Sales Forecasting Using Prediction Analytics Algorithm

Sales prediction plays a significant role in various firms that are engaged in merchandising, logistics, production and sales. This enables the users to allot resources efficiently for the estimation of sale revenue and set up an improved strategy for future growth of the firm. Machine-learning helps to find data accuracy in sales prediction. In this paper, forecasting the sales of a product is analyzed by using linear regression and K-means cluster approach that produces higher prognostic performance compared to the present predictive learning algorithms. A novel approach known as predictive analytic approach has been proposed. This approach is used to cluster the dataset based on consumable and non-consumable items. It selects the medium, high and low levels (central points) from the collected dataset. The performance of the proposed approach has been compared with linear regression and k-means cluster approaches. The experimental results prove that the proposed approach outperforms all other existing approaches.