Insight and algorithmic clustering for automated synthesis

• US 9,336,302 B1
• Filed: 03/14/2013
• Issued: 05/10/2016
• Est. Priority Date: 07/20/2012
• Status: Active Grant

• Current Assignee: OOL LLC
• Original Assignee: Zuci Realty LLC
• Inventors: Swamy, Gitanjali
• Primary Examiner(s): Hoang, Son T
• Application Number:
  US13/826,338
• Time in Patent Office: 1,153 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 16/217 Database tuning G06F16/2282 takes precedence; database performance monitoring G06F11/3409
• CPC Class Codes: G06F 16/2428 Query predicate definition using graphical user interfaces, including menus and forms G06F16/2423 takes precedence
• CPC Class Codes: G06F 16/2465 Query processing support for facilitating data mining operations in structured databases
• CPC Class Codes: G06F 16/248 Presentation of query results
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/3326 using relevance feedback from the user, e.g. relevance feedback on documents, documents sets, document terms or passages
• CPC Class Codes: G06F 16/3328 using graphical result space presentation or visualisation
• CPC Class Codes: G06F 16/355 Class or cluster creation or modification
• CPC Class Codes: G06F 18/24317 Piecewise classification, i.e. whereby each classification requires several discriminant rules
• CPC Class Codes: G16B 40/00 ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding

Abstract

A decision support system and method, which receives user inputs comprising: at least one user criterion, and at least one user input tuning parameter representing user tradeoff preferences for producing an output; and selectively produces an output of tagged data from a clustered database in dependence on the at least one user criterion, the at least one user input tuning parameter, and a distance function; receives at least one reference-user input parameter representing the at least one reference-user'"'"'s analysis of the tagged data and the corresponding user inputs, to adapt the distance function in accordance with the reference-user inputs as a feedback signal; and clusters the database in dependence on at least the distance function, wherein the reference-user acts to optimize the distance function based on the user inputs and the output, and on at least one reference-user inference.

Claims

1. A decision support system, comprising:

a user input port configured to receive user inputs from a first user comprising at least one user selection criterion and at least one user input tuning parameter representing user tradeoff preferences for defining an output of tagged data within a tagged data set from a database;

a reference-user input configured to receive at least one reference-user input parameter from an identified at least one reference-user different from the first user, representing at least the at least one reference-user'"'"'s classification inference based on analysis of the tagged data and the corresponding at least one user selection criterion, wherein the user and the at least one reference-user each being human;

at least one hardware processor configured to;

cluster the tagged data set from the database according to a distance function representing a quantitative correspondence of a tagged datum to a putative classification;

selectively produce an output of the tagged data from the database in dependence on at least the at least one user criterion, the at least one reference-user input tuning parameter, and the distance function;

identify the at least one reference-user from a plurality of possible reference-users based on at least prior quality of classification of data which clusters together with the tagged data by each respective possible reference-user;

adapt the distance function in accordance with the reference-user input tuning parameter from the identified at least one reference-user as a feedback signal, wherein the reference-user acts to optimize the distance function based on the user inputs and the output, and on the at least one reference-user'"'"'s classification inference;

receive new data into the database; and

re-cluster the database comprising the new data before another at least one reference-user input tuning parameter is received, in dependence on at least the adapted distance function; and

an information repository configured to replace the distance function with the optimized distance function.

2. A decision support method utilizing at least one computing device, the method comprising:

receiving user inputs from a user, comprising;

at least one user criterion, and

at least one user input tuning parameter representing user tradeoff preferences for producing an output;

selectively producing the output of tagged data from a database clustered according to classifications, the output being generated by at least one automated processor in dependence on the at least one user criterion, the at least one user input tuning parameter, and a distance function representing a quantitative correspondence of a tagged datum to a putative classification;

selecting at least one reference-user from a plurality of possible reference-users based on at least prior quality of classification of data which clusters together with the tagged data by each respective possible reference-user, the user and the plurality of possible reference-users each being human and distinct from each other;

receiving at least one reference-user input parameter, from the selected at least one reference-user distinct from the user, representing the selected at least one reference-user'"'"'s classification inference based on analysis of the tagged data and the corresponding user inputs, to adapt the distance function in accordance with the reference-user inputs as a feedback signal;

receiving new data into the database; and

re-clustering the database comprising the new data before another at least one reference-user input parameter is received, in dependence on at least the adapted distance function with the at least one automated processor,

wherein the reference-user acts to optimize the distance function based on the user inputs and the output, and on the at least one reference-user'"'"'s classification inference.

