| With the development of the Internet and the IoT(Internet of Things),the production as well as the access to information has become more diverse and convenient.In this age of "information overload",the emergence of recommendation algorithms has undoubtedly brought great convenience and benefits to users and enterprises.However,with the increasingly complex application scenarios,the quality of traditional recommendation algorithms is highly limited by the sparsity of the dataset,and these algorithms can’t provide accurate recommendations according to the changing process of user’ s interest.In order to solve the problem of data sparsity and the problem of perception to user’s interest change in traditional recommendation algorithms,a hybrid recommendation algorithm based on the retroactive inhibition theory is designed and implemented in this paper,the main work is as follows:(1)Aiming at the problem that the traditional recommendation algorithm is limited by the data sparsity which affects the accuracy of the algorithm,in this paper,an object-based embedding model is used to generate the object embedding vector as the implicit representation,which can improve the semantic information content of the object representation.(2)Aiming at the problem that the traditional recommendation algorithm can’t sense the change of user’s interest,this paper introduces the theory of retroactive inhibition to construct a user preference model,which gives the algorithm the ability of sensing the change of user’s interest.(3)Finally,the hybrid recommendation algorithm is obtained by fusing the former two algorithms in a hybrid way of "feature complement".In addition,based on the proposed hybrid recommendation algorithm,this paper designed and implemented a movie recommendation prototype system to visualize the algorithm results and provide basic recommendation services to users.Finally,in the scenario of rating prediction,the proposed hybrid recommendation algorithm is experimented on public datasets,and the experimental results show that the proposed algorithm reduces the error of rating prediction by about 9%compared with the tested traditional recommendation algorithm.Meanwhile,the movie recommendation prototype system designed and implemented in this paper can also display the algorithm results and provide recommendation services successfully. |
