LEARNING OPINION-RELATED PATTERNS FOR CONTEXTUAL AND DOMAIN-DEPENDENT OPINION DETECTION
First Claim
1. A method for extracting opinion-related patterns, comprising:
- receiving a corpus of reviews, each of the reviews in the corpus including an explicit rating of a topic;
partitioning at least a portion of the reviews among a predefined plurality of classes, based on the explicit ranking;
identifying syntactic relations in a text portion of each of the reviews in the portion of the reviews, each of the identified syntactic relations including a first term comprising an adjective and a second term comprising a noun, the adjective serving as a modifier or attribute of the respective noun;
generating a set of patterns, each of the patterns having at least one of the identified syntactic relations as an instance;
with a processor, clustering the patterns into a set of clusters based on a set of features, at least one of the features in the set of features being based on occurrences, in the predefined classes, of the instances of the patterns;
selecting a subset of the clusters and assigning a polarity to patterns in the respective clusters in the subset.
1 Assignment
0 Petitions
Accused Products
Abstract
A method for extracting opinion-related patterns includes receiving a corpus of reviews, the reviews each including an explicit rating of a topic. The reviews are partitioned among a predefined plurality of classes, based on the ranking. Syntactic relations are identified in each review. The syntactic relations may each include an adjective and a noun. A set of patterns is generated, each of the patterns having at least one of the identified syntactic relations as an instance and the patterns clustered into a set of clusters based on a set of features. At least one of the features is based on occurrences, in the predefined classes, of the instances of the patterns. A polarity is assigned to ones of the clusters and propagated to patterns in the respective clusters. The polarity-labeled patterns can each be instantiated as a contextual rule for opinion mining.
-
Citations
24 Claims
-
1. A method for extracting opinion-related patterns, comprising:
-
receiving a corpus of reviews, each of the reviews in the corpus including an explicit rating of a topic; partitioning at least a portion of the reviews among a predefined plurality of classes, based on the explicit ranking; identifying syntactic relations in a text portion of each of the reviews in the portion of the reviews, each of the identified syntactic relations including a first term comprising an adjective and a second term comprising a noun, the adjective serving as a modifier or attribute of the respective noun; generating a set of patterns, each of the patterns having at least one of the identified syntactic relations as an instance; with a processor, clustering the patterns into a set of clusters based on a set of features, at least one of the features in the set of features being based on occurrences, in the predefined classes, of the instances of the patterns; selecting a subset of the clusters and assigning a polarity to patterns in the respective clusters in the subset. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
-
-
21. A system for generating contextual rules for opinion detection comprising:
-
a review classifier for partitioning reviews among a predefined plurality of classes, based on an explicit ranking of a topic associated with each of the reviews; a parser for identifying syntactic relations in a text portion of each of the reviews in the portion of the reviews, each of the identified syntactic relations including a first term comprising an adjective and a second term comprising a noun, the adjective serving as a modifier or attribute of the respective noun; a pattern extractor for generating a set of patterns, each of the generated patterns having at least one of the identified syntactic relations as an instance; a clustering component for clustering the patterns into a set of clusters based on a set of features, at least one of the features in the set of features being based on occurrences, in the predefined classes, of the instances of the patterns; a contextual rule generator for generating contextual rules for the patterns in a plurality of the clusters, to which a polarity has been assigned; and a processor for implementing the review classifier, parser, pattern extractor, clustering component, and contextual rule generator. - View Dependent Claims (22, 23)
-
-
24. A method for generating contextual rules, comprising:
-
receiving a corpus of documents, each of the documents in the corpus being associated with an explicit rating of a topic; partitioning at least a portion of the documents among a predefined plurality of classes, based on the explicit ranking; identifying opinion instances in the documents, each of the opinion instances comprising an instance of a term in an associated polar vocabulary; identifying syntactic relations in the documents, each of the identified syntactic relations including a first term comprising an adjective that is not an instance of a term in the polar vocabulary and a second term comprising a noun, the adjective serving as a modifier or attribute of the respective noun; generating a set of patterns, each of the patterns having at least one of the identified syntactic relations as an instance; extracting features for each of the patterns, the extracted features including features based on occurrences, in the predefined classes, of the instances of the patterns; with a processor, clustering the patterns into a set of clusters based on the extracted features; generating contextual rules for the patterns in a plurality of the clusters for which a polarity has been assigned.
-
Specification