Neural networks application in software cost estimation: A case study

One of the topics during the last 30 years much research has been allocated to cost estimation for software projects. Important issues in the field of software engineering capabilities estimate size and effort required for development of software projects. Cost estimates must be made at the beginning of the project, and principally at the beginning of projects through cost and set new work requirements will be done. In this paper, we describe new case study for estimate software cost and will be presented new Neural Network (NN) architecture for prediction necessary effort for new software. The results indicate that the NN model that we offer the very best of algorithmic methods to predict and estimate software costs as smart deals. NN training results by NN indicating, that the test results indicate that over 90% of cases give much better estimates of NN algorithm model.