SVO-based taxonomy-driven text analytics

• US 9,465,790 B2
• Filed: 09/09/2013
• Issued: 10/11/2016
• Est. Priority Date: 11/07/2012
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Kimelfeld, Benny, Li, Yunyao, Vaithyanathan, Shivakumar
• Primary Examiner(s): PULLIAS, JESSE SCOTT
• Application Number:
  US14/021,155
• Publication Number:
  US 20140129213A1
• Time in Patent Office: 1,128 Days
• Field of Search: 704/1, 704/9, 704/10
• US Class Current: 1/1
• CPC Class Codes: G06F 40/211 Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars

Abstract

Textual data is organized into statement clusters. Sentences are extracted from textual data and parsed. A verb usage pattern is identified and an SVO triplet is determined. The SVO triplet is compared to a taxonomy associated with the domain of the data and a sentiment is derived. A statement cluster is constructed comprising a higher level SVO triplet sensitive to the taxonomy and verb usage pattern, as well as the derived sentiment. Accordingly, the statement clusters may be organized by grouping.

Claims

1. A method comprising:

receiving textual data, and storing the received data in memory;

analyzing the stored data, wherein the analysis comprises;

identifying at least one sentence from the stored data;

parsing parts of speech of each word of the at least one identified sentence using a linguistic parser, including parsing a verb from the at least one parsed sentence, and identifying a verb usage pattern in the at least one identified sentence;

forming a low level subject-verb-object (SVO) triplet for the at least one parsed sentence, including identifying a subject, verb, and object of the at least one parsed sentence, wherein the identification of the subject, verb, and object comprises joining the identified verb usage pattern with a form of the identified verb to ascertain linguistic taxonomy;

forming a high level SVO triplet for the at least one parsed sentence, including determining a subject category for the subject, a verb category for the verb, and an object category for the object based on the taxonomy; and

classifying the at least one parsed sentence based on the high level SVO triplet; and

summarizing the analysis of the stored data, including;

producing an analysis report reflective of the analysis; and

converting the produced analysis report into a summary report reflective of the analysis report, including clustering the received textual data into one or more statement clusters, wherein the summary report comprises the statement clusters.

2. The method of claim 1, further comprising deriving sentiment from the identified subject, verb, and object.

3. The method of claim 2, wherein the derived sentiment is determined from a predefined category selected from the group consisting of:

positive, neutral, and negative.

4. The method of claim 2, wherein the sentence is further classified with respect to the derived sentiment.

5. The method of claim 1, wherein the taxonomy is a domain specific taxonomy associated with the received data.

6. The method of claim 1, wherein the joining of the identified verb usage pattern with a form of the identified verb comprises providing a mapping from the verb usage pattern to the identified verb based on a reference to an existing linguistic resource.

7. The method of claim 1, wherein the form of the identified verb is a root form of the identified verb.

8. The method of claim 1, wherein the parsing comprises creating a parse tree to provide a tag for each token in the at least one identified sentence.