SQA Defect Prediction: An SVM Based In-Appendage Software Log Analysis

Software Quality Assurance is an important factor in IT industry.The work reported in this thesis arose as part of an idea whose goal is to develop an adaptive SQA by defect prediction.In this regard we use SVM machine learning approach to predict the degree of fault proneness of software modules. The machine learning technique for defect forecasting and handling SQA called appendage log training and analysis can be referred as ALTA.The proposed defect forecasting of in-appendage software development log works is to deal the forecasted defects accurately and spontaneously while developing the software.In defect prediction process we opt machine learning technique called least square support vector machines in short LSSVM. The defect prediction stage of the ALTA targets the development logs available as input to train the LSSVM for better predictions.The future extraction process that is part of SVM training Process can be done with support of mathematical model called Intensified worst particle based Quantum Particle Swarm Optimization(QPSO).The QPSO algorithm and LSSVM,works as an intelligent system to predict defects to improve the software quality.