Design And Implementation Of A Movie Hybrid Recommendation System Based On Spark

Because the speed expends of internet technology,the economic benefits brought by China’s film industry have continued to grow,in the 21 st century.The production of films is increasing year by year,the types are more abundant,and the number of people watching films is increasing substantially,but the quality of films is uneven.The propaganda of big film manufacturers is also in full bloom.How consumers can watch high-quality and suitable films in their free time plays a huge role in the evaluation of movie reputation and the box office of films.Therefore,in order to recommend a variety of high-quality films to users,a hybrid film recommendation system based on Spark is designed and implemented.In this paper,the theory,strengths and weaknesses of the current mainstream recommendation algorithm are studied,and the overall recommendation strategy is designed by combining a variety of recommendation algorithms with a variety of hybrid recommendation mechanisms.An offline recommendation component including statistical recommendation and model recommendation is devised.In view of the problem that the offline recommendation results are not updated in time,a real-time recommendation module was designed,and the collaborative filtering algorithm based on the project is improved.By strengthening the consideration of users’ recent rating,the preference of users in the short term is emphasized,and the recommendation results are updated quickly.In the end,a relatively stable and easy to use hybrid recommendation system was realized,through the use of Spark etc big data platform.The simulation experiments on the experimental data indicate the offline recommendation of the system has a excellent effect,the real-time recommendation can update the recommendation results in a timely manner,and the overall operation of the system is stable and good.