Collaborative Internet Data Mining Systems

The 1990's have been remarkable for the explosive growth of the Internet from an specialized system used by academia to a widespread medium for the transfer of information and electronic commerce.

The "Internet" was developed in the 1970's with funding from the Department of Defense to interconnect university computer systems. Until recently, Internet usage was largely confined to academic circles to send e-mail, chat and access remote files and computer resources. The Internet application programs to perform e-mail, chat and the remote access of data were, in large part, command intensive and did not provide an easy to use graphical user interface.

The explosive growth of the Internet has been fueled, in large part, by the development and wide adoption of the HyperText Transfer Protocol (HTTP). HTTP is the Internet protocol used to transfer documents and other Multipurpose Internet Mail Extensions (MIME) type data between systems. HTTP is the protocol on which the World Wide Web ("the web") is based. To the Internet user, the web is an easy to use graphical user protocol on which the World Wide Web ("the web") is based. To the Internet user, the web is an easy to use graphical user interface that provides "point-and-click" access to data from an enormous number of remote computers.

The communication technology of the web can be explained by analogy to the Open System Interconnect Model (OSI) for computer communication. HTTP resides above the Transport Control Protocol/Internet Protocol (TCP/IP) layers and provides a transfer protocol between the web server and the browser client. TCP/IP divides networking functionality into only four layers: (1) a network interface layer that corresponds to the OSI link layer, (2) an Internet layer which corresponds to the OSI network layer (3) a transport layer which corresponds to the OSI transport layer and (4) an application layer which corresponds to the session, presentation and application layers of the OSI model. The web browser (client) may correspond to the application layer of the OSI model and Hyper-Text Markup Language may correspond to the presentation layer.

The Hyper-Text Markup Language (HTML) is the software language in which most of the web is written. HTML is basically ASCII text surrounded by HTML commands in angled brackets. HTML commands are interpreted by a web browser to determine how to display a web page.

The web, as a whole, is made up of web page servers and web browsers that provide a hardware and operating system independent environment. A web browser is an application program that interprets and displays HTML pages. The web is hardware and operating system independent because of the common HTTP and HTML protocols and languages used between the web servers and the browser client applications.

HTML web pages usually contain links or HyperText that point to other HTML pages on the web. By pointing and clicking on these links, a user can skip or "surf" from page to page on the web.

A primary function of a web browser is to display the page located at an Universal Resource Location (URL) address. A URL is an address that includes the protocol to reference the data, the system path and data filename. The data file addressed by the URL data filename is located on a server.

One aspect of the way in which HTML supports the display of data is through the support of "frames." Frame support can be defined as the ability of a web browser to split the browser display area into separate "framed" display areas. Each display area, or frame, can contain information from a separate web page and/or point to a separate URL address. Frames can be created to present the user with a simultaneous coordinated presentation of multiple frames while maintaining the look-and-feel of a single web page.

Another feature in most web browsers is the ability to "bookmark" a page. Typically, the web browser stores a plurality of bookmarked pages in a non-volatile storage mechanism where they may be retrieved when the browser is reactivated. A bookmark is a reference to a single URL address.

The use of bookmarks presents a problem for web pages that are designed for display as multiple coordinated, or framed, web pages. A bookmark is a reference to a single URL address. A frame based web page, however, simultaneously displays multiple URL addressed web pages. Therefore, a bookmark created when viewing a frame based web page stores only one URL address, where multiple URL addresses are required to properly display the frame based data. When the user attempts to re-access the page with the bookmark, the browser display will only load one frame, which provides only part of the coordinated framed presentation of data.

Another service found on the web is the ability to search for information. Search services such as Yahoo.TM., Excite.TM., Lycos.TM., Infoseek.TM. and Hotbot.TM. provide a means for searching web pages and other information on the Internet that return references to URL address of web pages and other data that satisfy the search criteria. For the most part, these search services use a keyword search to find web pages and other information that satisfies the search request.

The web has created a forum that provides a very low cost way to publish information, views and opinion. This inexpensive way to publish information has resulted in an explosion in the amount of data available on the web. Ironically, the success of the web has created its own problems, namely how to separate informed views and authoritative information from uninformed views and unreliable information. The present invention addresses this problem by providing useful, novel and non-obvious methods and apparatus to point to and find quality information available on the Internet.

The present invention provides methods and apparatus for managing, implementing and creating a collaborative Internet data mining system. The collaborative data mining system is comprised of many human "guides" that maintain web sites on their respective topic areas. The guides may use conventional search services, their own knowledge and judgment and their knowledge of where information may be found on the Internet to construct high quality and authoritative web pages. The collaborative data mining system uses automated methods and apparatus to process the web pages created by the guides. The processing automatically "brands" the web pages by inserting uniform characteristics and information into the pages. The system may then sell advertising on the branded network and remunerate the guides based on predetermined criteria.

