Neural Network Aspect-Sentiment Analysis for Attention-Property-Aware Ratings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sentiment analysis approaches fail to extract the strength of user attention to specific properties of products or services, which can dominate a user's attitude and affect the generosity of their rating, leading to inaccurate feedback analysis.
Innovation Solution
A neural network-based method that extracts aspect-sentiment pairs from user reviews to estimate an attention-property-aware rating, incorporating both implicit and explicit features of the text, and performs a response based on the estimated rating.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing sentiment analysis approaches are used, then the analysis process is simple and fast, but the extraction precision of user attention strength to specific properties is poor
Solution Approach 1:
The patent segments the review text into aspect-sentiment pairs, where each pair consists of a specific product aspect and the user's sentiment toward it. This segmentation allows the system to identify and analyze user attention strength toward specific properties independently, improving extraction precision while maintaining manageable system complexity through structured data representation
Solution Approach 2:
The patent introduces a new dimension to sentiment analysis by extracting not just the sentiment polarity but also the attention strength dimension. The attention-property-aware rating model processes both the sentiment direction and the intensity of user attention toward different aspects, transforming traditional one-dimensional sentiment scoring into a multi-dimensional analysis that captures user attention strength
2Reliability
If existing sentiment analysis approaches are used, then the system is easy to implement, but the accuracy of feedback analysis is insufficient
Solution Approach 1:
The patent introduces aspect-sentiment pairs as an intermediary representation between the raw review text and the final rating prediction. This intermediary structure serves as a bridge that captures both sentiment information and attention strength, enabling more accurate feedback analysis while organizing the complexity into a manageable processing pipeline
Solution Approach 2:
The patent replaces traditional mechanical sentiment analysis methods with a neural network-based approach that processes aspect-sentiment pairs to generate attention-property-aware ratings. This substitution enables the system to capture non-linear relationships and subtle patterns in user feedback, improving accuracy while abstracting the implementation complexity into a unified deep learning model
3Loss of information
If traditional sentiment analysis is used, then the processing speed is high, but the information granularity of user attitudes is insufficient
Solution Approach 1:
The patent segments user feedback into granular aspect-sentiment pairs, where each pair represents a specific product aspect and the user's targeted sentiment. This segmentation preserves fine-grained information about user attitudes toward different properties while maintaining efficient processing through structured data formats that can be quickly analyzed by the neural network
Data Source
AI summary
Rating prediction systems and methods include extracting aspect-sentiment pairs from an input text. An attention-property-aware rating is estimated for the input text using the extracted aspect-sentiment pairs with a neural network that captures implicit and explicit features of the text. A response to the input text is performed based on the estimated rating.