販売履歴を用いたTPO商品向け推薦システム Sales Records Based Recommender System for TPO- Goods
This paper presents a recommender system for TPO (Time, Place, and Occasion)-dependent goods. The TPO-dependent goods have three features: many attributes, multiformity, and high-frequency update. In order to recommend alternatives of the goods, our system (a) abstracts and metrizes the user's preference implied in sales records, and (b) filters massive alternatives by three kinds of methods: High-Angle Search, Low-Angle Search and Neighbor Search. Additionally, this paper describes the improvement method of the recommendation accuracy by memory-based reasoning with user's preference to latter two kinds of search. The numerical simulation for 10,000 user's data and 400,000 sales records has shown their accuracy.
- 電気学会論文誌. C, 電子・情報・システム部門誌 = The transactions of the Institute of Electrical Engineers of Japan. C, A publication of Electronics, Information and System Society
電気学会論文誌. C, 電子・情報・システム部門誌 = The transactions of the Institute of Electrical Engineers of Japan. C, A publication of Electronics, Information and System Society 126(5), 661-666, 2006-05-01
The Institute of Electrical Engineers of Japan