Method and system for identifying data and users of interest from patterns of user interaction with existing data

• US 9,465,889 B2
• Filed: 12/07/2012
• Issued: 10/11/2016
• Est. Priority Date: 07/05/2012
• Status: Active Grant

• Current Assignee: Physion Consulting LLC
• Original Assignee: Physion Consulting LLC
• Inventors: Wark, Barry
• Primary Examiner(s): To, Baoquoc N
• Application Number:
  US13/708,701
• Publication Number:
  US 20140012870A1
• Time in Patent Office: 1,404 Days
• Field of Search: 707/705, 707/706
• US Class Current: 1/1
• CPC Class Codes: G06F 16/2465 Query processing support for facilitating data mining operations in structured databases
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/90335 Query processing
• CPC Class Codes: G06Q 10/101 Collaborative creation, e.g. joint development of products or services
• CPC Class Codes: G06Q 10/105 Human resources

Abstract

A system and method for identifying relevant data and experts from a large data set that are relevant to a researcher, scientific project, or other analysis project using a recommendation algorithm utilizing distance metrics based on structural, annotation, and usage information associated with data elements.

Claims

1. A computer-based recommendation system comprising:

a user interface connected to at least one processor; and

at least one set of memory for storing data element information, user information, and computer readable instructions that, when executed by the processor, cause the processor to perform the steps of;

a. receiving a user input query from the user interface;

b. calculating a distance between each data element and user based on;

i. similarity of structural data associated with each data element,

ii. similarity of annotation information associated with each data element, and

iii. usage information associated with data elements,

wherein at least some of the structural data is selected from a list comprising;

type of data, domain relationships between data items, experimental subject, subject species, parameters of acquisition, sequence or timing,

wherein at least some of the annotation information is selected from a list comprising;

keyword tags, key-value metadata, region-of-interest annotations in time and/or space, or user-inputted information, and

wherein at least some of the usage information is selected from a list comprising;

order of interaction, views, incorporation in analysis, sharing with others, publishing, recommending to the user, or user interaction with the data elements;

c. inputting the calculated distances between each data element into a recommendation algorithm to identify data elements or users based on the query input, wherein the recommendation algorithm is based in part on the magnitude of the calculated distances;

and

d. outputting a list of recommended data elements or users based on the query input.

2. The system of claim 1, further including the step of updating the calculated distances based on new or updated data element and user information that is uploaded to the at least one set of memory.

3. The system of claim 1 wherein the recommendation algorithm performs a clustering or co-clustering step.

4. The system of claim 1, wherein the usage information is synthesized information.

5. The system of claim 1, further configured to track and update usage information.

6. The system of claim 1, wherein calculating distance includes a weighting factor based on the type of structural, annotation, or usage information.

7. A computer implemented recommendation method comprising the steps:

creating a database comprising multiple users and data elements represented by nodes, each node in the database having a structural, annotation, and usage field associated therewith;

determining distances between each node of the database based on the structural, annotation, and usage field for each node,

wherein at least one information set for the structural field is selected from a list comprising;

type of data, domain relationships between data items, experimental subject, subject species, parameters of acquisition, sequence or timing,

wherein at least one information set for the annotation field is selected from a list comprising;

keyword tags, key-value metadata, region-of-interest annotations in time and/or space, or user-inputted information,

wherein at least one information set for the usage field is selected from a list comprising;

order of interaction, views, incorporation in analysis, sharing with others, publishing, recommending to the user, or user interaction with the data elements;

calculating a distance between the nodes and an input query, the distance being a function of the determined distances based on the structural, annotation, and usage field for each node, and wherein the step of calculating the distance includes application of a weighting factor;

selecting at least one data element or user based on the magnitude of the calculated distance to the input query; and

outputting to a display the selected data element or user.

8. A computer-based method for identifying a domain expert comprising the steps of:

receiving from a user, into a terminal connected to a database and a processor, an initial query related to a research subject;

identifying multiple data elements and users associated with the database, wherein each data element and user is represented by a separate and distinct node;

determining using a processor a distance between each node based on;

1) similarity of structural data associated with each node,

2) similarity of annotations information associated with each node, and

3) usage information associated with data elements;

determining using the processor a distance between comparing the initial query to the nodes of the large database based on weighted;

1) structural data information,

2) annotation information, and

3) usage information; and

outputting a recommendation node of a domain expert, based on the magnitude of the weighted distance step, to an output device.

9. The method of claim 8, wherein structural data is based in part on conditions under which nodes were acquired.

10. The method of claim 8, wherein annotations may include:

keywords, key-value metadata, notes, ratings, circling, symbolic marking, or highlighting.

11. The method of claim 8, wherein usage information is synthesized information.

12. The method of claim 11, wherein the synthesized information is obtained from the identified users.

13. The method of claim 8, wherein the usage information is obtained from the identified users.

14. The method of claim 8, wherein the recommendation system may be configured to mask data elements added by an inquiring user.

15. The method of claim 8, wherein the initial query is the user and is triggered by the user logging in.

16. The method of claim 15, further including automatically updating structural, annotation and usage information in the at least one database.

17. The method of claim 8, wherein the recommendation system is comprised of at least one recommendation algorithm.

18. The method of claim 17, wherein the recommendation algorithm performs a clustering or co-clustering step.

19. The method of claim 8, wherein a graph representation is created for use with the distance calculation.

20. The method of claim 19, wherein a graph search algorithm is used to calculate distance from the graph representation.