摘要 |
PROBLEM TO BE SOLVED: To predict purchase behavior of each customer with a high degree of accuracy.SOLUTION: A purchase behavior prediction model learning part 21 estimates: a probability distribution Pinnerof an internal reference price of a customer n; a probability distribution Pbof a price range of a product k; and a customer's interest-based probability Q(k|n) of purchase of the product k by the customer n. A purchase behavior prediction part 22 calculates a score Score(k|n) showing a possibility of purchase of the product k by a customer nto be predicted, with respect to each product, on the basis of: a product-price-based customer purchase prediction value sim(n,k) that is calculated from a probability distribution Pinnercorresponding to the customer nand the probability distribution Pb; and the probability Q(k|n). |