Back propagation network based prediction on map-reduce platforms

Abstract

In this paper, we provide a significant model illustrating the challenges to handle terabytes of data efficiently. Proposed paper aims to develop a model which predicts the success rate of a product before its in launch into the market. Artificial Neural Network using back propagation (BANN) has been implemented for product prediction. Huge volume of e-commerce customer reviews and rating dataset is used for testing this model. This nonlinear method uses Broyden-Fletcher-Goldforb-Shanno (BFGS) training on multiple network layer. Our method gives improved prediction accuracy. The results of this model is to introduce a sustainable and successful product into the market, which in turn helps to improve the quality of the product.