Demand Forecasting in Retail Industry for Liquor Consumption using LSTM

This paper focuses on forecasting demand of sales stock. Retailers face problems of stock out or over stock due to improper management or estimation skills. So, to help retailers overcome these loss causing problems, our system predicts the stock requirements based on the previous data. This proposed work has implemented various algorithms like Long Short-Term Memory (LSTM), Regressions and Support Vector Machine (SVM) to predict sales. Various methodologies are used for demand forecasting and an appropriate one is to be selected which is capable of providing a high estimation accuracy. Our experimental results on real data set suggests to follow LSTM as it considerably improves the forecasting performance.