrates a computer system 1300 upon which an embodiment of the invention may be implemented. Computer system 1300 includes a bus 1302 or other communication mechanism for communicating information, and a processor 1304 coupled with bus 1302 for processing information. Computer system 1300 also includes a main memory 1306, such as a random access memory (“RAM”) or other dynamic storage device, coupled to bus 1302 for storing information and instructions to be executed by processor 1304. Main memory 1306 also may be used for storing temporary variables or other intermediate information during execution of instructions to be executed by processor 1304. Computer system 1300 further includes a read only memory (“ROM”) 1308 or other static storage device coupled to bus 1302 for storing static information and instructions for processor 1304. A storage device 1310, such as a magnetic disk or optical disk, is provided and coupled to bus 1302 for storing information and instructions.
[0133] Computer system 1300 may be coupled via bus 1302 to a display 1312, such as a cathode ray tube (“CRT”), for displaying information to a computer user. An input device 1314, including alphanumeric and other keys, is coupled to bus 1302 for communicating information and command selections to processor 1304. Another type of user input device is cursor control 1316, such as a mouse, trackball, stylus, or cursor direction keys for communicating direction information and command selections to processor 1304 and for controlling cursor movement on display 1312. This input device typically has two degrees of freedom in two axes, a first axis (e.g., x) and a second axis (e.g., y), that allows the device to specify positions in a plane.
[0134] The invention is related to the use of computer system 1300 for entity resolution. According to one embodiment of the invention, entity resolution is provided by computer system 1300 in response to processor 1304 executing one or more sequences of one or more instructions contained in main memory 1306. Such instructions may be read into main memory 1306 from another computer-readable medium, such as storage device 1310. Execution of the sequences of instructions contained in main memory 1306 causes processor 1304 to perform the process steps described herein. In alternative embodiments, hard-wired circuitry may be used in place of or in combination with software instructions to implement the invention. Thus, embodiments of the invention are not limited to any specific combination of hardware circuitry and software.
[0135] The term “computer-readable medium” as used herein refers to any medium that participates in providing instructions to processor 1304 for execution. Such a medium may take many forms, including but not limited to, non-volatile media, volatile media, and transmission media. Non-volatile media includes, for example, optical or magnetic disks, such as storage device 1310. Volatile media includes dynamic memory, such as main memory 1306. Transmission media includes coaxial cables, copper wire and fiber optics, including the wires that comprise bus 1302. Transmission media can also take the form of acoustic or light waves, such as those generated during radio wave and infrared data communications.
[0136] Common forms of computer-readable media include, for example, a floppy disk, a flexible disk, hard disk, magnetic tape, or any other magnetic medium, a CD-ROM, any other optical medium, punchcards, papertape, any other physical medium with patterns of holes, a RAM, a PROM, and EPROM, a FLASH-EPROM, any other memory chip or cartridge, a carrier wave as described hereinafter, or any other medium from which a computer can read.
[0137] Various forms of computer readable media may be involved in carrying one or more sequences of one or more instructions to processor 1304 for execution. For example, the instructions may initially be carried on a magnetic disk of a remote computer. The remote computer can load the instructions into its dynamic memory and send the instructions over a telephone line using a modem. A modem local to computer system 1300 can receive the data on the telephone line and use an infrared transmitter to convert the data to an infrared signal. An infrared detector can receive the data carried in the infrared signal and appropriate circuitry can place the data on bus 1302. Bus 1302 carries the data to main memory 1306, from which processor 1304 retrieves and executes the instructions. The instructions received by main memory 1306 may optionally be stored on storage device 1310 either before or after execution by processor 1304.
[0138] Computer system 1300 also includes a communication interface 1318 coupled to bus 1302. Communication interface 1318 provides a two-way data communication coupling to a network link 1320 that is connected to a local network 1322. For example, communication interface 1318 may be an integrated services digital network (“ISDN”) card or a modem to provide a data communication connection to a corresponding type of telephone line. As another example, communication interface 1318 may be a local area network (“LAN”) card to provide a data communication connection to a compatible LAN. Wireless links may also be implemented. In any such implementation, communication interface 1318 sends and receives electrical, electromagnetic or optical signals that carry digital data streams representing various types of information.
