System and method for identifying records with valid address, but invalid name information

[0001] The present application claims the benefit under 35 U.S.C. § 119(e) of U.S. Provisional Application No. 60/332,823, filed Nov. 14, 2001. Said U.S. Provisional Application No. 60/332,823 is herein incorporated by reference in its entirety.

[0002] The present invention is generally related to mail processing systems commonly utilized for providing direct mail services and the like, and more particularly to a system and method for identifying mail records with valid addresses, but invalid name information.

[0003] Direct mail campaigns can be extremely effective for disseminating information to a large audience. The originators of such campaigns commonly employ computer-based information handling systems to maintain databases containing mailing information for individuals within an audience targeted to receive information. However these systems currently have no way of determining whether the party being addressed exists at the location or address obtained for that individual. If the targeted party does not reside at the given location, information sent to that location likely will not reach the party, wasting the resources of the originator of the campaign. Consequently, with direct mail expenses constantly increasing due to increased costs for postage, labor, materials and the like, and with mail security becoming a prevalent issue, it is desirable to verify the accuracy of mailing information obtained for the audience targeted to receive the direct mail campaign.

[0004] Known to the art are software applications which compare a database containing mailing addresses with a database containing known addresses to determine if an address is contained within a set of known addresses. These applications prevent processing of records not in the known set saving the material, labor and transportation costs. However, such applications cannot determine if the party (name) targeted to receive information is at the verified address.

[0005] Accordingly, it would be advantageous to provide a method for determining whether a party (name) is contained within the subset of names associated with an address contained within the set of known addresses. In this manner, it would be possible to prevent mailing of information to locations where an targeted party does not reside, saving further material, labor and transportation costs.

[0006] The present invention is directed to a novel method for automatically determining if a record address belongs to a set of known addresses and the associated record name does not belong to the subset of known names associated with the known address. In exemplary embodiments, the probability of match between the addresses is calculated and the probability of mismatch of names associated with the address is then computed. Heuristics are employed for computing the probability of match and mismatch.

[0007] The present invention may be utilized in mail processing systems as a complimentary technology in computer controlled printers such as an laser printer addressing letters to be mailed, where it is desirable to know if the intended recipient is at the address on the record to be printed. Similarly, the present invention may be utilized in conjunction with an ETL (Extract, Transform, and Load) database environment where it is desirable to know that the name and address information stored in the database is correct.

[0008] It is understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention claimed. The accompanying drawings, which are incorporated in and constitute part of the specification, illustrate an embodiment of the invention and together with the general description, serve to explain the principles of the invention.

[0013] Reference will now be made in detail to the presently preferred embodiments of the invention, examples of which are illustrated in the accompanying drawings.

[0014] Referring now to FIG. 1, an exemplary method for automatically identifying records with a valid address, but invalid name information in a information handling system is described. The method 100 may, for example, be implemented in a mail processing system or like information handling system (see FIGS. 2 and 3) when an event or condition within an application executed by the system requires input of a record with name and address for processing. When such a record is input at step 110, the address from the input record is compared against a set of records to determine if the address matches an address in the known set at step 130. Preferably, to determine if a match is made, heuristics may be determined for various operating systems, applications, application environments, and the like. These heuristics preferably lead to the correct determination of the address being a match to any record in the known set of addresses, by implementing a non-rigorous or self-learning approach. For example, in a preferred embodiment, the determination of a match may be based on the premise that a match is made if weights are assigned to specific address components (house number, street name and the like), and the sum of the weights for address components that match exceed a threshold value. For example the weighted address value (WAV) may be calculated as follows:
WAV=(W1·V1)+(W2·V2)+ . . . (Wn·Vn)
where W(1, . . . n) is the weight assigned to the various address components and V(1, . . . n) is 1 if the components match, 0 if the component is missing and −1 if the components do not match. If the weighted address value (WAV) is greater than the threshold address value (TAV) then a match is made. TAV may be determined by experimentation or may be set as a parameter by an operator using the invention. If the address is not in the set of known addresses it is marked as unknown address at step 160 and returned to the application step 110. If the address belongs to the set of known addresses, the name is then compared to the subset of names known to be associated with the known address at step 140. Preferably, to determine if a match is made, heuristics may be determined for various operating systems, applications, application environments, and the like. These heuristics preferably lead to the correct determination of the name being a match to any record in the known subset of names, by implementing a non-rigorous or self-learning approach. For example, in a preferred embodiment, the determination of a match may be based on the premise that a match is made if weights are assigned to specific name components (title, first name and the like), and the sum of the weights for name components that match exceed a threshold value. For example the weighted name value (WNV) may be calculated as follows:
WAV=(W1·V1)+(W2·V2)+ . . . (Wn·Vn)
where W(1, . . . n) is the weight assigned to the various name components and V(1, . . . n) is 1 if the components match and 0 if the components do not match. In comparison of components the first letter and single consonants may be used, dropping vowels. If WNV is greater than the threshold name value (TNV) then a match is made. TNV may be determined by experimentation or may be set as a parameter by an operator using the invention. If the name does not belong to the subset of names associated with the known address the record is marked as an unknown name at a known (valid) address (phantom) at step 170 and returned to the application step 110. If the name belongs to the subset of known names associated with the known address the record is marked as a known name and address at step 150 and returned to the application step 110.

