Intelligent and Soft Computing Systems for Green Energy

9Sales Demand Forecasting for Retail Marketing Using XGBoost Algorithm

M. Kavitha¹*, R. Srinivasan¹, R. Kavitha¹ and M. Suganthy²

¹Vel Tech Rangarjan Dr. Sagunthala R&D Institute of Science and Technology, Chennai, India

²Vel Tech Multi Tech Dr. Rangarajan Dr. Sakunthala Engineering College, Chennai, India

Abstract

Demand forecasting helps in estimating the requirements of products in the future, based on past and present data. The future depends on trends and other aspects of the market. Nowadays, organizations are not able to predict future product demand, which leads to unnecessary storage and decay of products. So by analyzing the past and current market data, future demands are well predicted; thus those demanded products can be manufactured in the near future. This paper is mainly focused on predicting the demand of any particular goods and services that are available at a particular time of the year to satisfy consumer needs. Demand forecasting can also be helpful to both the consumer and the retailer; the consumer satisfies his needs and therefore, the retailer gets profits. This paper mainly presents the forecasting of sales of data by using Time Series analysis and XGboost algorithm, where we can see the trend and recession of sales in different times of the year. The accuracy reached is “0.14” by reducing overfitting and underfitting. Here, Root Mean Square Percentage Error Value is used; the accuracy is calculated in the range of 0.1 to 0.9. Value closer to 0.1 ...