Comparison of conventional approaches and soft-computing approaches for software quality prediction
暂无分享,去创建一个
Managing software development and maintenance projects requires early knowledge about quality and effort needed for achieving a necessary quality level. Quality prediction models can identify outlying software components that might cause potential quality problems. Quality prediction is based on experience with similar predecessor projects constructing a relationship between the output-usually the number of errors-and some kind of input-here we use complexity metrics-to the quality of a software development project. Two approaches are presented to build quality prediction models: multilinear discriminant analysis as one example for conventional approaches and fuzzy expert-systems generated by genetic algorithms. Using the capability of genetic algorithms, the fuzzy rules can be automatically generated from example data to reduce the cost and improve the accuracy. The generated quality model-with respect to changes-provides both quality of fit (according to past data) and predictive accuracy (according to ongoing projects). The comparison of the approaches gives an answer on the effectiveness and the efficiency of a soft-computing approach.
[1] Lesley M. Pickard,et al. Statistical techniques for modelling software quality in the ESPIRIT REQUEST project , 1987 .
[2] Victor R. Basili,et al. The TAME Project: Towards Improvement-Oriented Software Environments , 1988, IEEE Trans. Software Eng..
[3] Christof Ebert,et al. Rule-based fuzzy classification for software quality control , 1994 .