[0015] An example illustrating the application of the method shown in FIG. 1 for identifying records with valid addresses, but invalid name information is now provided for illustration. The known set of addresses and names is as follows:

[0016]

		SUBSET OF
	ADDRESS	ASSOCIATED NAMES
	123 N. Main St Apt 1	John Smith
	Smalltown, NE 68106	Nancy Jones
	41 Corby Ave	Harry Guy
	Metro, NY 10112	Nada Guy
	702 State Rd 4	Otto Doors
	Pipeline, AK 99702	Rusty Doors

The address components that will be matched and their weights are:

[0017]

	ADDRESS COMPONENT	WEIGHT
	House number	2
	Pre-directional	1
	Street name	4
	Suffix	1
	Post-directional	1
	Secondary type	2
	Secondary number	2
	City name	4
	State	4
	Zip code	1

The name components and their weights are:

[0018]

	NAME COMPONENT	WEIGHT
	First name	2
	First initial	1
	Last name	3

[0019] By way of example only, if the threshold address value (TAV) and the threshold name value (TNV) are determined to be 13 and 4 respectfully, based on experimentation or data entered by an operator of a system employing the present method, then when the following name and address is presented the method would automatically notify the application system that the address is not in the set of known addresses:

• • [0020] Mary Lamb 402 Corby Ave, Metro, N.Y. 10111
    When matched against known address the matches would be as follows:

[0021]

KNOWN ADDRESS 1
Address		Input	Known	Match	Weight
Component	Weight	Record	Record 1	Value	Value
House number	2	402	123	−1 (No Match)	−2
Pre-directional	1		N	0 (Missing)	0
Street name	4	Corby	Main	−1 (No Match)	−4
Suffix	1	Ave	St	−1 (No Match)	−1
Post-directional	1			0 (Missing)	0
Secondary type	2		Apt	0 (Missing)	0
Secondary	2		1	0 (Missing)	0
number
City name	4	Metro	Smalltown	−1 (No Match)	−4
State	4	NY	NE	−1 (No Match)	−4
Zip code	1	10111	68106	−1 (No Match)	−1
				WAV =	−16
KNOWN ADDRESS 2
Address		Input	Known	Match	Weight
Component	Weight	Record	Record 2	Value	Value
House number	2	402	41	−1 (No Match)	−2
Pre-directional	1			0 (Missing)	0
Street name	4	Corby	Corby	1 (Match)	4
Suffix	1	Ave	Ave	1 (Match)	1
Post-directional	1			0 (Missing)	0
Secondary type	2			0 (Missing)	0
Secondary	2			0 (Missing)	0
number
City name	4	Metro	Metro	1 (Match)	4
State	4	NY	NY	1 (Match)	4
Zip code	1	10111	10112	−1 (No Match)	−1
				WAV =	10
KNOWN ADDRESS 3
Address		Input	Known	Match	Weight
Component	Weight	Record	Record 3	Value	Value
House number	2	402	702	−1 (No Match)	−2
Pre-directional	1			0 (Missing)	0
Street name	4	Corby	State Rd 4	−1 (No Match)	−4
Suffix	1	Ave		0 (Missing)	0
Post-directional	1			0 (Missing)	0
Secondary type	2			0 (Missing)	0
Secondary	2			0 (Missing)	0
number
City name	4	Metro	Pipeline	−1 (No Match)	−4
State	4	NY	AK	−1 (No Match)	−4
Zip code	1	10111	99702	−1 (No Match)	−1
				WAV =	−15

The input record does not have a WAV value equal to or greater than the threshold address value (TAV) of 13, thus the address does not match a known address.