[0139] Network link 1320 typically provides data communication through one or more networks to other data devices. For example, network link 1320 may provide a connection through local network 1322 to a host computer 1324 or to data equipment operated by an Internet Service Provider (“ISP”) 1326. ISP 1326 in turn provides data communication services through the world wide packet data communication network now commonly referred to as the “Internet” 1328. Local network 1322 and Internet 1328 both use electrical, electromagnetic or optical signals that carry digital data streams. The signals through the various networks and the signals on network link 1320 and through communication interface 1318, which carry the digital data to and from computer system 1300, are exemplary forms of carrier waves transporting the information.
[0140] Computer system 1300 can send messages and receive data, including program code, through the network(s), network link 1320 and communication interface 1318. In the Internet example, a server 1330 might transmit a requested code for an application program through Internet 1328, ISP 1326, local network 1322 and communication interface 1318. In accordance with the invention, one such downloaded application provides for entity resolution as described herein.
[0141] The received code may be executed by processor 1304 as it is received, and/or stored in storage device 1310, or other non-volatile storage for later execution. In this manner, computer system 1300 may obtain application code in the form of a carrier wave.
[0142] Data Importer 150 and Entity Resolver 152 each may be implemented in one or more computer programs, firmware, or a combination. The Data importer 150 and Entity Resolver 152 may be implemented in unitary fashion or as separate modules. Alternatively, a special-purpose computer can be configured with software, firmware, hardware, other logic, or a combination to implement Data Importer 150 and Entity Resolver 152. When Data Importer 150 and Entity Resolver 152 are implemented in software, program instructions may be loaded into electronic digital memory of a computer system for execution by a processor and thus the program instructions become a part of the electronic digital memory.
[0143] 4.0 Extensions and Alternatives
[0144] In the foregoing specification, the invention has been described with reference to specific embodiments thereof. It will, however, be evident that various modifications and changes may be made thereto without departing from the broader spirit and scope of the invention. The specification and drawings are, accordingly, to be regarded in an illustrative rather than a restrictive sense.
Claims
What is claimed is:
1. A non-transitory computer-readable storage medium comprising one or more stored sequences of instructions which, when executed by one or more processors, cause the one or more processors to perform:
receiving imported data comprising imported data entities each having one or more properties each having one or more values;
matching the imported data entities to existing data entities in a database using matching criteria, resulting in creating and storing matched entity information, wherein the matched entity information indicates that a subset of the imported data entities matched a subset of the existing data entities for a subset of the matching criteria, and at least another, different subset of the imported data entities matched another, different subset of the existing data entities for another, different subset of the matching criteria;
causing display of:
the matched entity information, and
for each entity of the imported data entities, options comprising:
a first option to select the entity for consolidation when the entity is matched correctly;
wherein consolidation of the entity comprises adding one or more values from the entity into a matched existing data entity;
a second option to select the entity for further resolution when the entity is to be further resolved;
receiving user inputs comprising a first user input and a second user input;
wherein the first user input selects, for consolidation, a first one or more data entities of the imported data entities that are matched correctly;
wherein the second user input selects, for further resolution, a second one or more data entities of the imported data entities that are to be further resolved;
after receiving at least the first user input, consolidating at least the first one or more data entities but not the second one or more data entities.
2. The non-transitory computer-readable storage medium of claim 1, wherein one of the plurality of matching techniques specifies no conflicts between values of properties of the imported data entities and the existing data entities.
3. The non-transitory computer-readable storage medium of claim 1, further comprising sequences of instructions which, when executed by the one or more processors, cause the one or more processors to perform:
receiving additional imported data for resolution to the existing data entities;
receiving a third user input that specifies a name of a stored criteria set;
loading the stored criteria set;
automatically performing the matching based on the stored criteria set.
4. The non-transitory computer-readable storage medium of claim 1, further comprising sequences of instructions which, when executed by the one or more processors, cause the one or more processors to perform:
receiving a third user input as a selection of one of the imported data entities and one of the existing data entities in the database;
receiving a fourth user input requesting an entity resolution of the selected imported data entity and the selected existing data entity;
loading a named criteria set;
automatically performing the matching based on the named criteria set.
