A Study on the use of Artificial Neural Network Approach to Software Cost Estimation

Enhancing Software Cost Estimation with Artificial Neural Networks

Abstract

Various cost estimating models are there, each of these models has advantages anddisadvantages when it comes to predicting the price tag and time involved in creating a product. Theestimate of software costs by use of Back-Propagation neural networks. The model is developed in away that accommodates and enhances the commonly used COCOMO model. As a result, software costestimates are more accurate, even when dealing with ambiguous or imprecise information. Threepublicly available datasets are used to test the model. In this study, the software cost estimatingexpertise is modeled using an artificial neural network technique, and the findings are compared withthe COCOMO model.