Feature Failure Prediction in Agile Development
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In agile software development, existing feature analysis tools cannot identify internal feature iterations failures in real-time, leading to inadequate testing and increased costs due to the inability to monitor features during development.
Innovation Solution
A method for predicting failure rates of system features by defining feature states during development stages, assigning defect class values, and creating a training set using machine learning to classify features as 'OK' or 'BAD' based on defect reports, enabling proactive testing and reducing testing backlogs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional feature analysis tools are used to monitor completed projects, then defect classification can be performed after project completion, but the tools cannot identify internal feature iterations failures during development
Solution Approach 1:
The patent creates training sets and builds classification models during the development process itself, rather than waiting for project completion. By preliminarily establishing the defect classification framework and continuously training it with emerging data from ongoing iterations, the system can detect defects in real-time during development, transforming the traditional post-completion analysis into a proactive, continuous monitoring process that identifies failures as they occur in internal feature iterations
Solution Approach 2:
The patent implements a feedback mechanism where defect data from ongoing development iterations is continuously fed back into the training set, which then retrains the classification model. This creates a closed-loop system where the model learns from actual defect patterns emerging during development and improves its predictive accuracy over time, enabling reliable defect detection during the development process rather than only after completion
2Reliability
If continuous testing is implemented during agile development iterations, then internal feature iterations failures can be identified, but testing resources and complexity increase
Solution Approach 1:
The patent enables the system to automatically create training sets, train classification models, and predict defect risks without requiring complex manual testing infrastructure. The model serves itself by continuously learning from defect data and automatically updating its predictions, reducing the need for extensive manual testing resources and simplifying the overall testing system architecture while maintaining high reliability in defect detection
Solution Approach 2:
The patent replaces traditional mechanical testing processes with an intelligent classification model that uses machine learning to predict defect risks. Instead of relying on extensive manual testing and complex testing frameworks, the system uses computational algorithms to analyze feature data and predict potential failures, substituting physical/testing complexity with intelligent computation that scales more efficiently
3Productivity
If traditional defect classification is used with binary OK/BAD labels, then simple classification is achieved, but the system cannot predict failure rates or prioritize testing efforts
Solution Approach 1:
The patent enhances the basic OK/BAD classification by introducing probability predictions and failure rate estimates for each classification. Instead of uniform treatment of all defects, the system provides localized quality assessment through probability scores that indicate the confidence and risk level of each prediction, allowing testing resources to be prioritized based on actual risk levels rather than treating all defects equally
Data Source
AI summary
The disclosed herein relates to a method for failure rate prediction of a feature of a system under development. The method is executed by a processor coupled to a memory. The method includes defining a feature state of the feature during a predetermined time interval, the predetermined time interval being associated with a development stage of the system. The method also includes assigning a first defect class value to the feature for the predetermined time interval, the first defect class value configured to indicate a first condition and selecting, when a defect is reported for the feature, a second defect class value indicating a second condition, the second condition being associated with a higher failure rate than the first condition. The method can be embodied in system and a computer program product.