Design And Implementation Of Drama Recommendation System Based On Social Information And Context Awareness

With the rapid development of social network,the recommendation system of city activity has become one of the hot spots in the research of recommendation system.At present,the most dramatic events recommendation algorithm based on the analysis between the user and participate in drama activities of a single interaction between similarity and user perspective,such as ignoring the drama activities caused by sparse data cold start,the user and the social relationship between context and drama activities of mutual influence and the change of user’s preferences.In order to solve the above problems,this is designs a drama activity recommendation system based on social information and context awareness.The system analyzes historical behavior data and real-time behavior information of users participating in drama activities by using a variety of improved recommendation algorithms,so as to make real-time dynamic recommendation for users.The main contents of this is are as follows:(1)In order to solve the problem of recommending cold start and sparse original data for drama activities.A hybrid drama activity recommendation model based on context-aware clustering is proposed.By integrating clustering algorithm and spatio-temporal context aware network,the model uses sparse data information to mine users’ real preferences,and carries out personalized Top-N drama activity recommendation for them.Experiments on Douban and Meetup data sets confirm that the model can alleviate the cold start problem to some extent.(2)In order to make full use of users’ social information and improve the recommendation performance of drama activity recommendation system.This is proposes a drama activity recommendation model based on social information.The model will be users to participate in the activities of each of the drama of the historical records and the attention of the user relationship as input,fully tap the user and social information of pluralistic interaction relationship between context and drama activities,build a user-the preference matrix characteristics,study it is concluded that the user preference score of drama activities,personalized Top-N drama activities for users is recommended.Experimental results show that the proposed model can improve the accuracy rate and NDCG recommendation indexes,and improve the recommendation performance of drama activities.(3)To capture users’ dynamic preference for dramatic activities and further improve the recommendation effect.A serialized drama activity recommendation model based on users’ dynamic preferences is proposed.This model embedded the time series of dramatic activities into the attention mechanism,assigned different preference values to each dramatic activity,and learned personalized Top-N dramatic activity recommendations by modeling users’ short-term preference and long-term preference.Finally,experiments are carried out on real data sets,and it is proved that the proposed model has some improvement in AUC and accuracy.(4)In order to integrate the studies of the first three research points,this is designs and implements a drama activity recommendation system,which can provide dynamic personalized drama activity recommendation service for users while alleviating the problem of data sparsity,improve the recommendation accuracy rate and meet users’ various needs.The client of the system uses ReactNative framework to build and write JavaScript,and tests each function.