Recommendation engine using inferred deep similarities for works of literature

• US 10,120,908 B2
• Filed: 05/06/2016
• Issued: 11/06/2018
• Est. Priority Date: 12/03/2013
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Allen, Corville O., Carrier, Scott R., Woods, Eric
• Primary Examiner(s): Hwa, Shyue Jiunn
• Application Number:
  US15/149,023
• Publication Number:
  US 20160253413A1
• Time in Patent Office: 914 Days
• CPC Class Codes: G06F 16/2457 with adaptation to user needs
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/313 Selection or weighting of terms for indexing
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G06F 16/335 Filtering based on additional data, e.g. user or group profiles filtering in web context G06F16/9535, G06F16/9536
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 16/38 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• CPC Class Codes: G06F 16/93 Document management systems

Abstract

A recommendation engine for works of literature uses patterns of flow and element similarities for scoring a first user-rated work of literature against one or more recommendation candidate works of literature. Cluster models are created using meta-data modeling the works of literature, the meta-data having literary element categories and instances within each category. Each instance is described by an index value (position in the literature) and significance value (e.g. weight or significance). Cluster finding process(es) invoked for each instance in each category find Similarity Concept clusters and Consistency Trend clusters, which are recorded into the cluster models representing each work of literature. The cluster model can be printed or displayed so that a user can visually understand the ebb and flow of each literary element in the literature, and may be digitally compared to other cluster models of other works of literature for potential recommendation to a user.

Claims

1. A method for comparing and optionally recommending works of digital literature, comprising:

performing, by a processor, deep semantic analysis to create at least a first digital heuristic model for at least a first work of digital literature, wherein the deep semantic analysis comprises machine learning by the computing platform, and wherein the heuristic model contains heuristics which are limited within a range of positional index values for one or more instances within one or more literary element categories;

determining by a processor a degree of similarity between the first digital heuristic model for a first work of digital literature and a second digital heuristic model for a second work of digital literature; and

producing, by a processor, via a user interface device a recommendation to a user regarding the degree of similarity;

wherein the first and second heuristic digital models reflect limited lengths of segments within each first and second respective work of digital literature, respectively, within which similar abstracted concepts, or similar abstracted relationships, or both similar abstracted concepts and similar abstracted relationships are detected using natural language processing, wherein the concepts and relationships were abstracted from actual concepts and relationships in the respective works of digital literature by deep semantic analysis.

2. The method as set forth in claim 1 wherein the first heuristic model comprises at least one literary element for which a first user has indicated at least one interest or disinterest.

3. The method as set forth in claim 2 wherein the second heuristic model comprises a literary element for which the first user has indicated at least one interest or disinterest.

4. The method as set forth in claim 2 wherein the second heuristic model comprises at least one literary element for which a second user has indicated at least one interest or disinterest.

5. The method as set forth in claim 1 wherein the accessing of a first heuristic model further comprises:

receiving by a processor meta-data modeling the first work of literature, the meta-data comprising one or more literary element categories, one or more instances within each literary element categories, and the positional index value and a significance value for each instance, wherein the significance value corresponds to a significance, weight or strength of an instance relative to other instances in the meta-data;

for each literary element category and for each instance within each literary element category of the first work of literature, invoking by a processor a heuristic finding process using a first control parameter limiting the range of index value variation and a second control parameter limiting the range of significance value variation in a found run length;

receiving by a processor from the heuristic finding process one or more heuristics found around one or more instances of one or more literary element categories; and

storing by the processor the one or more heuristics into the first heuristic model for the first work of literature.

6. The method as set forth in claim 5 wherein the first control parameter limiting the range of index value variation is set to a run length value.

