Method for Estimation of Software Requirements Using Neural Network Based Classification Technique

Effective management of software projects depends on ability to make accurate time-predictions. Nowadays, software companies need to deliver their solutions in expected time and budget. There are many factors influencing duration and cost of software projects. This paper provides innovative approach for estimations in early phase of software development. It shows usage of standard methods and its combination with soft computing technique called classification that is used for time-estimation of requirements using two-layer feed-forward neural network, which classifies requirements into time-groups.