ID3 has been proposed as a way of speeding up the acquisition of knowledge for expert systems. To segment the medium-large level computer market, the original ID3 algorithm must be modified since ID3 is not very robust in the face of noisy data. A modified algorithm of ID3 is proposed in this study and it contains prunning measure based on entropy. Also a measure of gained information after generating rule tree is proposed. The modified algorithm is applied to computer market. A method of market segmentation using modified algirithm is suggested.
A obtained Rule-tree classify the competitive products and market. To develop the marketing strategy, a table transforming Rule-tree is used.