Implementing Enhanced AdaBoost Algorithm for Sales Classification and Prediction

Abstract

In today's data driven economy, retail businesses rely on information systems that monitor and process their daily transactions. These huge amount of data being processed on a day-today basis can be utilized to forecast sales for inventory management, and decision-making. In this paper the AdaBoost algorithm is used in classification and prediction of data. While it is known to be capable of processing both variable and numerical values, it is quite certain that processing data, represented as facts, is faster in digital form. This allows the algorithm to process the conditions digitally. The original raw facts presented in this study are in variable forms. To better improve performance, the first part of the algorithm converts the facts and represent them in numerical, computable values. This allows the rest of the algorithm to process the entire set of data numerically, and evidently faster, resulting to a better performance of the algorithm. The use of this innovative technique improves the performance of the extraction methods used in data mining which is very important for business support and decision making. The future of this research can explore the development of a Decision Support System (DSS) for the purpose of predictive analysis and support of business decisions.