Product brand prediction in retail industry

Abstract

Nowadays, there are some challenges regarding to sales and retailing issues. There are some factors that impact to the retailer influences to their customer. It is including the competition from online businesses. Sometimes it is hard to achieve the consumer demands expectation. At the same time, the retailer may need to overcome the financing pressure and marketing challenges just to ensure that their brand could be survive in business matter. The retailing challenges that always facing by company including finding the financing to stay in business. The only ways to solve the problem is by making prediction to investigate their product sales by using the data in year 2015 and 2016. The objective is to study the existing data sales based on type of brand, to develop prediction model for each type of brand, and to evaluate the performance of Artificial Neural Network (ANN) and Exponential Smoothing (ES) model in selecting the best model. The measurement performance used to analyse the data are using RMSE, MAE and MAPE. The comparison for the best model is based on the actual and predicted data that helps to get the result for profit sales and loss sales for each brand. Result obtained shows that MLP model is better in prediction model compared to the ES model.