PCA based cost estimation model for agile software development projects

Agile software development has been attached much importance as a new software engineering methodology as it emphasizes on good communication between the developers, the rapid delivery of software, and change on demand. The metrics and methods from conventional lifecycle models cannot predict accurately for agile projects. We have proposed a new cost estimation model for agile software development projects. In this model we apply Principal Component Analysis to reduce the dimensions of the attributes required and identify the key attributes which have maximum correlation to the development cost; and then use constraint solving approach to satisfy the criteria imposed by agile manifesto. The proposed methodology is found to bet suitable for agile projects as it uses constraint programming to explicitly check for satisfaction of agile manifestos. On comparison with other approaches under research we find that our model provides a low MMRE value i.e.50.63. Our methodology can also be used in case of unavailability of historical data or expert opinion. Hence we can safely say that the proposed cost estimation approach increases the precision and accuracy of estimates; and hence is better suited for the Agile Software Development Projects.