[0022] If the following name and address is presented the method would automatically notify the application system that the address is in the set of known addresses, but the name is not in the subset of associated names:

• • [0023] Mary Lamb 41 Corby Ave, Metro, N.Y. 10111
    When matched against known address the matches would be as follows:

[0024]

KNOWN ADDRESS 1
Address		Input	Known	Match	Weight
Component	Weight	Record	Record 1	Value	Value
House number	2	41	123	−1 (No Match)	−2
Pre-directional	1		N	0 (Missing)	0
Street name	4	Corby	Main	−1 (No Match)	−4
Suffix	1	Ave	St	−1 (No Match)	−1
Post-directional	1			0 (Missing)	0
Secondary type	2		Apt	0 (Missing)	0
Secondary	2		1	0 (Missing)	0
number
City name	4	Metro	Smalltown	−1 (No Match)	−4
State	4	NY	NE	−1 (No Match)	−4
Zip code	1	10111	68106	−1 (No Match)	−1
				WAV =	−16
KNOWN ADDRESS 2
Address		Input	Known	Match	Weight
Component	Weight	Record	Record 2	Value	Value
House number	2	41	41	1 (Match)	2
Pre-directional	1			0 (Missing)	0
Street name	4	Corby	Corby	1 (Match)	4
Suffix	1	Ave	Ave	1 (Match)	1
Post-directional	1			0 (Missing)	0
Secondary type	2			0 (Missing)	0
Secondary	2			0 (Missing)	0
number
City name	4	Metro	Metro	1 (Match)	4
State	4	NY	NY	1 (Match)	4
Zip code	1	10111	10112	−1 (No Match)	−1
				WAV =	13
KNOWN ADDRESS 2 ASSOCIATED NAME 1
Name		Input	Associated	Match	Weight
Component	Weight	Record	Name 1	Value	Value
First name	2	Mary	Harry	−1 (No Match)	−2
First initial	1	M	H	−1 (No Match)	−1
Last name	3	Lamb	Guy	−1 (No Match)	−3
				WNV =	−6
KNOWN ADDRESS 2 ASSOCIATED NAME 2
Name		Input	Associated	Match	Weight
Component	Weight	Record	Name 2	Value	Value
First name	2	Mary	Nada	−1 (No Match)	−2
First initial	1	M	N	−1 (No Match)	−1
Last name	3	Lamb	Guy	−1 (No Match)	−3
				WNV =	−6
KNOWN ADDRESS 3
Address		Input	Known	Match	Weight
Component	Weight	Record	Record 3	Value	Value
House number	2	41	702	−1 (No Match)	−2
Pre-directional	1			0 (Missing)	0
Street name	4	Corby	State Rd 4	−1 (No Match)	−4
Suffix	1	Ave		0 (Missing)	0
Post-directional	1			0 (Missing)	0
Secondary type	2			0 (Missing)	0
Secondary	2			0 (Missing)	0
number
City name	4	Metro	Pipeline	−1 (No Match)	−4
State	4	NY	AK	−1 (No Match)	−4
Zip code	1	10111	99702	−1 (No Match)	−1
				WAV =	−15

Here the input address has a WAV equal or greater than the TAV for known address 2, however the WNV is less than the TNV for both names associated with known address 2.

