Reputation Evalution Using a contact Information Database

• US 20090265198A1
• Filed: 04/22/2008
• Published: 10/22/2009
• Est. Priority Date: 04/22/2008
• Status: Abandoned Application

• Current Assignee: Plaxo Incorporated (Comcast Corporation)
• Original Assignee: Plaxo Incorporated (Comcast Corporation)
• Inventors: Golub, Ben, Smarr, Joseph, Lester, Peter
• Application Number:
  US12/107,695
• Publication Number:
  US 20090265198A1
• Time in Patent Office: Days
• US Class Current: 705/7
• CPC Class Codes: G06Q 30/0185 Product, service or business identity fraud
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0282 Rating or review of business operators or products
• CPC Class Codes: G06Q 30/0609 Buyer or seller confidence or verification
• CPC Class Codes: H04L 12/1813 for computer conferences, e.g. chat rooms instant messaging H04L51/04; protocols for multimedia communication H04L65/1101; arrangements for multi-party communication H04L65/403; telephonic conference arrangements H04M3/56; television conference systems H04N7/15
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46

Abstract

A contact information database, including records such as those stored in a personal address book, is applied to evaluate the reputation of a user and/or conduct fraud or spam detection. A number of different factors selected for reputation/fraud prediction value can be used in a statistical model to evaluate reputation of an individual based on an identifier, such as an email address. The factors can include information useful in predicting the reputation of an individual, such as in how many address books the email address or other information appears, whether emails have been previously sent to that email address, whether any such emails have been returned as undeliverable, and so forth. These factors can be used to create a vector including scores for the user on the various factors, which can be stored in a vector database and updated regularly as the information changes. The information in the vector database can be accessed by parties for use in reputation evaluation, fraud detection, etc. for a particular email address or individual.

Claims

1. A computer-implemented method for reputation evaluation based on a contact information database, the method comprising:

selecting a plurality of factors for evaluating the reputation of each of a plurality of users, each of the factors based on one or more records about users stored in a contact information database or the absence of such records;

building a reputation model for determining which of the plurality of factors are predictive in evaluating the reputation of each of the plurality of users;

generating a vector for each of the plurality of users based on the results of the reputation model;

storing the vector for each of the plurality of users;

receiving a substantially unique identifier for identifying a first user conducting a transaction;

retrieving, based on the substantially unique identifier, the stored vector for the first user, wherein at least one of the factors included in the vector is independent of the transaction being conducted by the first user; and

providing the vector for the first user to a reputation engine for evaluating the reputation of the first user.

2. The method of claim 1, wherein retrieving, based on the substantially unique identifier, a stored vector for the first user further comprises:

selecting additional factors for evaluating the reputation of the first user; and

updating the vector to include the additional factors added to the record for the first user.

3. The method of claim 1, wherein the contact information database includes at least a million records that can be used in generating vectors.

4. The method of claim 1, wherein receiving a substantially unique identifier further comprises receiving a reputation evaluation request including the substantially unique identifier from an independent entity requesting evaluation of the reputation of the first user, the request including information regarding specific factors of interest to the independent entity, the specific factors being used to customize the vector for the first user.

5. The method of claim 1, wherein providing the vector for the first user to the reputation engine further comprises sending the vector or a portion of the vector to an independent entity having the reputation engine, the reputation engine applying the vector or portion of the vector received to evaluate the reputation of the first user.

6. The method of claim 1, wherein retrieving further comprises:

determining that no vector has been generated for the first user;

selecting the plurality of factors for evaluating the reputation of the first user; and

generating the vector for the first user based on the reputation model.

7. The method of claim 1, wherein retrieving further comprises:

determining that no vector has been generated for the first user;

selecting the plurality of factors for evaluating the reputation of the first user; and

using the absence of a vector or absence of the substantially unique identifier in the contact information database as a predictive factor regarding the user'"'"'s reputation.

8. The method of claim 1, wherein the vector generated in the method is a real-time generated vector that is generated in real-time upon the receiving of the substantially unique identifier, the real-time being then stored and provided to the reputation engine.

9. The method of claim 1, wherein one or more of the plurality of factors used in generating the vector for the first user is based on records stored in the contact information database for other users and wherein the transaction is the first transaction conducted by the first user of that transaction type, the vector for the first user not including any data on prior history of the first user conducting prior transactions.

10. The method of claim 1, wherein the plurality of factors for the first user are selected from a group consisting of:

whether the substantially unique identifier for the first user appears in records for other users, the number of records for other users in which the substantially unique identifier appears, the type of contact information database in which the substantially unique identifier appears, the composition of a social network of which the first user is a member, whether emails have been sent to the first user, whether emails sent to the first user have been received by the first user, and whether the first user has replied to any emails sent to the first user.

11. The method of claim 1, wherein building a reputation model further comprises:

applying binary logistic regression to build a statistical model in which the plurality of factors represent independent variables and wherein the model includes a dependent variable of fraud or not fraud; and

assigning each of the plurality of factors a score representing the relevance of the factor to the reputation of each of the users.

