A Retail Outlet Classification Model Based on AdaBoost

Abstract

This paper proposes a framework to get a stable classification rule under unsupervised learning, and the term ‘‘stable’’ means that the rule remains unchanged when the sample set increases. This framework initially makes use of clustering analysis and then use the result of clustering analysis as a referencestudying sample. Secondly, AdaBoost integrated several classification methods is used to classify the samples and get a stable classification rule. To prove the method feasible, this paper shows an empirical study of classifying retail outlets of a tobacco market in a city of China. In this practice, k-means is used to make clustering analysis, and AdaBoost integrated RBF neural network, CART, and SVM is used in classification. In the empirical study, this method successfully divides retail outlets into different classes based on the sales ability.This work has been partially supported by the National University Student Innovation Program of China and the National Natural Science Foundation of China (Grant No. 61272003)

Implementing Enhanced AdaBoost Algorithm for Sales Classification and Prediction

“…AdaBoost integrated several classification methods is used to classify the samples and get a stable classification rule. An empirical study of classifying retail outlets of a tobacco market in a city in China is done in order to prove the method is feasible [7].…”