ITEM RECOMMENDATION SYSTEM, METHOD, AND PROGRAM

PROBLEM TO BE SOLVED: To provide an item recommendation system, method and program capable of decreasing the number of users for which recommendation lists are created and reducing necessary capacity of a storage.SOLUTION: An item recommendation system 1 comprises a purchase information acquisition means 10 to acquire purchase information of each user and item, a user feature amount calculation means 20 to use each purchase information to calculate a user feature amount of each user, a representative user determination means 30 to differentiate a predetermined number of representative users on the basis of a representative determination value calculated by using the user feature mount and store the representative users and the representative determination values in a representative user storage means 77, a similar representative user detection menas 40 to calculate the representative determination value of a user and detect a representative user from among those in the representative user storage means 77 who has the most similar representative determination value as a representative user for the user, and a recommendation list creation means 50 to calculate a recommendation score of each item in the purchase information for the representative user and select a recommendation item to be recommended to the representative user on the basis of the recommendation score to create a recommendation list.