More specifically, the collaborative data mining system is accomplished through a unique computer based methodology for (1) selecting, training and policing Internet guides for pre-determined topic areas, (2) processing pre-determined forms, formats and commands to create co-branded web pages that provide a coordinated look and feel across many web pages and (3) an automated revenue distribution system for compensating guides based on a predetermined performance measurements.

One aspect of the present invention provides an automated system for use in conjunction with a pre-determined form or template based methodology to generate web pages that automatically maintain the simultaneous and coordinated presentation of framed based data.

Another aspect of the present invention is the use of server side includes to replace "hard coded" HTML with references to "library" objects thereby increasing efficiency of the coding process, page loading, and the propagation of changes to web pages.

Another aspect of the present invention is the creation of novel procedures, system templates, scoring methods, and support tools to identify and solicit quality web producers and web artists to affiliate themselves with present invention's branded Internet server.

Another aspect of the present invention is the use of an automated system for designating and managing a plurality of guides in training.

Yet, another aspect of the present invention is in a mass mentoring system in which to improve and develop large numbers of guides and potential guides to improve their sites and meet the standards required to be a guide.

Yet, another aspect of the present invention is providing a novel, economical and expeditious way in which to maintain the highest possible level of quality and compliance across a high volume network while maintaining low cost and efficiency in developing and manufacturing information "content."

FIG. 1 shows a block diagram of the collaborative Internet invention as having a guide acquisition system (2) an application processing system (4), a guide authoring system (6), a mass mentoring system (8), the global HTML data bank (10), a collaborative page generator system (12), an accounting system (14) and advertising database system (16), the frames system (18), the Internet in general, as denoted by reference (22), a quality control process (24), an automated training and management system (25) and at least one web browser (20).

FIG. 2 shows a detailed diagram of the guide acquisition system (2). It is understood that the elements in FIG. 2 denote processes that may be executed on the systems of the present invention.

FIG. 3 shows a detailed diagram of the application screening system (4) in which a guide identified from the guide acquisition system (2) receives further processing to determine, inter alia, whether to select a particular guide for guide training.

FIG. 4 shows a detail diagram of the mass mentoring system (8). Here, guides from the application processing system (4) receive training and feedback to use the templates and achieve the performance selection criteria demanded by the system. The pre-determined form methodology that interacts with the collaborative page generator system (12) and the global HTML data include files (10) to create on-line content.

FIG. 4A provides a detailed flow diagram of the application processing system and the mentoring system.

FIG. 5 shows the detailed diagram of the present invention's quality control process. Here, sites may be checked, observed and tested to "police" the performance, quality and activity of the guides of the present invention.

Figure 5A provides a detailed flow diagram of the logical steps that may execute after a guide graduates from training.

FIG. 6 shows the detailed diagram of the frames system (18) which provides a means for supporting frame based data.

FIG. 7 shows a block diagram of an example of how frame based data may be used.

FIG. 8 is a detailed diagram of the collaborative page generator (12) in which the global HTML data (10) and advertising data (16) are brought together to create page content for the frames system (18) and the Internet user at a web browser (20).

FIG. 9 shows the steps in the template production system starting from the guide's personal computer, through the processing steps and finally to the live production site.

FIG. 10 shows a detailed diagram of the "CHEWY" process and how it is used to change the guide template to a server ready HTML page.

FIG. 11 provides a detailed diagram of the logical file structure arrangement of the "zshare54z" shared directory structure.

FIG. 12 shows a logical diagram of the directory structure for a guide site.

FIG. 13A shows a functional diagram for the automated training and management system used to manage and control a collaborative data mining system.

FIG. 13B provides a detailed diagram of a control screen used to add a new web site to the taxonomy of a collaborative Internet data mining system.

FIG. 13C provides a detailed diagram of a control screen used to modify a web site, that is maintained or used with the collaborative data mining system.

FIG. 13D provides a detailed diagram of a control screen used to add an application to the training system.

FIG. 13E provides a detailed diagram of a control screen used to modify an application in the system.

FIG. 13F provides a detailed diagram of an ATMS control screen to group new applicants into classes and assign graduation dates.

FIG. 13G provides a detailed diagram of a control screen that may be used to modify account information.

FIG. 13H provides a detailed diagram of additional fields for the control screen provided in FIG. 13G.

FIG. 13I provides a detailed diagram of a control screen that may be used to keep track and/or control part of the mentoring program.

FIG. 13J provides a detailed diagram of a control screen that may be used to control and/or track contract information.

One aspect of the present invention provides a means for providing a "brand name" look and feel to a plurality of web pages by using frames to provide a consistent banner across the pages that reside on the network regardless of how a user "surfs" into the network. This aspect of the present invention provides a brand look and feel to the network while maintaining the ability to randomly surf to a web page of interest.