5. The non-transitory computer-readable storage medium of claim 1, further comprising sequences of instructions which, when executed by the one or more processors, cause the one or more processors to perform receiving a third user input that specifies one or more target filters for types of the existing data entities, wherein the matching is not performed for other types of data entities.
6. The non-transitory computer-readable storage medium of claim 1, further comprising sequences of instructions which, when executed by the one or more processors, cause the one or more processors to perform displaying the imported data entities in a histogram display that identifies the imported data entities and corresponding counts of the values of the properties of the imported data entities;
receiving a third user input as selections of one or more of the imported data entities in the histogram display.
7. The non-transitory computer-readable storage medium of claim 1, further comprising sequences of instructions which, when executed by the one or more processors, cause the one or more processors to perform:
displaying, in a first panel of a graphical user interface, the multiple subsets of the matching criteria that were matched, and counts of matched imported data entities in the multiple matched subsets of imported data entities associated with the multiple subsets of the matching criteria;
displaying, in response to a third user input, in a second panel of the graphical user interface, each matched imported data entity represented in the count of a selected subset of imported data entities.
8. The non-transitory computer-readable storage medium of claim 1, wherein the instructions that cause consolidating comprise instructions which, when executed by the one or more processors, cause the one or more processors to perform merging at least the first one or more data entities into a set of matched existing data entities from the existing data entities.
9. The non-transitory computer-readable storage medium of claim 1, further comprising sequences of instructions which, when executed by the one or more processors, cause the one or more processors to perform displaying, in a graphical user interface, results of consolidating the at least the first one or more data entities into a set of matched existing data entities from the existing data entities, the results comprising:
a first list of one or more associations of resolved entities each comprising one of at least the first one or more data entities, one of the set of matched existing data entities, and an associated subset of the matching criteria that was matched for the one or more associations;
a second list of one or more new entities to create in the database based on those of the imported data entities that did not match any of the existing data entities.
10. The non-transitory computer-readable storage medium of claim 1, further comprising sequences of instructions which, when executed by the one or more processors, cause the one or more processors to perform:
displaying, in a graphical user interface, four visually expandable and collapsible graphical regions associated with steps in an entity resolution process;
wherein a first graphical region when expanded comprises a first list of the imported data entities and one or more of the properties and one or more of the values;
wherein a second graphical region when expanded comprises a first panel comprising the matching criteria for matching the imported data entities to the existing data entities and a particular matching technique for each of the matching criteria, and a second panel comprising one or more target filters for types of the existing data entities;
wherein a third graphical region when expanded comprises a third panel comprising the multiple subsets of the matching criteria that were matched, and counts of matched imported data entities of the multiple matched subsets of imported data entities associated with the multiple subsets of the matching criteria, and a fourth panel comprising each matched imported data entity represented in the count of a selected subset of the imported data entities;
wherein a fourth graphical region when expanded comprises a fifth panel comprising a second list of one or more associations of resolved entities each comprising one of the imported data entities, one of the existing data entities, and an associated subset of the matching criteria that was matched for the one or more associations; and a sixth panel comprising a third list of one or more new entities to create in the database based on those of the imported data entities that did not match any of the existing data entities.
11. The non-transitory computer-readable storage medium of claim 1, further comprising instructions which when executed by the one or more processors cause the one or more processors to perform determining whether all the imported data entities were resolved, receiving modified matching criteria applicable to one or more unresolved data entities, and repeating the matching and consolidating steps for the one or more unresolved data entities.
12. The non-transitory computer-readable storage medium of claim 1, further comprising instructions which when executed by the one or more processors cause the one or more processors to perform receiving a third user input that selects the imported data entities from the imported data for resolution, wherein the third user input comprises a semantic criterion referencing one or more properties of the imported data entities, and wherein the imported data comprises one or more other data entities that are not selected for resolution.
