Expression Extraction Device for Sentiment Analysis

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing sentiment analysis techniques face challenges in efficiently extracting and analyzing evaluation expressions from text due to the variability of expressions across different fields, requiring manual effort for dictionary creation and relying on specific document structures or search engines, which limits processing efficiency and accuracy.

Innovation Solution

An expression extraction device that includes a registered expression storage unit, an expression extraction unit, a registered expression detection unit, and a polarity judgment unit to automatically extract and judge the polarity of evaluation expressions based on predefined polarities and conjunctions, using ordinary and adversative/concessive conjunctions to determine the polarity of series of expressions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual dictionary production is used, then evaluation expressions can be extracted, but the process is time-consuming and cannot handle diverse expressions across various fields

Engineering Contradiction:

Improveextraction accuracyVSAvoiddictionary production time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system automatically learns and expands the dictionary by extracting evaluation expressions from text and determining their polarities without manual intervention. The machine learning model processes diverse expressions across multiple fields autonomously, eliminating the need for manual dictionary creation while maintaining high extraction accuracy.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system changes the parameter of dictionary creation from manual to automated by using machine learning techniques. The model learns from text data and automatically expands the dictionary with new evaluation expressions and their polarities, transforming the time-consuming manual process into an efficient automated system.

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If search engine-based polarity judgment is used, then evaluation expressions can be classified, but processing efficiency is low and results depend on document contents

Engineering Contradiction:

Improvepolarity judgment accuracyVSAvoidprocessing efficiency

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The system replaces the mechanical search engine-based polarity judgment with a machine learning-based automated classification. The machine learning model directly processes text data to determine polarities, eliminating the need for search engine queries and significantly improving processing efficiency while maintaining accurate polarity judgment.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system extracts the polarity judgment function from the search engine dependency and integrates it directly into the machine learning model. This extraction eliminates the intermediate search engine step and allows for direct, efficient polarity determination based on the learned patterns from training data.

Inventive Principle:

Principle #2Taking out (Extraction)

3Measurement precision

If document structure dependency is used, then polarity can be determined, but the method is limited to specific document types and reduces adaptability

Engineering Contradiction:

Improvepolarity determination accuracyVSAvoiddocument type flexibility

Core Design Contradiction:

Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The machine learning model is trained to handle diverse evaluation expressions across multiple fields and document types universally. The model learns general patterns from varied text data, enabling it to accurately determine polarities in different document types without requiring type-specific processing methods, thus achieving both high accuracy and broad adaptability.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS7475007B2Expression extraction device, expression extraction method, and recording medium

Publication Date: 2009.01.06 SAP SE

AI summary

Provided is an expression extraction device for extracting evaluation expressions from text having descriptions on evaluations of a specific evaluation target, which includes a registered expression storage unit for registering an evaluation expression including a predetermined polarity as a registered expression, an expression extraction unit for extracting multiple evaluation expressions and a conjunction expression from the text, a registered expression detection unit for detecting the evaluation expression including the registered expression registered with the registered expression storage unit out of the multiple evaluation expressions, and a polarity judgment unit for judging that the evaluation expression, which is in conjunction with the evaluation expression including the registered expression by means of the conjunction expression in a form of ordinary conjunction, and the series of evaluation expressions, which are not in conjunction with the evaluation expression by means of the conjunction expression in any form of the ordinary conjunction and adversative/concessive conjunction and are not in conjunction with each other by means of the conjunction expression in any form of the ordinary conjunction and the adversative/concessive conjunction, are of the same polarity as the registered expression.