Group Recommendation Algorithm Of Symbolic Data

With the rapid development of e-commerce and the rapid popularization of Internet, recommend system has become an indispensable information service platform for the Internet and e-commerce company now. Compared with the individual recommendation algorithm which has been widely studied, the study of group recommendation algorithm is still not very deeply. Symbolic Data Analysis (SDA) is a method analyzing and gleaning useful information and knowledge from massive data, and SDA method can grasp the data characteristics at whole aspect. Based on the SDA, the paper does some research on the group recommendation algorithm.The symbolic data (distributed data and interval data) has been applied to the research of group recommendation algorithm in this paper. First, based on the distributed data, the paper do some research on the item-based collaborative filtering algorithm (item-based CF), group user list was constructed based on distributed data, and item-based collaborative filtering model was established to get similar items and recommended results. In the following, based on descriptive statistics of uniform and general interval symbolic data, elaborate a new interval data distance which calledμσdistance and K-means clustering. And last, the real application is considered about the recommend quality and application fields.The results show that the group recommendation algorithm based on SDA can effectively improve the recommend quality compared with the traditional analysis method, and also can grasp the group characteristics at whole aspect.