13. A machine-implemented method, comprising:
receiving imported data comprising imported data entities each having one or more properties each having one or more values;
matching the imported data entities to existing data entities in a database using matching criteria, resulting in creating and storing matched entity information, wherein the matched entity information indicates that a subset of the imported data entities matched a subset of the existing data entities for a subset of the matching criteria, and at least another, different subset of the imported data entities matched another, different subset of the existing data entities for another, different subset of the matching criteria;
causing display of:
the matched entity information, and
for each entity of the imported data entities, options comprising:
a first option to select the entity for consolidation when the entity is matched correctly;
wherein consolidation of the entity comprises adding one or more values from the entity into a matched existing data entity;
a second option to select the entity for further resolution when the entity is to be further resolved;
receiving user inputs comprising a first user input and a second user input;
wherein the first user input selects, for consolidation, a first one or more data entities of the one or more imported data entities that are matched correctly;
wherein the second user input selects, for further resolution, a second one or more data entities of the one or more imported data entities that are to be further resolved;
after receiving at least the first user input, consolidating at least the first one or more data entities but not the second one or more data entities;
wherein the method is performed by one or more computer systems.
14. The method of claim 13, wherein one of the plurality of matching techniques specifies no conflicts between values of properties of the imported data entities and the existing data entities.
15. The method of claim 13, further comprising:
receiving additional imported data for resolution to the existing data entities;
receiving a third user input that specifies a name of a stored criteria set;
loading the stored criteria set;
automatically performing the matching based on the stored criteria set.
16. The method of claim 13, further comprising:
receiving a third user input as a selection of one of the imported data entities and one of the existing data entities in the database;
receiving a fourth user input requesting an entity resolution of the selected imported data entity and the selected existing data entity;
loading a named criteria set;
automatically performing the matching based on the named criteria set.
17. The method of claim 13, further comprising displaying the imported data entities in a histogram display that identifies the imported data entities and corresponding counts of the values of the properties of the imported data entities;
receiving a third user input as selections of one or more of the imported data entities in the histogram display.
18. The method of claim 13, further comprising:
displaying, in a graphical user interface, four visually expandable and collapsible graphical regions associated with steps in an entity resolution process;
wherein a first graphical region when expanded comprises a first list of the imported data entities and one or more of the properties and one or more of the values;
wherein a second graphical region when expanded comprises a first panel comprising the matching criteria for matching the imported data entities to the existing data entities and a particular matching technique for each of the matching criteria, and a second panel comprising one or more target filters for types of the existing data entities;
wherein a third graphical region when expanded comprises a third panel comprising the multiple subsets of the matching criteria that were matched, and counts of matched imported data entities of the multiple matched subsets of imported data entities associated with the multiple subsets of the matching criteria, and a fourth panel comprising each matched imported data entity represented in the count of a selected subset of the imported data entities;
wherein a fourth graphical region when expanded comprises a fifth panel comprising a second list of one or more associations of resolved entities each comprising one of the imported data entities, one of the existing data entities, and an associated subset of the matching criteria that was matched for the one or more associations; and a sixth panel comprising a third list of one or more new entities to create in the database based on those of the imported data entities that did not match any of the existing data entities.
19. The method of claim 13, further comprising determining whether all the imported data entities were resolved, receiving modified matching criteria applicable to one or more unresolved data entities, and repeating the matching and consolidating steps for the one or more unresolved data entities.
20. The method of claim 13, further comprising instructions which when executed by the one or more processors cause the one or more processors to perform receiving a third user input that selects the imported data entities for resolution, wherein the third user input comprises a semantic criterion referencing one or more properties of the imported data entities, and wherein the imported data comprises one or more other imported data entities that are not selected for resolution.
21. The non-transitory computer-readable storage medium of claim 1, wherein the options further comprise a third option to select the entity for removal when the entity is matched incorrectly, and wherein the user inputs further comprise a third user input that selects, for removal, a third one or more data entities, of the imported data entities, that are matched incorrectly.
22. The non-transitory computer-readable storage medium of claim 1, further comprising instructions which when executed by the one or more processors cause the one or more processors to perform, after receiving the user inputs, resolving the second one or more entities that are selected for further resolution.
23. The non-transitory computer-readable storage medium of claim 1, wherein the matched entity information is a first matched entity information, further comprising instructions which when executed by the one or more processors cause the one or more processors to perform, after receiving the user inputs, resolving the second one or more entities that are selected for further resolution at least in part by:
matching the second one or more entities to the existing data entities using at least one matching criteria that is different from the multiple subsets of matching criteria, resulting in creating and storing a second matched entity information, wherein the second matched entity information indicates that a subset of the second one or more entities matched a subset of the existing data entities for the at least one matching criteria;
causing display of:
the second matched entity information, and
one or more options to select one or more correctly matched entities of the second one or more entities for consolidation;
receiving one or more user inputs selecting the one or more correctly matched entities of the second one or more entities for consolidation;
after receiving the one or more user inputs, consolidating the one or more correctly matched entities of the second one or more entities for consolidation.
