A Research On Location Based Collaborative Filtering Recommendation Method

With the development of the Internet,massive information makes it difficult for users to find their own content of interest,so the recommended system for information processing came into being,personalized recommendations for the majority of users has brought great convenience.The traditional collaborative filtering recommendation system mainly uses the user to score the information of the project,but less attention to the location information.With the rapid development of mobile Internet,it is difficult to meet the needs of users.Most of the location-based recommendation system research on the user location information has a strong dependence with the privacy of the user's attention,while the user's location information is generally not open,or open a little not accurate location information under users' agree,which makes the relevant research is difficult to apply to the actual.How to use the location information of the goods to provide users with more personalized,more in line with the user's habits of consumption recommendations,is the purpose of this study.This paper studies the recommendation method of location-based collaborative filtering service for data with no user location information,and only with commodity location information.This paper proposes a framework of collaborative filtering based on location.The framework consists of three modules.It adds product location information on the basis of traditional collaborative filtering,improves the accuracy of recommendation results and alleviates data sparseness.First of all,according to the location information of goods on the product location hierarchical clustering,divided them into different goods area with different levels,to simulate the real geographical area of ? ?the hierarchical relationship.Then,incombination with the user's visit history,the user area interest model is established to show the degree of preference of users in different levels.The model of regional attraction is established to show the attractiveness of goods in different levels.Finally,based on the framework of location-based collaborative filtering recommendation method,two methods of collaborative filtering are proposed by using two regional models.One is regard the two regional models as the regional characteristics of the user and the commodity,and the two regional models are added into the hierarchical weight to the matrix decomposition process,so that the recommendation result is combined with the influence of the location areainformation.Another method is based on the user region interest model,the neighborhood of interest of the user region is obtained according to the similarity degree of the user region interest,and then the user area neighborhood is combined with the matrix decomposition to produce the recommendation result that combines the position information.Finally,this paper uses the reptile technology to capture the food information of a certain area on a website as experimental data set.The results show that the method based on the position-based collaborative filtering recommendation method and the traditional recommendation method.The combination of location information in the filter can improve the recommended accuracy,indicating the effectiveness of the proposed method of location-based collaborative filtering.