ML-Based Negative Review Analysis for Hidden Software Issues
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing software technologies fail to detect and address hidden software issues that are intermittent, difficult to observe by providers, or not detected by automated health monitoring systems, leading to negative user experiences and frustration.
Innovation Solution
A method utilizing machine learning algorithms to analyze natural language text from various sources, identifying negative reviews, categorizing them, and recognizing software application names to pinpoint hidden issues, allowing for adjustments to be made to improve user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If automated software health tracking systems are used to monitor software operation, then software functionality can be detected, but hidden issues that affect user experience cannot be identified
Solution Approach 1:
The patent introduces natural language text from external sources as an intermediary carrier that bridges the gap between automated monitoring systems and user experience feedback. This intermediary contains information about user interactions and frustrations that automated systems cannot directly observe, allowing the system to detect hidden software issues while maintaining functional monitoring capabilities
Solution Approach 2:
The patent replaces traditional automated health tracking mechanisms with a machine learning-based natural language analysis system. Instead of relying solely on technical metrics from automated monitors, the system uses MLMs to process and interpret user-generated text, enabling detection of subjective user experience issues that mechanical monitoring systems cannot capture
2Measurement precision
If machine learning models process all natural language text, then comprehensive issue detection is achieved, but processing time and computational resources increase
Solution Approach 1:
The patent segments the natural language text processing task into distinct stages: pre-processing and cleaning by a first MLM, negative review extraction by a second MLM, and categorization by a third MLM. This segmentation allows each model to focus on specific aspects of the text, improving detection accuracy while enabling parallel processing that reduces overall processing time
Solution Approach 2:
The patent applies partial action by selectively processing only the most relevant portions of natural language text. The system identifies and focuses on negative reviews and sentiment-bearing text segments, rather than uniformly processing all input text. This approach maintains high detection accuracy for critical issues while reducing computational overhead on less relevant content
Data Source
AI summary
A method including preprocessing natural language text by cleaning and vectorizing the natural language text. A first machine learning model (MLM) extracts negative reviews. A first input to the first MLM is the natural language text and a first output of the first MLM is first probabilities that the negative reviews have negative sentiments. The method also includes categorizing the negative reviews by executing a second MLM. A second input to the second MLM is the negative reviews. A second output of the second MLM is second probabilities that the negative reviews are assigned to categories. The method also includes identifying, using a name recognition controller and based on categorizing, a name of a software application in the negative reviews and sorting the negative reviews into a subset of negative reviews relating to the name. The software application is adjusted based on the subset of negative reviews.