Social Network Based Context-aware Recommendation Algorithm

With the rapid development of information technology, human society has entered the era of information overload. In the face of massive information on the Internet, on the one hand, it is difficult for users to find information which interest them, on the other hand, the producer of information is also difficult to find users who are really interested in their information so that their information could get more attention.By means of the analysis of user behavior data, extract preferences of users, to provide users with personalized recommendation content, in a lot of network applications(such as e-commerce site Amazon, Taobao and social networking site Linked, Facebook, etc.), recommendation system has become a very promising tool for dealing with information overload.At present, the recommendation algorithm used frequently in the domain of recommendation system including user based collaborative filtering recommendation algorithm,item based collaborative filtering recommendation algorithm,latent factor model based recommendation algorithm, recommendation algorithm based on context information and social network information. One of the most widely used is collaborative filtering(CF)recommendation, which predicts the target user's preferences by using the historical data of similar users or items. Although the collaborative filtering recommendation algorithm has been widely used in the industry, the traditional collaborative filtering technology has only used the two-dimensional "user- item" and not considered other information. When the scale of information is increasing, the performance may face great challenges, such as the sparsity of data(i.e., the lack of sufficient number of similar users or items), and the decline of recommendation quality caused by the data sparsity and the homogeneity of the information sources.This paper mainly studies the context aware recommendation algorithm.The concept of context, the research status of context aware recommendation system, social network data and user behavior data were introduced in detail. The study focuses on the extraction of context information,the processing of various context information and social network data,the calculation of user similarity, and proposes a context extraction based perception recommendation algorithm and on the basis of it we introduce social network data proposing a social network based context-aware recommendation algorithm.There are many kinds of contextual information in the practical application, but different kinds of contextual information could not have the same impact on the preferences of users. The context aware recommendation algorithm based on context extraction is used to recognize the contextual segments that affect user preference by comparing the performance of traditional recommendation model in different segments of context, and the random decision tree algorithm is used to segment the scores of different types of contextual information. Since scores of the sub-matrix are in similar context, they have a higher degree of correlation with each other. In the leaf node of the tree, matrix factorization is used. To predict the rating of users for an item,we solve the objective function.Social network information is another kind of information that can have an important impact on user preferences. The social network based context-aware recommendation algorithm introduces a social regularization term, which predicts the user's preference by learning the preference of the user's friends. In order to identify the friends which have similar preferences,we introduce Pearson Correlation Coefficient(PCC) which contains context information to measure user similarity.The theoretical analysis and experimental results show that the performance of context extraction based perception recommendation algorithm and social network based context-aware recommendation algorithm is better than the traditional recommendation algorithm in terms of precision.