12. The method of claim 1, wherein building a reputation model further comprises calculating a reputation score for each user in the contact information database or social graph based on the records in the contact information database or the social graph that reference the user, where those references are themselves weighted by the reputation of a user from which these records came, and wherein this weighting continues recursively throughout the social graph

13. The method of claim 1, further comprising providing an interface through which an independent entity directly accesses the stored vectors.

14. A computer system for reputation evaluation based on a contact information database, the system comprising a computer configured to execute a method computer modules comprising:

a factor manager for selecting a plurality of factors for evaluating the reputation of each of a plurality of users, each of the factors based on one or more records about users stored in a contact information database or the absence of such records;

a model generator for building a reputation model for determining which of the plurality of factors are predictive in evaluating the reputation of each of the plurality of users;

a vector generator for generating a vector for each of the plurality of users based on the results of the reputation model;

a vector database for storing the vector for each of the plurality of users;

a transmission module for receiving a substantially unique identifier for identifying a first user conducting a transaction;

a retrieval module for retrieving, based on the substantially unique identifier, the stored vector for the first user, wherein at least one of the factors included in the vector is independent of the transaction being conducted by the first user; and

the transmission module for providing the vector for the first user to a reputation engine for evaluating the reputation of the first user.

15. The system of claim 14, wherein the reputation engine is a fraud engine for evaluating the likelihood that the first user is attempting a fraudulent transaction.

16. The system of claim 14, further comprising an update engine updating the vector to reflect additional information added to the record.

17. The system of claim 14, wherein the retrieval module is further configured for communicating with:

the factor manager for selecting additional factors for evaluating the reputation of the first user; and

the vector generator updating the vector to include the additional factors added to the record for the first user.

18. The system of claim 14, further comprising a historical data manager for storing information about message sending/receiving history of the user that is used in generating the vectors.

19. The system of claim 14, wherein the vector generated by the vector generator is a real-time generated vector that is generated in real time upon the receiving of the substantially unique identifier by the transmission module, the vector database being further configured to store the real-time generated vector and the transmission module being further configured to provide the real-time generated vector to the reputation engine.

20. The system of claim 14, wherein the transmission module is further configured for receiving a reputation evaluation request including the substantially unique identifier from an independent entity requesting evaluation of the reputation of the first user and wherein the reputation evaluation request for the first user includes information regarding specific factors of interest to the independent entity, the specific factors being used to customize the vector for the first user.

21. The system of claim 14, wherein one or more of the plurality of factors used in generating the vector for the first user is based on records stored in the contact information database for other users, and wherein the transaction is the first transaction conducted by the first user of that transaction type, the vector for the first user not including any data on prior history of the first user conducting prior transactions.

22. The system of claim 14, wherein the plurality of factors for the first user are selected from a group consisting of:

whether the record for the first user includes contact information besides an email address, whether the record for the first user is a new record, the composition of a social network of which the first user is a member, whether emails have been sent/received by the first user recently.

23. The system of claim 14, further comprising an interface through which an independent entity accesses a vector database in which the vectors are stored to obtain information about the first user.

24. A computer program product for reputation evaluation based on a contact information database, the computer program product comprising a computer-readable medium storing computer program code for performing the method comprising:

selecting a plurality of factors for evaluating the reputation of each of a plurality of users, each of the factors based on one or more records about users stored in a contact information database or the absence of such records;

building a reputation model for determining which of the plurality of factors are predictive in evaluating the reputation of each of the plurality of users;

generating a vector for each of the plurality of users based on the results of the reputation model;

storing the vector for each of the plurality of users;

receiving a substantially unique identifier for identifying a first user conducting a transaction;

retrieving, based on the substantially unique identifier, the stored vector for the first user, wherein at least one of the factors included in the vector is independent of the transaction being conducted by the first user; and

providing the vector for the first user to a reputation engine for evaluating the reputation of the first user.

25. The computer program product of claim 24, wherein the reputation engine is a spam engine for evaluating the likelihood that the first user is attempting to send spam messages.

26. The computer program product of claim 24, wherein retrieving, based on the substantially unique identifier, a stored vector for the first user further comprises:

selecting additional factors for evaluating the reputation of the first user; and

updating the vector to include the additional factors added to the record for the first user.

27. The computer program product of claim 24, wherein the contact information database includes at least a million records that can be used in generating vectors, the records being updated regularly as changes are made by users.

28. The computer program product of claim 24, wherein receiving a substantially unique identifier further comprises:

receiving a reputation evaluation request including the substantially unique identifier from an independent entity requesting evaluation of the reputation of the first user.

modifying the vector in near real time to customize the vector to the needs of the independent entity.

29. The computer program product of claim 24, wherein the plurality of factors for the first user are selected from a group consisting of:

whether a shipping address of a product requested by the first user is geographically near the first user'"'"'s physical address, whether and how the first users is linked in a social network to other users, and the message sending and receiving history of the first user.

30. The computer program product of claim 24, wherein building a reputation model further comprises calculating a reputation score for each user in the contact information database or social graph based on the records in the contact information database or the social graph that reference the user, where those references are themselves weighted by the reputation of a user from which these records came, and wherein this weighting continues recursively throughout the social graph.