| Estimation of effort and costs is crucial for successfully implementation of software projects. Project development time is an essential factor, both for project clients and project developers. The amount of money needed to invest in a project influences the decision whether to start a project or not or whether it will be completed successfully or not. In practice, the cost of a project is most often compared to the cost of similar projects, which have been successfully completed. The article proposes combining the experimental information of the Taguchi method with ANN (Artificial Neural Network) learning, and constructs a progressive Taguchi neural network model to decrease the number of experimental runs and time required. The main goal of this paper is to make generalization of the conditions and criteria for successful project realization based on a large number of experimental results that authors obtained so far. Furthermore, the most reliable and efficient artificial intelligence model for effort and cost estimation is identified. |
*** Title, author list and abstract as seen in the Camera-Ready version of the paper that was provided to Conference Committee. Small changes that may have occurred during processing by Springer may not appear in this window.