Research On Collaborative Filtering Recommendation Algorithm Based On User Reference Price Tendency

"Information overload" problem with the Internet growing, the traditional search engine for the user personalized information retrieval is more and more difficult to meet demand. At present, in order to solve this problem, recommender systems are being widely used in many field. Collaborative filtering recommendation is a widely and successful used recommended strategy.The basic data of the current collaborative filtering recommendation investigation mainly used for scores of items, based on the evaluation on the basis of similarity for users or items. People shopping in reality, however, depends not only on the evaluation of goods, more need based on factors such as consumer oneself condition and living habits, some users tend to buy more expensive goods, while other users tend to buy cheaper goods, each user has its own reference price of goods, the reference price to users tend to have great influence on the final purchase decisions. In general, the reference price tendency degree level is similar to the user in the final choice when buying items on the prices of the goods tends to more similar, therefore in the recommended decision factors taken into account in the process of commodity prices is necessary, while referring users to the most users of consumer psychology price products recommended to the user, can partly improve the accuracy of recommendation. At present, few from the user for reference price of items from the Angle of inclination degree of filtering recommendation strategy research.Based on the electric field characteristics, according to the above problem, was proposed based on the user reference price tendency degree study of the collaborative filtering recommendation algorithm. In this paper, we first introduced this topic research background, the research significance of topic. And related literatures, this paper introduces the electrical contractor recommended by the domestic and foreign research present situation; Secondly, the paper introduces the subject research involves the related technology and theory, including consumer psychology, consumer reference price, the price level. On this basis, this paper proposes a collaborative filtering recommendation based on user reference price tendency degree is proposed. Buy the idea according to the user in the history of the combination of commodity price range, form the user for reference price tendency degree matrix of items, from the perspective of the user reference price tendency degree seeking similarity between users, and according to the level of prices with the user for the price of the goods tend to be weighted score predicts. At the same time, the paper also puts forward the comprehensive weighted based on user reference price tendency degree and the traditional recommendation algorithm based on similarity user ratings.Through this topic research has shown that, based on the user reference price tendency of research and improvement of the collaborative filtering recommendation algorithm can more accurate reflection of consumer value tendency and practical improved the precision of prediction recommended, recommended for electrical contractor recommendation system of this mechanism is of great significance.