Prediction of Success or Failure of Software Projects based on Reusability Metrics using Support Vector Machine

Abstract

In the field of computer science & engineering and software industry the term reusability means usage of existing software assets or previously developed code in the software development process. The assets of the software are products and by-products of the product development life cycle which includes code, test cases, software designs and code documentation. The process of modifying the existing assets as per the need and specific requirements is called leveraging. But the reusability process creates a new version of the existing assets. So always reusability is preferred rather than leveraging. To identify the quality of the reusability of the software components various software metrics are available. But the framework or model that can predict the reusability of the software assets are needed to be developed. The reusability metrics must be identified during the design or coding phase and that can be used to reduce the rework needed develop a similar software module. This can much improve the productivity due to the probabilistic increase in the reuse level. In this study various software metrics representing the software reusability nature of the software components are collected in relation with a particular software project to form a database. The database is divided in to training and test set and Support Vector Machine is trained using the Radial Basis Function (RBF) to predict whether the software component can be reused or not.