3. The method according to claim 2, wherein the at least one reference-user'"'"'s classification inference causes the re-clustering to distribute members of a cluster of the database into a plurality of different clusters.

4. The method according to claim 2, wherein the tagged data comprises a valuation or rating.

5. The method according to claim 4, wherein at least the distance function is further adapted based on the received new data after at least one reference-user input parameter is received.

6. The method according to claim 2, wherein the reference-user'"'"'s classification inference represents at least one of a valuation and a validation.

7. The method according to claim 2, wherein the user input tuning parameter comprises a quantitative variable that impacts a plurality of dimensions hidden by the clustering of the database.

8. The method according to claim 7, wherein the hidden dimensions comprise at least one of completeness, timeliness, correctness, coverage, and confidence.

9. The method according to claim 2, wherein the user input tuning parameter balances completeness and correctness of the tagged data in the output.

10. A decision support system, comprising:

a user input port configured to receive user inputs from a user comprising at least one user criterion and at least one user input tuning parameter representing user tradeoff preferences for producing an output from a system which selectively produces an output of tagged data from a tagged data set within a multidimensional data space comprising a database, in dependence on the at least one user criterion, the at least one user input tuning parameter, and a distance function representing a quantitative measurement of a distance relationship of respective locations within the multidimensional data space;

at least one hardware processor configured to select at least one reference-user from a plurality of possible reference-users based on at least an accuracy of at least one prior reference-user input parameter according to an objective criterion, the user and the plurality of possible reference-users each being human and distinct from each other;

an automated reference-user agent comprising at least one hardware processor, configured;

to receive at least one reference-user input parameter representing the selected at least one reference-user'"'"'s inference based on an analysis of the tagged data and the corresponding user inputs comprising the at least one user criterion and the at least one user input tuning parameter representing the user tradeoff preferences; and

to optimize the distance function in accordance with the at least;

the reference-user input parameter as a feedback signal;

the user inputs comprising the at least one user criterion and the at least one user input tuning parameter representing the user tradeoff preferences; and

the output of tagged data;

to cluster the database according to the optimized distance function;

to receive new data into the database; and

to re-cluster the database comprising the new data before another at least one reference-user input parameter is received, in dependence on at least the optimized distance function; and

an information repository configured to store the output of tagged data.

11. A decision support method utilizing at least a computing device, the method comprising:

receiving user inputs from a user, comprising;

at least one user criterion, and

at least one user input tuning parameter representing user tradeoff preferences;

selectively producing an output from an automated system of tagged data in dependence on the at least one user criterion, the at least one user input tuning parameter, and a distance function, the distance function representing a quantitative metric of a correspondence of a tagged datum to a putative classification;

automatically selecting at least one human user from a plurality of possible users based on at least an accuracy according to an objective criterion of prior classification of at least one tagged datum, the selected at least one human user being at least one reference-user and distinct from the user;

receiving by an automated reference-user agent at least one reference-user input parameter representing the at least one reference-user'"'"'s inference derived from at least analysis of the tagged data and the corresponding user inputs;

adapting the distance function in accordance with the at least one reference-user input parameter as a feedback signal, to optimize the distance function based on at least the user inputs and the output, and the at least one reference-user input parameter, wherein the reference-user acts to optimize the distance function based on the user inputs and the output, and on the at least one reference-user'"'"'s inference;

clustering the tagged data according to the adapted distance function;

receiving new data into the database after the clustering; and

re-clustering the database comprising the new data before another at least one reference-user input parameter is received, in dependence on at least the adapted distance function; and

storing the clustered tagged data in an information repository.

