Research And Design Of Personalized Movie Recommendation System

With the advent of the era of big data,the amount of data is increasing day by day,and the recommendation system has also ushered in development and challenges.For example,the recommendation system exposed problems such as high percentage of coincidence points,cold start and data sparsity.According to the characteristics of collaborative filtering recommendation algorithm,paper studies various recommendation techniques in collaborative filtering algorithm.Different solutions are proposed for the above problems,namely the optimization of the similarity calculation formula.Firstly,in order to solve the problem that the proportion of coincidence points is too high,two optimizations are proposed for the similarity calculation formula of the collaborative filtering recommendation algorithm.Two similarity calculation formulas were obtained: Euclidean-CPARW and Concurrence-RW.Tested on the MovieLens100 K dataset,using the root mean square error as the evaluation index,the recommended accuracy of the formula Euclidean-CPARW is about 1% higher than that of the other four similarity formulas.The formula Concurrence-RW uses the accuracy as the evaluation standard,and the accuracy of the recommended results is about 3% higher than the other four similarity formulas.Secondly,in order to alleviate the cold start problem and data sparseness problem,an optimized hybrid collaborative filtering algorithm is proposed.The algorithm combines the movie label information,based on the optimized mixed similarity formula BAJ-DCMS,adaptively selects a similarity formula that is beneficial to the current score by the number of common scores in the score table.The formula BAJ-DCMS proves the effect of the algorithm on the MovieLens dataset.Using the root mean square error as the evaluation index,the calculated recommendation accuracy is improved by about 2% compared with the other two similarity formulas.Including Spark Big Data Processing and Scrapy Web Crawler,the recommendation engine uses the optimized mixed similarity formula BAJ-DCMS described above.System framework using SSM framework,database MySQL and the distributed storage system HDFS are used for the storage of data,the recommended personalized movie list even each movie poster,home page and other information can be showed to user on Web by the interaction between the Web and the user.