A Neighbor-Aware Group Recommendation Algorithm

With the rapid development of e-commerce, how to recommend the groups that are most likely to buy a certain kind of items to the merchants accurately has become an increasing research of scholars. However, the existing group recommendation technology rarely considers the influence of the close relationship between users on user preferences. Thus we propose a group recommendation model NGRN to generate groups and make recommendations. First we extract k-core groups on the social network, the groups meet the conditions that each user has k neighbors at least. Then we get the recommendable probability of candidate groups under different items. At the same time, the validity of this method is verified on two public datasets. Experiment shows our model has a great improvement for recommendation accuracy compared with other models.