24. The non-transitory computer-readable storage medium of claim 1, further comprising instructions which when executed by the one or more processors cause the one or more processors to perform:
receiving a third user input that selects the imported data entities from the imported data for resolution, wherein the imported data comprises one or more other data entities that are not selected for resolution.
25. The non-transitory computer-readable storage medium of claim 1, further comprising instructions which when executed by the one or more processors cause the one or more processors to perform:
receiving a third user input that specifies the matching criteria for matching the imported data entities to the existing data entities, wherein each of the matching criteria comprises a particular matching technique from a plurality of matching techniques.
26. The non-transitory computer-readable storage medium of claim 1, further comprising instructions which when executed by the one or more processors cause the one or more processors to perform:
receiving a third user input that selects the imported data entities from the imported data for resolution, wherein the imported data comprises one or more other data entities that are not selected for resolution;
receiving a fourth user input that specifies the matching criteria for matching the imported data entities to the existing data entities, wherein each of the matching criteria comprises a particular matching technique from a plurality of matching techniques
based on the third user input and the fourth user input, storing a named criteria set for use in subsequent entity resolution operations.
27. The non-transitory computer-readable storage medium of claim 1, wherein the matched entity information is organized in multiple matched entity data sets respectively associated with multiple different subsets of the matching criteria that were matched.
28. The method of claim 13, wherein the options further comprise a third option to select the entity for removal when the entity is matched incorrectly, and
wherein the user inputs further comprise a third user input that selects, for removal, a
third one or more data entities of the imported data entities that are matched incorrectly.
29. The method of claim 13, further comprising, after receiving the user inputs, resolving the second one or more entities that are selected for further resolution.
30. The method of claim 13, wherein the matched entity information is a first matched entity information, further comprising, after receiving the user inputs, resolving the second one or more entities that are selected for further resolution at least in part by:
matching the second one or more entities to the existing data entities using at least one matching criteria that is different from the multiple subsets of matching criteria, resulting in creating and storing a second matched entity information, wherein the second matched entity information indicates that a subset of the second one or more entities matched a subset of the existing data entities for the at least one matching criteria;
causing display of:
the second matched entity information, and
one or more options to select one or more correctly matched entities of the second one or more entities for consolidation;
receiving one or more user inputs selecting the one or more correctly matched entities of the second one or more entities for consolidation;
after receiving the one or more user inputs, consolidating the one or more correctly matched entities of the second one or more entities for consolidation.
31. The method of claim 13, further comprising:
receiving a third user input that selects the imported data entities in the imported data for resolution, wherein the imported data comprises one or more other data entities that are not selected for resolution.
32. The method of claim 13, further comprising:
receiving a third user input that specifies the matching criteria for matching the imported data entities to the existing data entities, wherein each of the matching criteria comprises a particular matching technique from a plurality of matching techniques.
33. The method of claim 13, further comprising:
receiving a third user input that selects the imported data entities in the imported data for resolution, wherein the imported data comprises one or more other data entities that are not selected for resolution;
receiving a fourth user input that specifies the matching criteria for matching the imported data entities to the existing data entities, wherein each of the matching criteria comprises a particular matching technique from a plurality of matching techniques based on the third user input and the fourth user input, storing a named criteria set for use in subsequent entity resolution operations.
34. The method of claim 13, wherein the matched entity information is organized in multiple matched entity data sets respectively associated with multiple different subsets of the matching criteria that were matched.
35. The method of claim 13, further comprising receiving a third user input that specifies one or more target filters for types of the existing data entities, wherein the matching is not performed for other types of data entities.
36. The method of claim 13, further comprising:
displaying, in a first panel of a graphical user interface, the multiple subsets of the matching criteria that were matched, and counts of matched imported data entities in the multiple matched subsets of imported data entities associated with the multiple subsets of the matching criteria;
displaying, in response to a third user input, in a second panel of the graphical user interface, each matched imported data entity represented in the count of a selected subset of imported data entities.
