Discovery, characterization, and analysis of interpersonal relationships extracted from unstructured text data

Patent metadata

Field	Value
Publication number	US-11003716-B2
Application number	US-201715402287-A
Country	US
Kind code	B2
Filing date	Jan 10, 2017
Priority date	Jan 10, 2017
Publication date	May 11, 2021
Grant date	May 11, 2021

Abstract

Embodiments for discovery and analysis of interpersonal relationships from a collection of unstructured text data by a processor. A relationship between one or more entities and extracted text data from a plurality of unstructured text data may be identified such that the relationship includes a sentiment of the relationship, a type of relationship, temporal information, or a combination thereof. The one or more entities may be associated with a knowledge graph based on an ontology of concepts representing a domain knowledge. The extracted information and the identified relationship may be automatically aggregated into a multi-graph representation.

First claim

The invention claimed is:

1. A method, by a processor, for discovery and analysis of interpersonal relationships from a collection of unstructured text data, comprising:
identifying a relationship between one or more entities and extracted text data from a plurality of unstructured text data such that the relationship includes a sentiment of the relationship, a type of relationship, and temporal information indicative of a timeframe of events occurring over a duration of the relationship, wherein the relationship and a connection between the relationship and a partial name, title, or role of the one or more entities is identified by inference according to references contained only within the plurality of unstructured text data notwithstanding whether the relationship between the one or more entities is explicitly named within the extracted text data, and wherein the one or more entities may be associated with a knowledge graph based on an ontology of concepts representing a domain knowledge; and
automatically aggregating the extracted text data and the identified relationship into a multi-graph representation, wherein the aggregating includes enhancing a social graph of the multi-graph representation according to aggregated sentiment information extracted over a defined time period, the sentiment information including qualitative descriptions of the relationship, wherein the qualitative descriptions include an intensity, a formality, and the duration of the relationship as evidenced by the temporal information indicative of the timeframe of events occurring over the duration of the relationship.

2. The method of claim 1 , further including:
linking the one or more entities to the extracted text data to create the relationship between one or more entities and extracted text data using the knowledge graph; or
associating an entry in the knowledge graph to each of the one or more entities in the extracted text data to analyze the relationship between one or more entities.

3. The method of claim 1 , further including assigning a confidence score to the sentiment of the relationship, the type of relationship, and the temporal information of the extracted data.

4. The method of claim 1 , further including, prior to the identifying, parsing the plurality of unstructured text data in order to extract the extracted data.

5. The method of claim 1 , further including detecting each semantic indication referencing the sentiment of the relationship, the type of relationship, and the temporal information in order to determine the relationship.

6. The method of claim 1 , further including analyzing additional text data associated with the extracted text data in the plurality of unstructured text to assist in determining the sentiment of the relationship, the type of relationship, and the temporal information.

7. The method of claim 1 , further including:
identifying an indication in the plurality of unstructured text data that references the one or more entities;
assigning the indication to the one or more entities;
performing a co-reference resolution for the one or more entities referenced in the plurality of unstructured text data by one or more incomplete semantic names relating to the indication;
performing a reference by role resolution for the one or more entities by annotating the indication with a defined role of the one or more entities;
detecting a semantic reference of one or more interpersonal relationships between the one or more entities from the plurality of unstructured text data;
analyzing text data preceding and subsequent to the semantic reference in the plurality of unstructured text data to characterize a semantic relationship according to the sentiment, a frequency, emotional state, relationship hierarchy, or a combination thereof; or
displaying the multi-graph representation in an interactive graphical user interface (GUI).

8. A system for discovery and analysis of interpersonal relationships from a collection of unstructured text data, comprising:
one or more computers with executable instructions that when executed cause the system to:
identify a relationship between one or more entities and extracted text data from a plurality of unstructured text data such that the relationship includes a sentiment of the relationship, a type of relationship, and temporal information indicative of a timeframe of events occurring over a duration of the relationship, wherein the relationship and a connection between the relationship and a partial name, title, or role of the one or more entities is identified by inference according to references contained only within the plurality of unstructured text data notwithstanding whether the relationship between the one or more entities is explicitly named within the extracted text data, and wherein the one or more entities may be associated with a knowledge graph based on an ontology of concepts representing a domain knowledge; and
automatically aggregate the extracted text data and the identified relationship into a multi-graph representation, wherein the aggregating includes enhancing a social graph of the multi-graph representation according to aggregated sentiment information extracted over a defined time period, the sentiment information including qualitative descriptions of the relationship, wherein the qualitative descriptions include an intensity, a formality, and the duration of the relationship as evidenced by the temporal information indicative of the timeframe of events occurring over the duration of the relationship.

9. The system of claim 8 , wherein the executable instructions:
link the one or more entities to the extracted text data to create the relationship between one or more entities and extracted text data using the knowledge graph; or
associate an entry in the knowledge graph to each of the one or more entities in the extracted text data to analyze the relationship between one or more entities.

10. The system of claim 8 , wherein the executable instructions assign a confidence score to the sentiment of the relationship, the type of relationship, and the temporal information of the extracted data.

11. The system of claim 8 , wherein the executable instructions, prior to identifying the relationship, parse the plurality of unstructured text data in order to extract the extracted data.

12. The system of claim 8 , wherein the executable instructions detect each semantic indication referencing the sentiment of the relationship, the type of relationship, and the temporal information in order to determine the relationship.

13. The system of claim 8 , wherein the executable instructions analyze additional text data associated with the extracted text data in the plurality of unstructured text to assist in determining the sentiment of the relationship, the type of relationship, and the temporal information.

14. The system of claim 8 , wherein the executable instructions:
identify an indication in the plurality of unstructured text data that references the one or more entities;
assign the indication to the one or more entities;
perform a co-reference resolution for the one or more entities referenced in the plurality of unstructured text data by one or more incomplete semantic names relating to the indication;
perform a reference by role resolution for the one or more entities by annotating the indication with a defined role of the one or more entities;
detect a semantic reference of one or more interpersonal relationships between the one or more entities from the plurality of unstructured text data;
analyze text data preceding and subsequent to the semantic reference in t

Assignees

• IBM

Inventors

• Bonin Francesca
• Daly Elizabeth M
• Deleris Lea A
• Deparis Stephane
• Hou Yufang
• Jochim Charles A
• LASSOUED Yassine

Classifications

• G06N5/025 Primary

  Extracting rules from data · CPC title

• G06F16/9024 Primary

  Graphs; Linked lists (G06F16/9027 takes precedence) · CPC title

• G06F40/30

  Semantic analysis · CPC title