FIG. 1 provides an overview of the system elements of the collaborative data mining system. The present invention achieves its co-branded look and feel through the use of the guide authoring system (6), in conjunction with the collaborative page generator system (12) and the frames system (18). The guide authoring system (6) provides a guide with predetermined templates that are developed in conjunction with the global HTML data (10) and the frames system (18). The templates are developed by creating a finished web page and then removing the global brand elements and replacing them with "include" comments. The remaining page, with global sections replaced with the "include" comments and section blocked off for the guide to insert content form the basis for a HTML template.

Another aspect of the present invention is its ability to locate a very large number of Internet savvy guides in a short period of time through the guide acquisition system (2). The guide acquisition system is represented as block 2 in FIG. 1 and works in conjunction with the application processing system (4), the guide authoring system (6), the mass mentoring system (8) and the Automated Training and Management System (ATMS) (7). As provided above, guides may be contractors responsible for the content on specific topic areas on the present invention's web sites. The guide acquisition system (2) may, under direction, contact Internet "surfers" who may be connected to the Internet. An "Internet surfer" is a person or entity that is very familiar with content available on the Internet. The present invention provides an Internet surfer with a framework for searching the Internet, e.g., topics to search and a detailed set of instructions with the specifics of how to find quality sites on the web.

The present invention may use a set of standard measurements that indicate the likelihood that a site has the qualities that will result in a worthwhile contact for the application processing system (4). The surfers may send the guide acquisition system (2) screen captures of their search results. The screen captures may be evaluated based on predetermined quality criteria standards.

The application processing system (4) may generate a form e-mail to guide candidates inviting them to work with the collaborative data mining system. Candidates who express an interest may be directed to the application processing system 4 where they may begin the application process.

The application processing system (4) may efficiently convert a large number of applications from its outbound guide acquisition system (2) or inbound marketing recruitment. Using conventional labor intensive methods to review these applicants may not be cost efficient and indeed may be cost prohibitive for the processing volume required by a collaborative data mining system. Applicants may be required to download, complete and submit a template application. The application may require the submission of individual creative content, a detailed curricula vitae of the applicant, and answers to many questions about their particular interest, background, software, equipment and the like. The application may be designed to identify and isolate the qualities essential to a guide. The completed applications may be downloaded, entered into a database, and screened by entry level personnel through the use of standardized criteria.

Another aspect of the present invention is the mass mentoring system (8). The mass mentoring system (8) may use a small team of individuals to coach and coordinate the mentoring process for large volumes of guides and guide trainees identified and screened by application processing system (4). One-to-one development and training may be inefficient given the number of affiliates or guides needed by a reasonably sized collaborative data mining system. The mass mentoring system (8) may employ a training process, discussed further below, which may last up to three weeks. During the training process the present invention may send out predetermined e-mails to the guides covering specific topics key to their development. These e-mails may be specifically tailored to address the typical growth curve and problems of a "proto-guide."

The mass mentoring system may establish a project schedule for training guides on the collaborative data mining system. The mass mentoring system (8) may identify weekly milestones for the process and assigns tasks the guides-in-training should try to accomplish in the first week. The mass mentoring system may also schedule the guides for group chat sessions to discuss their questions and issues. The mass mentoring system may also establish on-line bulletin boards to post questions to the staff at the mass mentoring system or to read the questions and answers of other guide trainees.

Once a guide is accepted for training they may be considered "affiliates" with the collaborative data mining system. The affiliates may be grouped in classes and the progress of the affiliates inside the class may be compared to others inside the class and to standards developed from prior classes. The classes give a means of peer support and may reduce the need for staff guidance.

The facilitated communication may result in guides solving the problems of guides. The mass monitoring system (8) may also assign an experienced guide to each class to act as a peer mentor and give advice and/or guidance to the guides-in-training. Through this novel process, the present invention may mentor up to a couple of hundred guides-in-training with a staff of small mentors, which may provide an economical way to train many guides.

Another aspect of the present invention is the quality control process (24). The present invention employs methods and systems to identify, screen, and develop guides for its Internet service. One key to establishing and keeping its consumer brand image for the present invention's network of guides is maintaining high quality standards across the many, many sites of the network. It may be cost prohibitive to rely on brute force methods of one-to-one inspection and review of individual sites. To wit, the present invention may use new approaches to the quality control (QC) process. First, it may use methods and programming code to make daily automated checks for bad links, breaks in standard template requirements, file download size, and organizational structure of a site. Second, it may use a standardized checklist to check for quality issues such as proper grammar, required site maintenance, timeliness of content, depth of contextual links and other criteria. The guide may receive a standardized e-mail "report card" with feedback on their site as well as specific tips for site enhancement.

Another aspect of the present invention is the frames system (18). The frames system (18) may provide a means for verifying that frame based data, or data that is formatted for presentation in framed format, is in fact loaded with the appropriate coordinated frame of data.

