A bug prediction method for mobile app versioning using Gradient Descent Algorithm

Abstract

Abstract Mobile software (apps) is increasing day by day and human being is mostly dependent on these apps. New methodologies, tools, and technologies have continuously made the lifecycle for the mobile apps development more technology dependent. Very few research in the field of mobile app bug targets on bug evaluation, bug recognition, and bug prediction. Prediction of a bug in software field is almost contemporary research area that includes adopting numerous approaches such as fuzzy logic, artificial intelligence, data mining etc. Nevertheless, bug prediction of mobile apps is the very latest area of research and existing literature shows that there are very superficial work has done in this area. Bug detection and correction is complex phenomena for mobile apps. There are various advantages to measuring the mobile app such as accuracy of estimation, mobile app cost to boost the quality of apps. The purpose of this study is to provide a model that predicts the number of bugs in a future version of a mobile app compared to an earlier version of the app, and we used a gradient descent algorithm to accomplish this task. The latest version of the app may have advance feature, new content, modified design etc. The aim of our study is to anticipate the new bugs recognized in the latest version of the app by evaluating the category of modifications in an objective and recognizing various types of bugs. Mobile app developer and project manager both can get help from bug predictor’s model. The bug predictor model helps to improve the quality codes and minimize the test time.