Integrating CBR and ID3 data mining tools for predicting supermarket sales

Abstract

Business organizations generate and collect a large amount of data in their daily operations. However, despite this wealth of data, many companies have failed to fully capitalize on its value because information implicit in the data is not easy to discern. In this paper, sales in the supermarket is predicted by integrating two data mining techniques such as the CBR (Case Based Reasoning) and the ID3 induction method. POS (Point of Sale) data from a real-life large retailer are analyzed. The data has chaotic and noisy time series patterns. The results of the prediction imply that the integrated method provides better prediction performance and is easy to use.