[0025] If the following name and address is presented the method would automatically notify the application system that the address is in the set of known addresses, and the name is in the subset of associated names:

• • [0026] N Guy 41 Corby Ave, Metro, N.Y. 10111
    When matched against known address the matches would be as follows:

[0027]

KNOWN ADDRESS 1
Address		Input	Known	Match	Weight
Component	Weight	Record	Record 1	Value	Value
House number	2	41	123	−1 (No Match)	−2
Pre-directional	1		N	0 (Missing)	0
Street name	4	Corby	Main	−1 (No Match)	−4
Suffix	1	Ave	St	−1 (No Match)	−1
Post-directional	1			0 (Missing)	0
Secondary type	2		Apt	0 (Missing)	0
Secondary	2		1	0 (Missing)	0
number
City name	4	Metro	Smalltown	−1 (No Match)	−4
State	4	Ny	Ne	−1 (No Match)	−4
Zip code	1	10111	68106	−1 (No Match)	−1
				WAV =	−16
KNOWN ADDRESS 2
Address		Input	Known	Match	Weight
Component	Weight	Record	Record 2	Value	Value
House number	2	41	41	1 (Match)	2
Pre-directional	1			0 (Missing)	0
Street name	4	Corby	Corby	1 (Match)	4
Suffix	1	Ave	Ave	1 (Match)	1
Post-directional	1			0 (Missing)	0
Secondary type	2			0 (Missing)	0
Secondary	2			0 (Missing)	0
number
City name	4	Metro	Metro	1 (Match)	4
State	4	NY	NY	1 (Match)	4
Zip code	1	10111	10112	−1 (No Match)	−1
				WAV =	13
KNOWN ADDRESS 2 ASSOCIATED NAME 1
Name		Input	Associated	Match	Weight
Component	Weight	Record	Name 1	Value	Value
First name	2		Harry	0 (Missing)	0
First initial	1	N	H	−1 (No Match)	−1
Last name	3	Guy	Guy	1 (Match)	3
				WNV =	2
KNOWN ADDRESS 2 ASSOCIATED NAME 2
Name		Input	Associated	Match	Weight
Component	Weight	Record	Name 2	Value	Value
First name	2		Nada	0 (Missing)	0
First initial	1	N	N	1 (Match)	1
Last name	3	Guy	Guy	1 (Match)	3
				WNV =	4
KNOWN ADDRESS 3
Address		Input	Known	Match	Weight
Component	Weight	Record	Record 3	Value	Value
House number	2	41	702	−1 (No Match)	−2
Pre-directional	1			0 (Missing)	0
Street name	4	Corby	State Rd 4	−1 (No Match)	−4
Suffix	1	Ave		0 (Missing)	0
Post-directional	1			0 (Missing)	0
Secondary type	2			0 (Missing)	0
Secondary	2			0 (Missing)	0
number
City name	4	Metro	Pipeline	−1 (No Match)	−4
State	4	NY	AK	−1 (No Match)	−4
Zip code	1	10111	99702	−1 (No Match)	−1
				WAV =	−15

Here the input address has a WAV equal or greater than the TAV for known address 2 and the WNV is equal or greater than the TNV for the second name associated with known address 2.

[0028] Referring now to FIG. 2, a block diagram depicting an exemplary computer controlled printer (Inkjet, Laser, or the like) mail processing system in accordance with the present invention is shown. The mail processing system 200 may include a computer (minicomputer, mainframe computer, personal computer, or the like) 220 interconnected with an input device (tape drive) 210, a non-volatile storage device (disk drive) 230 and one or more printing devices 240, 250 and 260. The mail processing system 200 may have a client/server architecture which makes use of distributed intelligence to treat the computer (or server) 220 and all devices (or clients) 210,230, 240, 250 and 260 as intelligent programmable devices, thus exploiting the full information processing capacity of each device. This may be accomplished by dividing the processing of data between the server 220 and the attached devices 210, 230, 240, 250 and 260. For example the server 220 may run administrative software that controls requests for data from and to all devices 210, 230, 240, 250 and 260 to the mail processing system 200, while each device 210, 230, 240, 250 and 260 may run software that is optimized to the device function.

[0029] As shown in FIG. 2, the mail processing system 200 may operate in a mainframe computer environment where the mail processing system 200 would determine if an input record is to be printed, how the input record is to be printed (e.g., on a letter, an envelope, a mailer, or the like) and where the input record is to be printed (e.g., on which printing device 240, 250 or 260). In this manner the mail processing system 200 minimizes the material, labor and transportation costs.

