Research On Software Defects Predict Methods Based On PCA

Number of faults is a key indicator of software quality. It's also a hotspot of software engineering research. To decrease faults' number and improve software quality are always the purpose of any software development organization. Testing is a common way to reduce software defects today. But it is costive and only available during the procedure of programming or after the executable product is available, which is too late for development organizations to remove all faults when testing is completed.Under the limited resources, in order to reduce the cost and improve the products' quality more effectively, the research of software defect prediction in software engineering was beginning. Hypothesis of software fault prediction is that complex software modules are more likely to have faults than simple one. Statistical metrics can represent the complexity of software modules. So it is possible to predict faults in software modules or whether a module is fault-prone based on statistical metrics that are easy to get. With software fault prediction, the defects will be found and cleaned more efficiently, and the quality as well as reliability of software will be increased obviously.Now software defects prediction is one of proven technology to improve software quality and reliability. And the number of research in this area is increasing quickly.Besides deep research on the correlative conceptions of software defects and software defects prediction, this paper also analysed and compared the software defects prediction methods which are popular used today. Based on these work, this paper proposed to employ Principle Component Analysis method in software defects prediction. PCA which is one of important methods in statistics can be used to predict the general capability of software systems. Otherwise, in allusion to the fault of PCA in the condition of multiple variable, the paper proposed a improved PCA to predict software defects.On the bases of the work above, the paper made use of the data from a middle software company to predict software defects with PCA, and the predicting results were validated by two classical models. Results of these experiments proved that software predict method based on PCA did competent job on predicting the general capability of software systems.