12. An information access method utilizing at least one computing device, the method comprising:

receiving a semantic user input comprising an explicit semantic indication of interest in information from a human user;

determining a context of the user distinctly from the semantic user input comprising an implicit indication of interest in information;

automatically clustering a database in dependence on a range of context-dependent distance functions, wherein a plurality of distance functions are provided for the same database, each respective context-dependent distance function representing a quantitative correspondence of a respective tagged datum to a putative cluster classification subject to a respective context corresponding to a respective cluster;

producing an output of at least a portion of the tagged data from a database, in dependence on at least the semantic user input, the determined context, and a context-appropriate distance function;

monitoring a user interaction with the output by the user;

classifying a plurality of users including the user;

modifying the plurality of distance functions in dependence on at least the monitored user interaction; and

distinguishing between different classes of users with respect to the selection and modifying of respective ones of the plurality of distance functions;

reclassifying the at least portion of the tagged data into different clusters based on the plurality of modified distance functions and the different classes of users.

13. The method according to claim 12, further comprising selecting at least one commercial advertisement extrinsic to the tagged data from the clustered database for presentation to the user, in dependence on at least:

at least one of the semantic user input and the output of tagged data, and the determined context.

14. The method according to claim 13, wherein the selecting the at least one commercial advertisement is further dependent on the modified distance function.

15. The method according to claim 14, wherein the monitoring comprises monitoring the user interaction with the at least one commercial advertisement, wherein the commercial advertisement is selected in dependence on the modified distance function, and the distance function is modified based on the user interaction with a selected advertisement.

16. The method according to claim 12, further comprising re-clustering the database in dependence on the plurality of modified distance functions.

17. The method according to claim 12, further comprising determining at least one reference-user from a set of users distinct from the user, based on at least one objective fitness criterion, and selectively modifying the distance function dependent on a reference-user input in preference to a non-reference-user input.

18. The method according to claim 17, wherein a user-reference input is associated with a respective reference-user in dependence on the determined context of the user.

19. An information processing method utilizing at least one computing device, wherein the method comprising:

clustering a database comprising a plurality of information records according to semantic information contained therein and a plurality of context-dependent distance functions, wherein the same information record within the database is classified in a plurality of different clusters in dependence on a respective context-dependent distance function corresponding to a respective cluster, wherein a common semantic query to the database yields different outputs over a range of contexts;

receiving at least a user semantic input from a user;

determining a contextual ambiguity from the at least user semantic input;

soliciting resolution information to the contextual ambiguity from the user;

producing an output identifying information records from the database in dependence on the at least user semantic input, and a determined context with an associated context-dependent distance function;

receiving user feedback on the output;

modifying the associated context-dependent distance function in dependence on the user feedback and the determined context; and

producing a follow-up output identifying information records from the database in dependence on at least the at least user semantic input, the resolution information, and at least one context-dependent distance function selected from the plurality of context-dependent distance functions in dependence on the resolution information.

20. The method according to claim 19, further comprising selecting at least one commercial advertisement extrinsic to the information records identified by the follow-up output in the database for presentation to the user, in dependence on at least:

the at least user semantic input, and the resolution information.

21. The method according to claim 20, wherein the selecting the at least one commercial advertisement is further dependent on the determined context and the selected at least one context-dependent distance function.

22. The method according to claim 20, wherein the selecting the at least one commercial advertisement is dependent on at least:

the at least user semantic input, the resolution information, the determined context, and an appropriate context-dependent distance function.

23. The method according to claim 19, wherein the monitoring comprises monitoring a user interaction with at least one commercial advertisement presented to the user as part of the output, wherein the commercial advertisement is selected in dependence on the determined context and at least one context-dependent distance function, and the at least one context-dependent distance function is modified based on the user interaction with at least one selected advertisement.

24. The method according to claim 19, further comprising re-clustering the database in dependence on the at least one modified context-dependent distance function.

25. The method according to claim 19, further comprising classifying a plurality of users, and distinguishing between different classifications of users with respect to the selection and modifying of respective ones of a plurality of context-dependent distance functions.

26. The method according to claim 19, further comprising assigning a reference-user status to at least one user within a set of users, based on at least one fitness criterion, and selectively weighting a user contribution to a modification of a respective context-dependent distance function dependent on at least the reference-user status of the respective user.

27. The method according to claim 26, wherein the reference-user status is assigned with respect to a context, and a user input of the respective reference-user is associated with a respective context-dependent distance function in dependence on the context.