Mining Competences of Expert Estimators

This paper reports on a study conducted with intention to identify competences of employees engaged on software development projects that are responsible for reliable effort estimation. Execution of assigned project tasks engages different human characteristics and effort estimation is integral part of development process. Competences are defined as knowledge, skills and abilities required to perform job assignments. As input data we used company internal classification and collection of employee competences together with data sets of task effort estimates from ten projects executed in a department of the company specialized for development of IT solutions in telecom domain. Techniques used for modeling are proven data mining methods, the neural network and decision tree algorithms. Results provided mapping of competences to effort estimates and represent valuable knowledge discovery that can be used in practice for selection and evaluation of expert effort estimators.