37. The method of claim 13, wherein consolidating comprises merging at least the first one or more data entities into a set of matched existing data entities from the existing data entities.
38. The method of claim 13, further comprising displaying, in a graphical user interface, results of consolidating at least the first one or more data entities into a set of matched existing data entities from the existing data entities, the results comprising:
a first list of one or more associations of resolved entities each comprising one of at least the first one or more data entities, one of the set of matched existing data entities, and an associated subset of the matching criteria that was matched for the one or more associations;
a second list of one or more new entities to create in the database based on those of the imported data entities that did not match any of the existing data entities.
Patent Citations (9)
| Patent | Date | Inventor | Cited By |
|---|---|---|---|
| US6006242(A) | 1999-12-01 | Poole et al. | Examiner |
| US7403942(B1) | 2008-07-01 | Bayliss | Examiner |
| US7757220(B2) | 2010-07-01 | Griffith et al. | Examiner |
| US7912842(B1) | 2011-03-01 | Bayliss | Applicant |
| US2004/0006523(A1) | 2004-01-01 | Coker | Applicant |
| US2004/0210763(A1) | 2004-10-01 | Jonas | Examiner |
| US2007/0067285(A1) | 2007-03-01 | Blume | Applicant |
| US2008/0005063(A1) | 2008-01-01 | Seeds | Applicant |
| US2010/0070531(A1) | 2010-03-01 | Aymeloglu et al. | Applicant |
Non-Patent Literature (10)
- Lim, Ee-Peng, et al., Resolving Attribute Incompatibility in Database Integration: An Evidential Reasoning Approach [online]. Published by Department of Computer Science, University of Minnesota. pp. 1-10. [Retrieved on Sep. 24, 2008]. Retrieved from the Internet: <URL:http://reference.kfupm.edu.sa/content/r/e/resolving—attribute—incompatibility—in—d—531691.pdf>.Applicant
- A Real-World Problem of Matching Records. Entity Resolution. Published by the Universitat Pompeu Fabra [online]. Nov. 2006 [retrieved on Sep. 24, 2008]. pp. 1-16. Retrieved from the Internet: <URL: http://grupoweb.upf.es/bd-web/slides/ullman.pdf>. Center for research, Chile, South America.Applicant
- L.W. Young, USPTO, International Search Report & Written Opinion, mailed by US Patent & Trademark Office and published by USPTO, United States and/or International Bureau of WIPO, Geneva, Switzerland, Dec. 5, 2008, 8 pages.Applicant
- Litwin, Witold, Multidatabase Interoperability, IEEE Computer, Dec. 1986, pp. 10-18, retrieved from the Internet: http://ceria.dauphine.fr/mdb-interoperability.pdf Published by IEEE, New York, NY US.Applicant
- Qiang, Bao-Hua, a Mutual-Information-Based Approach to Entity Reconciliation in Heterogeneous Databases, Proceedings of 2008 Int'l Conf. on Comp. Sci. & Software Eng., published by IEEE Computer Society, New York, NY US, pp. 666-669. Retrieved from the Internet: ftp://ftp.computer.org/press/outgoing/proceedings/csse08/data/3336a666.pdf.Applicant
- Zhao, Huimin, Entity matching across heterogeneous data sources: An approach based on constrained cascade generalization, Data & Knowledge Engineering 66 (2008), May 4, 2008, pp. 368-381. Published by Elsevier B.V., Amsterdam, Netherlands. Retrieved from the Internet: http://hem.hj.se/˜life/db/Entity%20matching%20across%20heterogeneous%20data%20sources%20-%20An%20approach%20based%20on%20constrained%20cascade%20generalization.pdf.Applicant
- International Searching Authority, “International Search Report”, PCT/US08/79072, dated Dec. 5, 2008, 8 pages.Applicant
- Claims, PCT/US08/79072, 11 pages.Applicant
- European Patent Office, “Exam Report” In application No. 08 839 003.4—1951, dated Jun. 12, 2013, 6 pages.Applicant
- Current Claims in application No. 08 839 003.4—1951 dated Jun. 2013, 5 pages.Applicant