Model Performance Evaluation: Sales Prediction

Abstract

AbstractThese days, every organization wants to uplift its revenue as well as profit. There are various departments in an organization including the Production department, IT department, Sales & Marketing department and many more. The work is done in hierarchal order but at the end of the day, one of the main objectives of any organization is to increase sales which will lead to increase in profit. For consistent performance in market, the people sitting there at upper level need to take various decisions for the organization’s well being. They need to tell the people of production and sales department about the quantity and quality of product that they need to produce within a given time window. Various factors including the location of their stores, population density, direct competitors, etc. play a vital role in prediction of sales in upcoming times. According to a broad study, organizations doing sales predictions correctly are approximately 10% more likely to enhance their revenue growth year-by-year. In this paper, we have used one such data of a very popular store. The major focus is on hypothesis formation, feature engineering and applying various Machine learning Algorithms. At last we have compared the results of all the different algorithms used on the basis of RMSE evaluation metrics and finally predicted the sales as well as the major factors that play significant role in their sales upliftment.KeywordsMachine learningRMSEHypothesisFeature engineeringRevenue