Software Fault Prediction: a Review

Software defect prediction in software engineering is one of the most interesting research fields. To improve the quality and reliability of the software in less time and in minimum cost, it is the most relevant key area where various researchers have been done. When the size and complexity of software increases then faults prediction in the software became more difficult. To maintain the high level of quality of the software, there is need of a model or system which can classify the software in two prone modules as faulty and non-faulty prone. In the process of predicting faulty and non-faulty prone, the prediction of faulty prone modules incurs more cost and time than prediction of non-faulty prone modules. In this literature survey, the process of defect prediction is analyzed. There are different techniques to predict faults and evaluate performance of the predictors. These predictors may be a model, system, techniques or algorithm. In this review it is studied that what type of progress has been done up to now and which type of software metrics have been used to design the fault predictor. It is also discuss about the cross project fault prediction which are more demanding in today’s scenario. Keywords– Software Bug prediction, Public Dataset, Neural Network