Software Fault Prediction Using Machine Learning Algorithms

Abstract

Software quality, development time, and cost can all be improved by finding and fixing bugs as soon as possible.Machine learning (ML) has been widely used for software failure prediction (SFP), but there is a wide range in how well different ML algorithms predict SFP failures.The impressive results that deep learning can produce are useful in many different fields of study, including computer vision, natural language processing, speech recognition, and many others.This investigation into Multi-Layer Perceptrons (MLPs) and Convolutional Neural Networks seeks to address the factors that may affect the performance of both methods (CNNs).The earlier software errors are found and fixed, the less time, money, and energy are wasted and the higher the likelihood of success and customer satisfaction.While machine learning (ML) and deep learning (DL) have been widely applied to SFP, the results that different algorithms produce can be somewhat inconsistent.This research uses ANN-MLP-based boosting models like XGBoost and CatBoost to enhance accuracy on NASA datasets (Artificial Neural Network-Multi Layer Perceptron).We will use a voting ensemble consisting of ANN-MLP and booster models such as XGBoost and CatBoost to increase precision.