Software Quality Prediction Using Machine Learning

Abstract

Since twenty-first century, software quality is considered as a vital factor in the global competitive position for any software product in order to ensure quality and to ensure the dependency of software product. Software fault proneness has made tremendous progress in predicting the faults. On the other hand, the prediction model helps to create an accurate model and also helps the developers to deliver the software in time to their customers [4]. A software quality prediction model seeks to predict the quality factor that whether the software is prone to fault or not. In the early stage of software development, the users use the fault prediction model to detect the faults. So, our main aim should be to create reliable, portable, and robust software that minimizes the errors that occur when a program runs. Software quality prediction model helps us to know which components are at fault which can be corrected by detailed testing. To improve the quality, reliability, efficiency, and maintenance cost, the software fault should be predicted beforehand. It is difficult to develop fault prediction software. The cost of detecting and correcting the errors becomes extremely higher as we move from requirement analysis to maintenance phase, where defects might even lead to loss of life [3]. Fault detection in the beginning stages help the stakeholders [5] to converge their resources on modules that are likely to cause fault in the beginning phase.