Interactive design on recommender system

With the arrival of the information age, the increasing volume of data, how to find effective information from the complex data to becomes an important part of everyone's daily work and life. This issue has promoted the birth and development of personalized recommendation system. At the same time, visual analysis technology based on large data presentation and analysis has become a hotspot. With the increasing sophistication of visualization technology, complex multidimensional data can also be showed well, so as to help people understand the nature of the data and useful information. There are three problems in the traditional recommendation system because of the lack of attention to the system interaction: one is the user preference guidance problem in the cold start of the recommender system; the second is the black box and the unexplainable problem in the recommendation process; the third is the lack of recommendation result Presentation and feedback mechanism of user interaction. In this paper, the above three issues, combined with the characteristics of visual analysis technology has been proposed. Firstly, an improved attribute-oriented algorithm based on attribute is proposed to visualize the media project data such as movies. Secondly, the explanatory visualization and interactive feedback scheme of collaborative filtering algorithm are studied. Finally, on the basis of the previous research, the general framework of recommendation visualization system and the UI design model based on project visualization are put forward. A film recommendation visualization prototype system is designed and implemented as the experimental basis of the visual interactive research program.