Software defect prediction using software metrics - A survey

Traditionally software metrics have been used to define the complexity of the program, to estimate programming time. Extensive research has also been carried out to predict the number of defects in a module using software metrics. If the metric values are to be used in mathematical equations designed to represent a model of the software process, metrics associated with a ratio scale may be preferred, since ratio scale data allow most mathematical operations to meaningfully apply. Work on the mechanics of implementing metrics programs. The goal of this research is to help developers identify defects based on existing software metrics using data mining techniques and thereby improve software quality which ultimately leads to reducing the software development cost in the development and maintenance phase. This research focuses in identifying defective modules and hence the scope of software that needs to be examined for defects can be prioritized. This allows the developer to run test cases in the predicted modules using test cases. The proposed methodology helps in identifying modules that require immediate attention and hence the reliability of the software can be improved faster as higher priority defects can be handled first. Our goal in this research focuses to improve the classification accuracy of the Data mining algorithm. To initiate this process we initially propose to evaluate the existing classification algorithms and based on its weakness we propose a novel Neural network algorithm with a degree of fuzziness in the hidden layer to improve the classification accuracy.