The frame system (18) also assures that the appropriate frame set for the designated topic is loaded. The data mining system of the present invention organizes data into topic areas such as, for example, health and business. Each topic area may have sub-topics. Each topic may have multiple pages (a "page set") contained inside a frame set. Each page may have a unique URL. Someone desiring to link to a page in the page set other than the first page in the set, e.g., the topic home page, would normally use that page's URL for the link. However, a link to a page other than the topic home page may not load the frame set correctly. Accordingly, the page may not be viewable as intended and could result in frame sets inside a frame set or no frame set at all. The frame system (18) provides a means for assuring that the systems frame based data is properly loaded with the appropriate frame or frames.

Another aspect of the present invention is the collaborative HTML processing and collaborative HTML page generator system (12). The data mining system of the present invention may require a clear navigation system across multiple interest areas and related topics. The system architecture is, therefore, reasonably flat with many similarly designed pages existing at the same level. In its simplest form, the "taxonomy" of the network consists of one home area (layer 1) leading down to thirteen plus or minus interest areas (layer 2) leading down to thousands of topics (layer 3). Each layer may share certain elements of a characteristic design across areas and topics. However, there are other elements that may vary depending on layer or topic. The time and cost to hard code each of these elements during creation or modification may be cost prohibitive and stifle the creative vision of the network. The collaborative page generator system (12) provides a means for processing data input from the guide network to produce complete HTML document for use with the live network.

A final aspect of the present invention is the automated training management system (ATMS) (7) that provides a means for if automatically tracking and managing the assets on the data mining network. The ATMS provides a report on the progress of each guide in training as well as providing a means for designating which web pages and topic areas are ready for the "live" network.

The network production process is shown in FIG. 9. A guide who is trained to use the template system (discussed further below) creates HTML documents (804) for her topic area on her personal computer (802). The guide may upload the template based HTML documents to a directory called "/mcupload" (824) on a server computer (806). The uploading of files may be accomplished with the file transfer protocol (FTP) or other conventional methods for transferring data files.

The design container processing tool or "CHEWY" process (812) may be a Windows NT service, a UNIX deamon, or similar computer program that is designed to continuously execute or periodically poll the /mcupload directory or equivalent directory (808). When the CHEWY process (812) finds a file in the /mcupload directory the CHEWY process may open the document and processes the commands contained in the file, described further below, in conjunction with the container definitions (810) to generate "completed HTML" pages. The CHEWY process may then output the completed HTML pages to a "test" directory (814). It is understood that the "test" directory may be any directory that is accessible or able to be viewed on the guide's browsers (816). The guide may then access the complete HTML files to verify that the CHEWY process accurately parsed and processed the HTML template documents (804). If the guide finds a problem in the completed HTML files, the guide may attempt to repair the completed HTML documents (816) or correct the template HTML documents (804) and re-submit the documents to CHEWY through the /mcupload (808) directory. Once the guide is satisfied with the completed HTML documents, the guide may transfer the HTML documents to the production site (818). The "live" production site (820) may serve the completed and approved HTML documents (822).

FIG. 10 provides a more detailed diagram of the CHEWY process. Once a HTML document is delivered to the /mcupload directory, the CHEWY process may open and process the document (850). HTML document (850) may contain codes in angled brackets "<>" that may be interpreted by the CHEWY process. The CHEWY process may also use the HTML document filename as information as to what global information may be used to generate completed HTML pages. For example, if the filename is "mbody.htm" the name of the file designates that CHEWY "tells" the document where to get design information and where to insert it in the template (854). One source of the design information may be the shared components repository--where `##` is the template version number for the "zshare##z" (858) directory, template (860) subdirectory and "##stndrd" (862) subdirectory which contains the "mbody.hi," "mbody.INChi," "mbody.INClo," and "mbody.lo" files. The CHEWY process may insert the files in the designated locations in the HTML page to create a web page that is suitable to "serve" (870). The suitable web page (870) may contain the appropriate server side "include" commands. The HTML include command is like the "#include" compiler instruction for the conventional "c" programming language. When a browser accesses the web page, the browser may receive a HTML page that contains the HTML commands necessary to display a complete HTML page, e.g., the #include instruction may be replaced with the HTML code referenced by the #include command (874).

FIG. 11 provides a logical diagram for the file structure for the shared files in the collaborative data mining invention. The "zshare##z" (858) file is shown with the subdirectories depicted one level below in the HTML (890), images (892), js (894), nav (898), notice (898), search (900), ssw (902), events (904) and template (860) subdirectories. These subdirectories may in turn contain subdirectories of their own such as the subdirectories under the images (892) subdirectories including arts, business, careers, computer, culture, family, health, bobbies, issues, living, local, sports and travel.

