Design And Implementation Of A Personalized Movie Recommendation Engine Based On Collaborative Filtering Algorithm

With the rapid development of Internet and electronic commerce,information overload has become a serious problem that internet users are facing.The internet has provided so much product information which makes it difficult for users to quickly find items they want.At the same time,the internet enterprises also need to consider how to provide users with interesting products,to win the user's affection.Recommendation system is an effective means to solve this problem.Collaborative filtering technology is one of the popular recommendation methods that many website is using.In this paper,we analyst and improved the traditional collaborative filtering recommendation algorithm and take movie recommendation as an example,studied how to use collaborative filtering algorithm to design and implement a personalized recommendation engine,the specific contents are as follows:(1)Firstly,we make an introduction of the research history,present domestic and international research background of recommendation system.Then,we elaborates the concept,theory,technology and related applications on the recommendation system,the concrete content includes recommendation algorithm classification,similarity measurement,evaluation of recommender system technology,recommend application of system etc..(2)Secondly,taking the movie recommendation system as an example,we use the Movielens dataset to build a movie recommendation engine,the recommendation engine realizes three main functions,respectively include: the dataset construction,mining of neighbors,movie recommendation.The engine implements the user-based filtering and item-based filtering recommendation algorithms.Traditional collaborative filtering algorithm calculates the difference of scores only for the common items of users while modeling the similarity of two users,and without taking into account the number of items together,which makes the calculation results deviate from the real situation.So we proposed a new algorithm,the new algorithm take the difference of scores and the numbers of common items into consideration while calculating the similarit y of users.In addition,considering the similarity evaluating effect of hot items is not as good as unpopular items,the similarity weight of hot items is reduced.The experimental results show that,the recommendation quality of new algorithm is improved by more than 1 times than traditional algorithm in both precision and recall.(3)finally,we studied the movie recommendation system on analysis,designing,implementation and testing of the system from the perspective of designing software engineering project.