Efficient prediction of new product growth rate: A comparison between Artificial Neural Network and conventional method

To achieve business growth of a company or organization it is needed to forecast growth rate of the new products. Most of the existing models for forecasting are linear in nature. As a result these models fail to predict correctly when the input attributes are nonlinear. To overcome this limitation this study develops a Functional Link Artificial Neural Network (FLANN) and Non-liner Regression (NLR) for identifying new product growth rate for three different industries from manufacturing sectors such as Electronic, Construction, Auto and Auto component. The desire nonlinearity to the FLANN model is introduced by trigonometric expansion of the input features. Further, comparison of performance between the two proposed models reveals that the FLANN model is preferable to the NLR model.