FIG. 12 provides an overview of a directory structure that may be used by a guide in a collaborative data mining system. The structure may be described as subdirectories depending from a root directory (912). The root directory (912) may be the default or starting point for the guide. The first subdirectory may be a "/library" (914) of further subdirectories to useful guide data, such as, weekly data (924), graphics (926), personal directories (928) and an archive directory (930). The next subdirectory may store subdirectories of information that is useful for the system (916). Information for the system (932) may include: a subdirectory of site specific pictures, graphics and/or images; a subdirectory of hub specific pictures, graphics and/or images; pages used by the system; search parameters that may be useful for outside search services to help find the site; a subdirectory for chat room parameters; programs, and/or other chat room related data; a subdirectory for boards such as board parameters, programs, and/or other board related data; a subdirectory for dynamic data for programmatically created web pages; a subdirectory for a template configuration file; a subdirectory for site parameters such as URL, hub and navigation parameters, a subdirectory for advertising data; a subdirectory for navigation parameters; a subdirectory for content ratings, and a locked subdirectory for system only access. The guide directory structure may include the upload (808) and test (918) directories described in further detail herein. The guide directory structure may also have a subdirectory for guide control center information (920).

Finally, the example of a guide directory structure may include a delivery (922) directory and the zshare##z directory described further herein.

Thus, the present invention employs a novel method that may eliminate the requirement to individually hard code HTML instructions at each of its sites for different logos, colors, artwork, etc. The present invention may also employ a novel software approach that builds into the web pages HTML instructions to "look for" missing site construction information, e.g., the color set for the site, at the server level. During the build process, the server is "told" the topic area of the site and the build process responds with information for insertion into a site's HTML code. Because of this novel approach, each site can be "manufactured" without hard coding all the design elements. This facilitates the scaleability of the design across many sites. It may also make possible collaborative work on the sites in a new way because certain elements of the system do not reside in the site's code but instead reside in a server working with many sites. This methodology may also make it possible to propagate changes to sites or groups of sites at the server level without the need to edit HTML code at all the sites thereby greatly increasing design and maintenance flexibility.

The automated guide recruitment process of the guide acquisition system (2) is further described in FIG. 2. The automated acquisition system may begin with an assignment of priorities and topics for the outbound search (102). The system may also use a set of guidelines in selecting a site manager such as a predetermined academic background or predetermined experience standards. Using these parameters, the universe of Internet sites (100) may be searched for candidates as web guides (104). A list of sites (106) is selected from the universe of Internet sites (100). The list of sites is passed (108) to a guide selection process (110). The guide selection process (110) may use predetermined subjective or objective criteria to select potential guides. A standardized e-mail may be issued (112) to the selected potential guide (114). The potential guide may respond (116) to the standardized e-mail (114) and request more information, decline the invitation or accept the invitation and be invited to apply (118). When a potential guide is invited to apply, (118) a standardized e-mail (120) may be sent conveying the invitation (120). The potential guide may reply to the invitation (112) by declining the offer or submitting to the application process (124).

FIG. 3 may provide a diagram of the process steps used to further process a potential guide's application to the network. Prospective guides may enter the system and are instructed to go to the application site (200). At the application site, the potential guide may receive the site application (202). The potential guides may complete and submit the application (204). The application may then be objectively or subjectively scored (210) by the data mining system and/or staff (206). The top scoring applications may be passed up for a higher level review (208). The application process may then reach a decision point (214) to determine a course of action for the guide application. If the guide application appears promising, the guide applicant may be issued an e-mail urging improvement and reapplication (216). If the guide application does not meet the performance criteria then the applicant may be sent an e-mail declining the application. If the application passes the performance criteria then e-mail may be sent to the applicant accepting the application (218). Once a guide application is accepted the applicant may enter guide training (220).

FIG. 4 may provide a detailed diagram of the guide training process. Guides may enter the guide training process at block (402). Here, guides are assigned passwords and may be given access to the training system. A standard work schedule may be assigned, the guide may be assigned to a class, and a mentor may be assigned to the class at block (406). The guide applicant may be given the full template downloads to begin constructing their sites and following the training schedule (404). Standard e-mail messages may be sent to properly inform the guides of their responsibilities (408). The guide applicant may begin to construct their sites and receive assistance from the mentor (410). The guide applicant may view special sites built for trainees (412) and e-mail questions to the mentor (414). The system may track the applicant's progress through system checklists (416). The guide applicants may then receive feedback and instruction on their site construction in the mentoring and monitoring phase of training (418). Guide applicants may be scheduled to participate in on-line chat groups on scheduled topics (420). Guide applicants may also log on on-line bulletin boards to retrieve FAQs and the like to assist the guides with site construction. Once a guide completes the template HTML documents and successfully submits them to the CHEWY process, described above (428), guide performance may be evaluated against the performance of other guides (430). If the guide has produced an acceptable site then the site is accepted and transferred to the live site (436). If a guide continues to unsatisfactorily perform, then the guide is sent a thank you and dismissed (434). Guides that produce promising sites may be sent back for more training (402).

