Analogy-Based Software Development Effort Estimation in Global Software Development

Abstract

Context: Software development has always been characterised by certain parameters. In the case of global software development, one of the important challenges for software developers is that of predicting the development effort of a software system on the basis of developer details, size, complexity, and other measures. Objective: The main research topics related to global software development effort estimation are the definition and empirical evaluation of a search-based approach with which to build new estimation models and the definition and empirical evaluation of all available early data. Datasets have been used as a basis to carry out an analogy-based estimation using similarity functions and measures. Method: Many of the problems concerning the existing effort estimation challenges can be solved by creating an analogy. This paper describes an enhanced analogy-based model for the estimation of software development effort and proposes a new approach using similarity functions and measures for software effort estimation. Result: A new approach for analogy-based reasoning with which to enhance the performance of cost estimation in distributed or combined software projects dealing with numerical and categorical data. The proposed method will be validated empirically using The International Software Benchmarking Standards Group dataset as a basis. Conclusion: The proposed estimation model could be a useful approach for early stage effort estimation on distributed projects.