Thesis:

Predicting software defects using Ensemble Classifier and feature selection algorithm

Automated estimation of software from the point of view of error is very important in the software engineering community, and researchers using damped data constantly focus on accurate and reliable forecasts. In order to deliver quality software on time, software project managers, quality managers And software developers constantly need to monitor, identify and correct software flaws at all stages of the development process. If the software defects are detected in a timely manner, the quality of the software increases and reduces software maintenance costs. As a result, the use of machine learning, which is a branch of artificial intelligence, is trying to diagnose software earlier, in order to increase software quality. In this thesis, we compared the non-dimensional classification model with the applied dimensionality model. We also used SVM, Naïve Bayes, Decision Tree methods and analyzed the results with Relief filters. Experiments have been conducted to identify faulty software units, and the results have shown that dimensionality methods always influence the detection of faulty units.

Keywords: Keywords: Error, Software Disadvantages, Classification, Dimension Reduction.