In recent years, with the rapid development of Internet, massive informationare spreading quickly and miscellaneous goods come to people. It is a new need forpeople to find what they want from these information and goods. Therefore, therecomes recommendation system. Recommendation system offers information anditems according to people’s behavior custom and item’s properties. For example,news recommendation system recommends news information on the basis ofpeople’s interests and reading habits. E-commerce websites recommendation systemrecommends appropriate items to people after analyzing users’ buying preference.Music and video websites recommends songs and videos according to people’staste.And social networks recommends new friends for people.Nowadays, Recommendation System has been frequently applied into people’sreal-life. With improving users’ acceptance, recommendation system has broughthuge commercial interests. However, as the original data is incomplete and somerecommendation algorithms have their own special ways of processing data, currentrecommendation system sometimes cannot work very well. For example, somerecommendation systems are bothered with cold-start, complex interestrecommendations or bad implementation. There already have some methods onsolving these problems. However, they all have their restrictions. As far as I know,there haven’t effective methods for complex-interest recommendations.In the paper, we proposed a slider-tag based recommendation system model. Inthe model, we will get the precise information of users’ need by the simple way ofusing slider-tags. Then we will generate recommendations using modified currentrecommendation algorithm. At last, we will show the recommendations to the userswith implementation in the way of slider-tags. In addition, it is feasible for users togive their’ feedback to the system and get more accuracy recommendations. At lastwe conducted a series of comparative experiments, which proved the effectivenessand feasibility of the techniques in this paper. |