Comparison between regression analysis and artificial neural network in project selection

A common problem faced by managers is that of project selection, to decide which project out of the lots should be undertaken. This paper aims at comparing results of the application of two approaches - respectively regression analysis, a parametric method and artificial neural network, a non-parametric technique. To demonstrate these methods, the models were illustrated using Oral, Kettani and Lang´s data on 37 R&D projects for their success. From statistical analysis, it was discovered that artificial neural network showed superiority to deciding how projects should be ranked and selected.