FIG. 4A may also provide an application processing system and mentoring system for the collaborative data mining system. Here, a potential applicant may download an application from a web site or HyperText link from the Internet (450). The applicant may receive a web page or a file download by FTP or other conventional file transfer means that contains the application materials. Once the applicant receives the application, the applicant may provide the application details and submit the application to the collaborative data mining system. Once the application is received by the collaborative data mining system, the data may be entered manually or automatically into a database (452). The application may then be reviewed either automatically or through intervention to determine whether the application fits into the taxonomy of the collaborative data mining system (454). The taxonomy of the collaborative data mining system is the structure of the universe of data or topics sponsored by the system, e.g., the genus and species of the topics supported. If the application does not fit into the taxonomy the application may be rejected (456). If the application does fit into the taxonomy then the application may be reviewed (458). If the application is accepted, applications may be judged by predetermined criteria as described above, and if appropriate the applicant is assigned to a class (460). If the application is rejected, a rejection letter or e-mail may be sent to the applicant (456). Once an applicant has been accepted for training by the network, the site may be marked as not available for further applications (462). This step may encourage the applicant to complete the guide training process. After an application is accepted a welcome letter or e-mail may be sent to the applicant (464). The welcome letter or e-mail may contain an address for the training cite and a password and user identification. The guide applicant may now be called an "affiliate" by the network and may begin building a site with the template and collaborative techniques described herein (466). If the affiliate fails or quits the process at this juncture, the system may automatically or through intervention note that the site is again available for applications (468). Hopefully, however, the affiliate will successfully complete the guide training process and proceed to graduation from the training process(470).

FIG. 5A may provide a block diagram of the functions the system may perform at the graduation of a guide. Initially, the guide is either recommended for graduation by the mentor and/or the editor (350). If a guide is not recommended for graduation, the affiliate may be terminated and the database updated to show the availability of this site topic (352). A termination letter or termination e-mail may be sent at this step (352). If a guide is graduated from the training process several steps may occur within the system. First, a contract may be sent to the affiliate which may provide the terms and conditions of the relationship with the system (352). Second, art may be created and uploaded onto the system that may be necessary to support the affiliate (356). Third, keywords and description files for the search and/or support for the site may be created (358). Finally, a photo of the affiliate may be input into the system to provide a picture of the guide to the system's users (360). After these steps are performed, the network may provide the site for a final quality control check by the mentor and/or the editor (362). The quality control check may include the automated site checking process described herein. After the quality control check, the site may be subject to a final edit and review by the hub editor (364). The hub editor may be assigned with editorial responsibility for a specific genus of topics and related topical sites on the system. The system may then promote the site to the production server, create a DNS entry and set permissions for the chat room, bulletin board and mail box subsystems (366). The site may be entered in the navigation system for the collaborative data mining system so that users may navigate to the site from within the network (368). Finally, a welcome letter may be sent to the guide (370).

The application processing, guide training, and the collaborative data mining system in general may be monitored and controlled by the ATMS system. FIG. 13A may provide the main screen of an ATMS system for a collaborative data mining system (1000). The main screen may provide a status bar to show the status of the system database (1001). In general, the ATMS main screen may be divided into the following logical components: Initial site setup (1002) which may include the subtopics to add a site, modify a site and build the taxonomy of the collaborative data mining system; Site Management (1004) which may include subtopics to update related tables, display a site, and miscellaneous functions; Applications (1006) which may include subtopics such as adding an application or modifying a guide application; Management (1008) which may include subtopics to assign classes, modify accounts, graduate guides and distribute and/or modify contracts; Systems Management (1012) which may include subtopics to access a manual META filemaker, access a manual META modifier, access a manual navigation filemaker, a keyword description creator, a global ASA creator and access automated processes; Reporting (1014) which may include subtopics to provide a site parameter snapshot and provide boards, chat and/or newsletter snapshots and; Additional (1010) which may include a subtopic to recycle accounts.

FIG. 13B may provide a diagram for adding a site to the collaborative data mining network taxonomy (1016). The template may include the name of the hub on the network (1018). The hub may provide a first horizontal division of the topical taxonomy. Within the hub a further delineation of the structure may be made with the section name (1020). The structure may in turn may be further delineated into the exact site name (1024). A data field may be provided to indicate the character site identification (1024) and the site navigation name (1026). The "live" data field may provide a "radio" button indication of the status of the site (1028). Site status indications may include whether the site is live and active, whether there is no interest in adding such a site to the taxonomy, whether the site topic is interesting, whether the site has been eliminated "X" from the taxonomy, and whether the site is active "A" and a guide is in training. A data field may be provided that indicates whether the site has a newsletter (1030), a chat room (1032), a bulletin board (1034) and/or classified adds (1036). A data field may indicate whether the site has disclaimers such as legal, medical, financial and/or official game. A final data field may be provide that indicates whether a copyright notice in on the site (1040). The add new site function may include a submit query bar (1042) to submit a query to the network. A reset function may be included to reset the form (1044).

