Application of Artificial Neural Network for Procedure and Object Oriented Software Effort Estimation

Abstract

Software effort estimation guides the bedding, planning, development and maintenance process of software product. Software development uses different paradigm like: procedure oriented, object oriented, Agile, Incremental, component based and web based etc. Different companies use different techniques for their software project development. The available estimation techniques are not suitable for all types of software development techniques. So there is a need of estimation technique that can be applied on all type of software. This paper we are evaluating the application of artificial neural networks in prediction of effort in conventional and Object Oriented Software development approach. We have used feed-forward neural network created using MATLAB10 ( NN tool kit) and applied on two different types of datasets, one for conventional software and another for object oriented software. The simulation results were studied and we found that artificial neural network model works very accurately on both types of software development techniques.