[0030] In exemplary embodiments, computer (server) 220 or alternately a client device 210, 230, 240, 250 or 260 may execute an application 270 implementing method 100 shown in FIG. 1. In this manner, system 200 may identify records with valid address information, but invalid name information. As described in the discussion of FIG. 1, this may be accomplished by determining if the address of the record to be printed is in a set of known addresses and, if the address of the record is determined to be in the set of known addresses, determining if the name of the record to be printed is in a subset of known names associated with the address by calculating a weighted name value for the name and comparing the weighted name value with a predetermined threshold name value. If the address is determined to be in the set of known addresses and the name is determined to not be in the subset of known names associated with the address, the record may be marked as having valid address information, but invalid name information to prevent printing of the record.

[0031] Turning now to FIG. 3, a block diagram depicting an information handling system implementing a computer controlled database ETL (Extract, Transform and Load) environment. The information handling system 300 may include a computer (minicomputer, mainframe computer or personal computer) 320 interconnected with an input device (tape drive) 310, a non-volatile storage device (disk drive) 330, a database server 340, which in turn may be connected to one or more disk drives 350, 360 and 370. The information handling system 300 may have a client/server architecture which makes use of distributed intelligence to treat the computer (server) 320 and all devices (or clients) 310, 330, 340, 350, 360 and 370 as intelligent programmable devices, thus exploiting the full information processing capacity of each device. This may be accomplished by dividing the processing of data between the server 320 and the attached devices 310, 330, 340, 350, 360 and 370. For example, the server 320 may run administrative software that controls requests for data from and to devices 310, 330, and 340 to the information handling system 300, while the database server 340 controls requests for data from and to devices 350, 360 and 370 to the server 320, while each device 310, 330, 340, 350, 360 and 370 may run software that is optimized to the device function.

[0032] As shown in FIG. 3, the information handling system 300 may operate in a mainframe computer environment where the information handling system 300 would determine if an input record is to be stored, how it is to be stored and where it is to be stored. In this manner the information handling system 300 minimizes the storage and processing costs.

[0033] In exemplary embodiments, computer (server) 320 or alternately, database server 340 or a devices 310, 330, 350, 360 or 380 may execute an application 390 implementing method 100 shown in FIG. 1. In this manner, system 300 may identify records with valid address information, but invalid name information. As described in the discussion of FIG. 1, this may be accomplished by determining if the address of the record to be printed is in a set of known addresses and, if the address of the record is determined to be in the set of known addresses, determining if the name of the record to be printed is in a subset of known names associated with the address by calculating a weighted name value for the name and comparing the weighted name value with a predetermined threshold name value. If the address is determined to be in the set of known addresses and the name is determined to not be in the subset of known names associated with the address, the record may be marked as having valid address information, but invalid name information. The marked record may then be deleted from the database of records being analyzed if so desired.

[0034] Although the invention has been described with a certain degree of particularity, it should be recognized that elements thereof, including different heuristics, may be altered by persons skilled in the art without departing from the spirit and scope of the invention. One of the embodiments of the invention can be implemented as sets of instructions resident in the main memory of one or more computer systems such as computers 220 and 320 shown in FIGS. 2 and 3, respectively. Until required by the computer system, the set of instructions may be stored in another computer readable memory such as the computer system's auxiliary memory, for example in a hard disk drive or in a removable memory such as an optical disk for utilization in a CD-ROM drive, a floppy disk for utilization in a floppy disk drive, or a personal computer memory card for utilization in a personal computer card slot. Further, the set of instructions can be stored in the memory of another computer and transmitted over a local area network or a wide area network, such as the Internet, when desired by the user. Additionally, the instructions may be transmitted over a network in the form of an applet that is interpreted after transmission to the computer system rather than prior to transmission. One skilled in the art would appreciate that the physical storage of the sets of instructions or applets physically changes the medium upon which it is stored electrically, magnetically, chemically, physically, optically or holographically so that the medium carries computer readable information.

[0035] It is believed that the present invention and many of its attendant advantages will be understood by the foregoing description, and it will be apparent that various changes may be made in the form, construction and arrangement of the components thereof without departing from the scope and spirit of the invention or without sacrificing all of its material advantages. The form herein before described being merely an explanatory embodiment thereof, it is the intention of the following claims to encompass and include such changes.