Suggestion Mining System for Customer Review Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current opinion mining techniques fail to effectively extract suggestions for improvement from customer reviews, as they are often buried within unstructured text and go beyond the scope of traditional sentiment analysis, making manual extraction time-consuming and inefficient.
Innovation Solution
A system and method using natural language processing (NLP) that employs a structured terminology, a thesaurus of terms related to suggestions, and specific patterns to identify and extract suggestions for improvement from text, including a parser for syntactic analysis and a suggestion review component to output relevant suggestions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional sentiment analysis is used to analyze customer reviews, then the process is simple and quick, but suggestions for improvement are not extracted
Solution Approach 1:
The patent segments the review text into different types of information (sentiment expressions and suggestion expressions) by using distinct patterns. The pattern matching process divides the analysis into two independent tracks: one for sentiment analysis and another for suggestion extraction, allowing both to be processed simultaneously without interfering with each other.
Solution Approach 2:
The patent introduces an intermediary pattern-matching layer that mediates between the raw review text and the extraction process. This intermediary layer uses syntactic patterns as filters to identify and extract suggestion expressions, acting as a bridge that transforms unstructured text into structured suggestion data without losing information.
2Measurement precision
If manual extraction of suggestions is performed, then extraction precision is high, but time consumption is excessive
Solution Approach 1:
The system performs self-service by automatically extracting suggestions using pattern matching algorithms that operate independently of human intervention. The pattern-based approach enables the system to autonomously identify and extract suggestion expressions from review text, eliminating the need for manual processing while maintaining high precision through carefully designed syntactic patterns.
Solution Approach 2:
The patent replaces the mechanical manual extraction process with an automated computational system. Instead of human analysts manually reading and extracting suggestions, the system uses pattern matching algorithms that computationally identify suggestion expressions based on syntactic patterns, dramatically reducing time consumption while maintaining extraction precision.
3Productivity
If automated extraction methods are implemented, then processing speed increases, but extraction accuracy decreases
Solution Approach 1:
The patent changes the parameters of the extraction process by using syntactic pattern matching instead of simple keyword searching. The patterns are designed to capture specific grammatical structures and syntactic relationships that characterize suggestion expressions, allowing the system to achieve both high processing speed and high extraction accuracy by leveraging linguistic structure rather than relying solely on statistical methods.
4Loss of information
If complex NLP processing is applied to extract suggestions, then extraction completeness improves, but system complexity increases
Solution Approach 1:
The patent extracts only the essential syntactic patterns needed for suggestion identification from the complex NLP processing pipeline. By focusing on specific syntactic patterns (such as conditional structures, modal verbs, and comparative phrases) rather than implementing a complete NLP system, the solution achieves high suggestion completeness while keeping the system relatively simple and maintainable.
Data Source
AI summary
A system and method for extraction of suggestions for improvement form a corpus of documents, such as customer reviews, are disclosed. A structured terminology provided or a topic includes a set of semantic classes, each including a set of terms. A thesaurus of terms relating to suggestions of improvement is provided. Text elements of text strings in the documents which are instances of terms in the structured terminology are labeled with the corresponding semantic class and text elements which are instances of terms in the thesaurus are also labeled. A set of patterns is applied to the labeled text strings to identify suggestions of improvement expressions. The patterns define syntactic relations between text elements, some of which are required to be instances of one of the terms in a particular semantic class or thesaurus. A set of suggestions for improvements is output based on the identified suggestions of improvement expressions.