| 摘要 |
PROBLEM TO BE SOLVED: To provide a recommended item filtering method and a recommended item filtering program that recommend an item based on an estimated evaluation value weighted by an advanced user. SOLUTION: The recommended item filtering method comprises: a step that advanced user extracting means extracts an advanced user; a step that the extracted advanced user is weighted; a step that evaluation matrix between a user and an item creating means creates an evaluation matrix between a user and an item; a step that evaluation matrix between a user and an item weighting means weights an evaluation value of the evaluation matrix between a user and an item; a step that user similarity calculating means calculates the similarity between a sample user and a recommended user; a step that item evaluation estimated value calculating means calculates an estimated evaluation value of each item to a recommended user; and a step that recommended item replying means replies to the recommended user that an item having a high estimated evaluation value is a recommended item. COPYRIGHT: (C)2012,JPO&INPIT |
