INTERNET SYSTEM WHICH RECEIVES ITEM THROUGH YES/NO EVALUATION – BASED ON STATISTICAL SIMILARITY FOR USER’S EVALUATIONS

摘要

The present invention relates to a system that, if a user of an Internet website evaluates plural items with a yes/no answer, searches for persons who match the corresponding user’s taste most properly and recommends an item to which the searched persons prefer based on the evaluation. The item recommendation system according to the invention includes: a DB that records “a corresponding user ID”, “an item title,” and “a yes/no evaluation value (for the item)” if the user evaluates the item with a yes/no answer; a sympathy probability function that calculates the similarity of the yes/no evaluation between users from the DB; a user list that automatically arranges the users who have the most similar taste according to the sympathy probability; a recommendation item chart that automatically arranges the items that the users with similar taste like according to preference; a sympathy user search function that searches for the users whose evaluation of a specific item matches the user’s evaluation from the DB, and a recommendation item chart arranges the items that the searched users like.