An efficient soft computing technique for software defect prediction

Abstract

Now a days the software systems have become very complex and versatile. So that it is necessary to identify and solve the software defect is very important. The software defect prediction is the one of the most active research areas in software engineering. Therefore the software defect prediction plays an important role in improve the software quality. It also helps to reduced the time and cost for software testing. So, that it is used in many organizations to predict software defects in order to save time, improve quality, software testing and planning resources to meet the timelines. The software defect prediction is to predict the defects in historical data base. So, in real world, it is difficult to predict because it requires more number of data variables, metrics and historical data. As the size of software projects becomes larger, defect prediction techniques will play an important role to support developers as well as to speed up time to market with more reliable software products. In this research we introduce the soft computing techniques for predicting software flaws. Our proposed technique predicts the software defects and provides the accurate results in effective way. In our proposed method the defect database is extracted first it acts as the input. After that the extracted input (data) is clustered byclustering technique. For this purpose we use Modified fuzzy c means algorithm. Therefore, data are clustered. Then the clustered data is classified by effective classification algorithm. For this reason we use hybrid neural network. Therefore, the software defects are predicted and these predicted defects are optimized by using MCS algorithm. Our proposed method for software defect prediction is implemented in JAVA platform and the performance measures are measured by various parameters such as prediction rate and execution time. newline