Software quality prediction techniques: A comparative analysis

There are many software quality prediction techniques available in literature to predict software quality. However, literature lacks a comprehensive study to evaluate and compare various prediction methodologies so that quality professionals may select an appropriate predictor. To find a technique which performs better in general is an undecidable problem because behavior of a predictor also depends on many other specific factors like problem domain, nature of dataset, uncertainty in the available data etc. We have conducted an empirical survey of various software quality prediction techniques and compared their performance in terms of various evaluation metrics. In this paper, we have presented comparison of 30 techniques on two standard datasets.