A study of recommendation considering cancellation — Case study of golf EC website

Recently Electronic Commerce (EC) have attracted much attention and widely popular. Whence recommendations which understand user's feelings and sort out the goods the user desires attract attention. On the other hand, cancellation behavior can be large risk at EC companies which handles use of service at facilities such as hotel rooms and golf courses as products. We try analyzing to decrease the canceling ratio caused by the inner factors which are come from customer's mind. We recommend item which understands customer's true needs by analyzing reservation and data of behavior log including actual cancellation. We also inspected the utility of the cancellation data and the constrain conditions for model building using a Bayesian network. We show that it is possible to contribute to accuracy improvement of a recommendation for the cancellation data.