FIG. 13C may provide a diagram of the functions that may be used to modify a site on the collaborative data mining system (1046). The functionality of the modify a site control may be similar to the add a site control described above. The modify site control may include a hub data field (1048), a section name data field (1050) and an exact site data field (1054). A data field for the site navigation name may be provided (1056). A data field with radio buttons to indicate the status of the site may be provided (1058) as described herein. Data fields may be provided to indicate and/or modify whether the site has a newsletter (1060), a chat room (1062), boards (1064) and/or classified advertising (1066). A data field may be provided to indicate and/or modify the sites disclaimers such as legal, medical, financial and/or official game (1068). A data field may be included to indicate and/or modify whether the cite has a copyright notice (1070). A data field may be provided to indicate whether the site maintains a Citibank.TM. profile (1072) and whether the site has business listings (1074). The custom profile feature such as the profiles in (1072) and/or (1074) may direct a collaborative data mining system to specifically "brand" the site or advertise on the site with the designated profile. The present invention may contain profiles for a plurality of customer profiles to create virtual collaborative data mining networks or subnetworks within a larger collaborative data mining network. The modify site menu may have a submit query function that may also modify the data entry (1076). And finally, the modify site menu may have a reset button to reset the menu (1078).

FIG. 13D may provide a guide management function to add an application (1080). This guide management function may be used to add a new guide application to the system described herein to begin the guide training process. For example, the guide management system may provide an "Air Travel" site application (1082). The site application may provide a username data field (1084) to receive a user name. A date field may be provided to receive or provide the current date (1086). Data fields for the title (1088), first name (1090), last name (1092), legal first name (1094), legal last name (1098) and e-mail address (1098) may be included to receive information about the applicant may be included. Data fields may be provided for a secondary e-mail address (1100), an addresses line one (1102), an address line two (1104), a city data line (1106), a state data line (1108), a zip code (1110), country (1112) and telephone information (1114). A data field may be provided to indicate what operating system the guide uses and/or intends to use (1116). A data field may be provided to track when the application was received (1120) and when and if the application was rejected (1122). The status of the application may be tracked with a pull down data field (1124). A data field may track whether the applicant was a referral and by whom the applicant was referred (1126). A referrer name data field may also be provided to track the name of a referrer (1128). A data text field may be provided to receive general comments on the guide application (1130). The add application menu may provide a submit function button (1132) to submit the application to the system and a reset button (1134) to reset the guide application screen.

FIG. 13E may provide a guide application modification menu to modify the data or status of a guide application. The guide modification menu tracks the add guide menu described above in FIG. 13D and its description is incorporated herein.

FIG. 13F may provide a means for assigning guides to classes for guide training (1136). A data menu may be provided (1138) to move accounts from an un-assigned status to an assigned status. A data field may be provided to list un-assigned accounts (1140). Function buttons may be provided to move accounts to an assigned status (1144) and assigned account back to an un-assigned status (1146). A data field may be provided to list assigned accounts (1142). A class name pull down menu may be provided to select a class name (1148). A due date data field may be provided to indicate a due date for the class (1150). A pull down menu may be provided to select a peer mentor (1152) and a pull down menu may be provided to select an assistant editor (1154) for the class. A function button may be provided to submit the form for data entry (1156). The class assignment menu may function by highlighting a list of accounts from the un-assigned data field (1140) and clicking the add function button (1144) to move the highlighted accounts to the assigned accounts data list. Once accounts are in the assigned data field, a user may assign a peer mentor and editor to the class. When the submit function button (1156) is pressed the ATMS may create a new class for the guide training program described herein.

FIG. 13G may provide a guide management function for tracking and modifying the guide accounts on a collaborative data mining system. An account tracking and modification menu may be provided (1158) to track and/or modify each guide in the network. The account tracking menu may provide identification information such as a tracking number (1160), a user name data field (1162), a current date data field (1164) a site name data field (1165), a title data field (1166) and guide information such as the first name (1168) , the last name (1170) , e-mail address (1172), a secondary e-mail address (1174) and a telephone number (1176). Data fields may be provided to include a class name (1178), class mentor (1180) and/or peer mentor (1181). Data fields may be provided to track and/or schedule the guide's progress through the development process such as a due date field (1182), a review date field (1184), an active date field (1186), a graduation date field (1188) and a termination date field (if any)(1190). A pull down menu may be provided to indicate the reason (if any) a guide was terminated (1192). A pull down menu may be provided to indicate the application status (1194) and whether there is any reason the application is on hold (1196). Data fields may be provided to track the dates for the system's review such as the final quality control date (1198) , the final edit date (1200), the final review date (1202), the hold date (1204) and the promotion date (if any) (1206). Data fields may also be provided to track when a contract was sent out to the guide (1208) and whether the contract was sent back (1210). Radio button type data fields may be provided to indicate whether a photograph has been received (1212), whether the navigational links for the site are in place (1214) and whether the art for the site is ready for deployment (1215). FIG. 13H may provide additional data fields for the guide management account tracking and/or modification system. Text data fields may be provided to receive mentor comments (1138) , editor comments (1140) and general comments (1142) about the guide and/or the account. In general, the functionality provided in this menu provides a means for tracking the guide's and a site's progress through the mentoring system. The menu provides a function button (1222) to submit and/or track the relevant data in the network. The system also provides a reset button (1224) to reset the data menu.

