Implementation of Data Mining Techniques for Software Development Effort Estimation
Deepti Gupta, Sushma Malik · 2020
Software development process now comprises a number of phases such as collecting requirements, designing, developing, testing, implementing and maintaining the software product. Each phase has its own implications and is interdependent. Each phase of the software development process generates a wide variety of data. Numerous types of data are generated during the software development process, including text, graphs, facts and figures. The increased data availability allows us to implement the data mining (DM) analysis techniques such as association, classification, clustering and many more to dig out meaningful information from complex data. The extracted information helps in optimizing and estimating the cost and efforts of software development process. The functioning of software engineering phases is improved by using the DM techniques. In this chapter, we illustrate the various data sources in software engineering and discuss the various techniques of DM that are implemented on data, especially to extract effort estimation in software development. DM techniques considered for this study include clustered (K-mean and K-nearest neighbors), regression (multivariate adaptive regression splines, ordinary least square and classification and regression tree) and classification (support vector machine and case-based reasoning).