Experimental Investigation of Machine Learning Techniques for Predicting Software Quality

Abstract-There are several points in the software development process when estimating software quality is useful. Quality assurance planning and benchmarking are two potential applications. Multiple criterion linear programming and quadratic programming are two approaches that have been utilised in prior research to estimate software quality. In addition, we tested out C5.0, SVM, and a Neutral network to determine how best to estimate quality. The precision of these research is rather poor. The purpose of this research was to enhance estimate precision by using important properties of a large dataset. In order to improve accuracy, we used a feature selection technique and a correlation matrix. We have also tried our hands at several newer techniques that have proven effective in other prediction challenges. Xgboost, Random Forest, and Decision Tree are only few of the machine learning algorithms used to analyse the data and draw conclusions about the software's quality and its relationship to its development qualities. Results from experiments demonstrate that machine learning algorithms can accurately predict the quality of softwar