Method and system for creating and maintaining a database of user profiles and a related value rating database for information sources and for generating a list of information sources having a high estimated value

• US 7,263,529 B2
• Filed: 08/29/2003
• Issued: 08/28/2007
• Est. Priority Date: 08/29/2003
• Status: Expired due to Fees

• Current Assignee: Pitney Bowes Incorporated (Böwe Systec AG)
• Original Assignee: Pitney Bowes Incorporated (Böwe Systec AG)
• Inventors: Heiden, Richard W., Cordery, Robert A., Lang, Christopher C.
• Primary Examiner(s): WOO, ISAAC M
• Application Number:
  US10/651,911
• Publication Number:
  US 20050050094A1
• Time in Patent Office: 1,460 Days
• Field of Search: 707100-1041, 707200-205
• US Class Current: 1/1
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: Y10S 707/99942 Manipulating data structure, e.g. compression, compaction, compilation
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface
• CPC Class Codes: Y10S 707/99944 Object-oriented database structure
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing

Abstract

A method and system for creating and maintaining databases of user profiles and value ratings for information sources, and for generating lists of information sources having high estimated values. User controlled systems for accessing information sources through the internet or other networks are monitored by a server. The server updates the user profiles and computes value ratings associated with the user profiles for various information sources in accordance with monitored information. Estimated values for information sources to a particular user are obtained as a function of the value ratings associated with profiles of other users and the degree of similarity between the particular user'"'"'s current profile and other user profiles. Previous profiles are not deleted when updated but are retained and treated substantially as other profiles in computing value estimates. Information sources can be other users.

Claims

1. A method utilizing a computer for generating a list of information sources having an estimated value to a user wherein at least one information source is present comprising the steps of:

a) receiving fixed information for each of a plurality of users;

b) initializing profiles for each of said users in a user profile database with said fixed information;

c) receiving monitored information relating to activities of a user;

d) determining if said monitored information relates to a particular information source and, if so, determine a value rating associated with said user for said particular information source in accordance with said monitored information;

e) creating an updated profile for said user in said profile database in accordance with said monitored information;

f) if a value rating is determined, associating said value rating with said updated user profile and recording said value rating in said related value rating database; and

g) determining if a list is to be generated; and

if so

g1) calculating an estimated value to said user of an information source as a function of said user'"'"'s profile, profiles for others of said users, and value ratings for said information source associated with said other users profiles;

g2) repeating step g1 until estimated values have been calculated for all of said information sources;

g3) selecting and outputting to said user, or saving for later output, information sources having estimated values so that said user receiving the list of information sources will have a quantitative estimate of the likely value of each of the sources.

2. A method as described in claim 1 where said estimated value is calculated as:

E_i,n=Σ

_x(S_x,n·

V_i,x)/X

where;

a) E_i,n is the estimated value to user n of information source i in accordance with the values and interests of user n as indicated by user n'"'"'s current updated profile P_n;

b) coefficients S_x,n are a measure of similarity between a profile P_x of another user x and said profile P_n;

c) V_i,x is a value rating for said information source i associated with said profile P_x;

d) Σ

_x indicates summation over all profiles P_x for which V_i,x is defined; and

e) X is the total number of profiles P_x.

3. A method as described in claim 2 where, as data is accumulated in said databases, statistical methods are used to adjust said coefficients S_x,n.

4. A method as described in claim 1 where said estimated value is calculated as:

E_i,n=Σ

_kC_i,k·

Z_k,n

where;

a) E_i,n is an estimated value to user n of information source i in accordance with the values and interests of said user n as indicated by said user n'"'"'s profile P_n;

b) C_i,k are coefficients generated using linear regression methods and relating said information source i to kth variables in profiles;

c) Z_k,n is the value of said kth variable in said profile P_n; and

d) Σ

_k indicates summation over all values of k.

5. A method as described in claim 4 where, as data is accumulated in said databases, linear regression methods are used to adjust said coefficients C_i,k.

6. A method as described in claim 1 where said estimated value is calculated by:

a) generating a neural network associated with information source i;

b) training said network using values Z_k,x and V_i,x, from profiles in a set {P_x} of profiles for other users x, and for which values V_i,x are defined;

c) determining E_i,n by applying Z_k,n to said network;

where

d) E_i,n is an estimated value to user n of information source i in accordance with the values and interests of said user n as indicated by said user n'"'"'s profile P_n;

e) Z_k,n is the value of said kth variable in said profile P_n; and

c) V_i,x is a value rating for said information source i associated with a profile in said set {P_x}.

7. A method as described in claim 6 where, as data is accumulated in said databases, said network is further trained.

8. A method as described in claim 1 where said calculating step is carried out using said user'"'"'s current, updated profile P_n⁺.

9. A method as described in claim 1 where said information source is in contact with another user.

10. A method as described in claim 9 where said estimated value is calculated as:

EFC_n,x=1/Σ

_k(W_n,k(Z_k,x−

Z_k,n)²)

where;

a) EFC_n,x is an estimate of the fruitfulness of contact with other user x by user n;

b) W_n,k are coefficients provided by said user n;

c) Z_k,n is the value of a kth variable in said user n'"'"'s profile P_n;

d) Z_k,x is the value of a kth variable in said other user x'"'"'s profile P_x; and

e) Σ

_k indicates summation over all values of k.

11. A method as described in claim 10 where said other user x can select either a private or a public status and will only be included in said list if he or she selects said public status.

12. A system comprising:

a) a server;

b) a profile database;

c) a value rating database;

d) said server communicating with a profile database and a value rating database;

e) said server also communicating with a plurality of monitors for monitoring a corresponding plurality of users;

f) said server being programmed to;

f1) receive fixed information for each of a plurality of users;

f2) initialize profiles for each of said users in said profile database with said fixed information;

f3) monitor a user and receive monitored information relating to said user;

f4) determine if said monitored information relates to a particular information source wherein at least one particular information sources is present and, if so, determine a value rating for said particular information source in accordance with said monitored information;

f5) create an updated profile for said user in said profile database in accordance with said monitored information;

f6) if a value rating is determined, associate said value rating with said updated user profile and record a value rating in said related value rating database; and

f7) determine if a list is to be generated; and

if so

f7A) calculate an estimated value to said user of an information source as a function of said user'"'"'s profile, profiles for others of said users, and value ratings for said information source associated with said other users profiles;

f7B) repeat step f7A until estimated values have been calculated for all of said information sources; and

f7C) select and output to said user, or save for later output, information sources having estimated values so that said user receiving the information sources will have a quantitative estimate of the likely value of each of the sources.

13. A method utilizing a computer readable storage medium having embedded instructions for providing instructions to a server, said instructions controlling said server to perform the steps of:

a) receiving fixed information for each of a plurality of users;

b) initializing profiles for each of a users in said user profile database with said fixed information;

c) monitoring a user and receive monitored information relating to said user;

d) determining if said monitored information relates to a particular information source wherein at least one particular information sources is present and, if so, determine a value rating for said particular information source in accordance with said monitored information;

e) creating an updated profile for a user in said profile database in accordance with said monitored information;

f) if a value rating is determined, associating said value rating with said updated user profile and record a value rating in said related value rating database; and

g) determining if a list is to be generated; and

if so

g1) calculating an estimated value to said user of an information source as a function of said user'"'"'s profile, profiles for others of said users, and value ratings for said information source associated with said other users profiles;

g2) repeating step gi until estimated values have been calculated for all of said information sources; and

g3) selecting and output to said user, or save for later output, information sources having estimated values so that said user receiving the information sources will have a quantitative estimate of the likely value of each of the sources.