7. A computer program product for comparing and optionally recommending works of digital literature, comprising:

a tangible, computer readable memory storage device;

program instructions stored by the computer readable memory storage device for causing a processor to, when executed, perform steps of;

perform deep semantic analysis to create at least a first digital heuristic model for at least a first work of digital literature, wherein the deep semantic analysis comprises machine learning by the computing platform, and wherein the heuristic model contains heuristics which are limited within a range of positional index values for one or more instances within one or more literary element categories;

determine a degree of similarity between the first digital heuristic model for a first work of digital literature and a second digital heuristic model for a second work of digital literature; and

produce via a user interface device a recommendation to a user regarding the degree of similarity;

wherein the first and second digital heuristic models reflect limited lengths of segments within each first and second respective work of digital literature, respectively, within which similar abstracted concepts, or similar abstracted relationships, or both similar abstracted concepts and similar abstracted relationships are detected using natural language processing, wherein the concepts and relationships were abstracted from actual concepts and relationships in the respective works of digital literature by deep semantic analysis.

8. The computer program product as set forth in claim 7 wherein the first heuristic model comprises at least one literary element for which a first user has indicated at least one interest or disinterest.

9. The computer program product as set forth in claim 8 wherein the second heuristic model comprises at least one literary element for which the first user has indicated at least one interest or disinterest.

10. The computer program product as set forth in claim 8 wherein the second heuristic model comprises at least one literary element for which a second user has indicated at least one interest or disinterest.

11. The computer program product as set forth in claim 7 wherein the program instructions for accessing of a first heuristic model further comprise program instructions for, when executed, causing a processor to:

receive meta-data modeling the first work of literature, the meta-data comprising one or more literary element categories, one or more instances within each literary element categories, and the positional index value and a significance value for each instance, wherein the significance value corresponds to a significance, weight or strength of an instance relative to other instances in the meta-data;

for each literary element category and for each instance within each literary element category of the first work of literature, invoke a heuristic finding process using a first control parameter limiting the range of index value variation and a second control parameter limiting the range of significance value variation in a found run length;

receive from the heuristic finding process one or more heuristics found around one or more instances of one or more literary element categories; and

store the one or more heuristics into the first heuristic model for the first work of literature.

12. The computer program product as set forth in claim 11 wherein the first control parameter limiting the range of index value variation is set to a run length value.

13. A system for comparing and optionally recommending works of digital literature, comprising:

a computer processor;

a tangible, computer readable memory storage device accessible by the computer processor;

program instructions stored by the computer readable memory storage device for causing the computer processor to, when executed, perform steps of;

perform deep semantic analysis to create at least a first digital heuristic model for at least a first work of digital literature, wherein the deep semantic analysis comprises machine learning by the computing platform, and wherein the heuristic model contains heuristics which are limited within a range of positional index values for one or more instances within one or more literary element categories;

determine a degree of similarity between the first digital heuristic model for a first work of digital literature and a second digital heuristic model for a second work of digital literature; and

produce via a user interface device a recommendation to a user regarding the degree of similarity;

wherein the first and second digital heuristic models reflect limited lengths of segments within each first and second respective work of digital literature, respectively, within which similar abstracted concepts, or similar abstracted relationships, or both similar abstracted concepts and similar abstracted relationships are detected using natural language processing, wherein the concepts and relationships were abstracted from actual concepts and relationships in the respective works of digital literature by deep semantic analysis.

14. The system as set forth in claim 13 wherein the first heuristic model comprises at least one literary element for which a first user has indicated at least one interest or disinterest.

15. The system as set forth in claim 14 wherein the second heuristic model comprises at least one literary element for which the first user has indicated at least one interest or disinterest.

16. The system as set forth in claim 14 wherein the second heuristic model comprises at least one literary element for which a second user has indicated at least one interest or disinterest.

17. The system as set forth in claim 13 wherein the program instructions for accessing of a first heuristic model further comprise program instructions for, when executed, causing the computer processor to:

receive meta-data modeling the first work of literature, the meta-data comprising one or more literary element categories, one or more instances within each literary element categories, and the positional index value and significance value for each instance, wherein the significance value corresponds to a significance, weight or strength of an instance relative to other instances in the meta-data;

for each literary element category and for each instance within each literary element category of the first work of literature, invoke a heuristic finding process using a first control parameter limiting the range of index value variation and a second control parameter limiting the range of significance value variation in a found run length;

receive from the heuristic finding process one or more heuristics found around one or more instances of one or more literary element categories; and

store the one or more heuristics into the first heuristic model for the first work of literature.

18. The system as set forth in claim 17 wherein the first control parameter limiting the range of index value variation is set to a run length value.