Review Clustering System for App Feedback Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
App developers face difficulties in effectively monitoring and analyzing the large volume of user reviews due to the sheer influx of feedback, making it challenging to identify and address issues raised by users.
Innovation Solution
A computer-implemented method and system that organizes user reviews by sampling, identifying unique terms, calculating frequency values, adjusting for rarity, determining similarities, and grouping reviews into clusters, allowing for more effective data analysis and ranking of clusters based on similarity and rating.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all user reviews are analyzed individually, then comprehensive feedback coverage is achieved, but analysis time and resource consumption increase significantly
Solution Approach 1:
The patent segments the large volume of user reviews into smaller clusters based on similarity in ratings and review text. Reviews with similar characteristics (rating values and text content) are grouped together, allowing developers to analyze cluster representatives rather than every individual review, thus reducing analysis time while maintaining comprehensive feedback coverage.
Solution Approach 2:
The patent merges similar reviews into clusters by identifying reviews with comparable rating values and text similarities. By combining multiple similar reviews into single clusters and selecting representative reviews from each cluster, the system reduces the total number of reviews needing detailed analysis while preserving the essential feedback information.
2Reliability
If every review is examined in detail, then all issues are identified, but developer productivity decreases due to overwhelming volume
Solution Approach 1:
The patent extracts representative reviews from each cluster that best summarize the cluster's characteristics. By selecting one or more representative reviews from each cluster and focusing analysis on these representatives, developers can identify issues comprehensively without examining every single review, thus maintaining issue identification completeness while improving productivity.
Solution Approach 2:
The patent applies partial action by analyzing only the representative reviews from each cluster rather than all reviews. This selective approach focuses computational and human resources on the most informative samples, achieving sufficient issue identification without the excessive effort of examining every review in detail.
3Loss of information
If reviews are organized without clustering, then individual review details are preserved, but pattern recognition and trend analysis become difficult
Solution Approach 1:
The patent segments reviews into clusters based on similarity metrics including rating values and text content. This segmentation organizes reviews into meaningful groups where patterns and trends become visible at the cluster level, while still preserving access to individual review details within each cluster for when needed.
Solution Approach 2:
The patent introduces a new organizational dimension by grouping reviews along similarity axes (rating similarity and text similarity). This creates a multi-dimensional organization structure that enables pattern recognition across multiple reviews simultaneously, while maintaining the ability to access individual review details when required.
Data Source
AI summary
Computer-implemented methods and systems are disclosed for organizing user reviews, especially computer app reviews, into clusters and ranking the clusters so that the reviews may be more meaningfully analyzed.