A Multilayer Collaborative Filtering Recommendation Method in Electricity Market
Li Tao, Minghai Jiao, Yong Dai, Chuncheng Gao · 2016
Transaction price accurate recommendation is a hot issue for buyer and seller on bidding information services in Chinese electricity market, a novel multilayer collaborative filtering algorithm is proposed to solve the bidding prices accurate mining problem. A three-tier relationship model of user-item-attribute is described to accommodate the real electricity transaction on bidding price service mining and recommendation. The fuzzy evaluation method is presented to improve candidate items sets of similar neighbors, integrating user fuzzy preference by attributes into membership degree evaluation. And then, the similarity function is improved to determine better proportions in items Pearson coefficients. A case study is done to give an application example for electricity market. And the experiment is also implemented to prove that the model and the algorithm are efficient and robust for application value by performance of experiment results.