FIG. 13I may provide a means for guide management for affiliate graduations (1226). This tracking and status menu may provide identification information like that described above and incorporated herein, to identify the guide, the class and other appropriate information as described above. In this example, the guide has been placed on hold and the reasons therefore are indicated in the hold reasons data field (1232). The example also shows text information that more specifically describes the status in the mentor comments data text field (1334). This guide management function provides the submit function button (1228) and the reset function button (1230) to submit and /or reset the function respectively.

FIG. 13J provides a means for tracking and managing the details on the particular contract with each guide in the network. A guide management menu may be provided for managing and tracking guide contracts (1340). The contract management menu may provide a data field for the guide identification number assigned by the system (1342), a guide username (1344) and the current date (1346). Data fields may be provided for identifying the guide such as the guide's first name (1348), last name (1350), legal first name (1352), legal last name (1354), e-mail address (1356) and secondary e-mail address (1358). A data field may be provided to identify the exact site name (1360) as well as historical information such as when the application was received (1362). A data field may be provided to identify a contract number (1364) and whether the associate agreement has been sent out (1366) and whether the associate agreement has been received back (1372). Data fields may be provided to identify a license (1374) and the payout amount for the guide's services (1376). A data text field may be provided to note any contract addendum(s) (1378) and general comments (1380). The guide management function also provide data fields to note a termination date (1382) and a reason for the termination (1384). This guide management function may also provide a submit function key to submit and/or track data (1386). A reset button is also provided to reset the data form (1388).

The quality control process of the present invention may be used to automatically check the quality of web sites that are managed by the collaborative data mining system. Thus, the data mining system of the present invention has taken what was once a subjective human-resource intensive process and refined it to a checklist that may be completed in 20 minutes per site. This degree of refinement, as well as the technical enhancements for e-mail and tracking allows the present invention to ensure quality across the whole network by allowing the system to check all sites biweekly.

FIG. 5 provides a detailed diagram of a quality control process that may be used by system element (24). The quality control process may begin with a list of the live sites maintained by the network (302). The quality criteria of the sites may used to create a predetermined quality control checklist (304) and (306). The quality control process may perform spot checks (308) of the list of the web sites maintained by the system. The quality control process may use a software routine to automatically check a site for dead links that reference other web pages (310). The quality control process may use a software routine to check when the site was last updated to assure that the guide is actively participating (312) in the network. The quality control process also may check for feedback from other web users (314). From these quality control subroutines a list of action items may be generated for the guides (316). After a predetermined time, the sites may be re-evaluated to check for compliance (318) with the list of action items generated above (316). The quality control process may also maintain a confidential or public on-line forum for peer review (320). The total quality control scoring and tracking of the number of "hits" may be used to adjust the financial compensation for the guide (324).

FIG. 6 provides a detailed diagram of the frame system. The frame system assures that the proper frame set is displayed at the end user's web browser no matter how that user entered into the network of sites in the collaborative data mining system. More specifically, a page may arrive at a web browser (502). At that time, embedded java script code may be executed to query the "frames" object. If the frames object is greater than one then the java script may ask the object for the name of frame number one. If the name of frame number one designates a predetermined frame then the system knows the appropriate banner is already displayed (508) and the frame system does nothing more (506). If, however, the name of the frame is not the predetermined frame (510) then the system dynamically builds the frame set for the requested page (512). The frame system may then pass the frame set and appropriate data to the browser where the browser can process the frame set and cause the appropriate banner and page data display (514). The frame system may then exit (516).

FIG. 7 depicts a very simple frame based data format. The banner frame is shown (602) above the content frame (604). In the typical application, the banner frame provides the branded look and feel to the web site and the content frame (604) provides the topical content.

The present invention provides a means for creating, managing, maintaining and automating a collaborative data mining

The present invention provides a means for creating, managing, maintaining and automating a collaborative data mining system. This disclosure provides an exemplary disclosure of this system and other ways to implement and/or modify the execution of the present invention are within both the spirit and scope of this disclosure.