Research And Application Of Personalized Intelligent Recommendation Engine Algorithm

The Internet has been gradually going into people’s lives, the globalization ofinformation has become a trend, through the network to obtain information becomecommon.In recent years, the Internet information grow quickly, vast amounts ofinformation filled with the Internet, users feel more and more difficult to accurately findthe information they need,"information overload" has become a huge challenge for thedevelopment of the Internet.Personalized recommendation technology is one of the most effective tools to solvethe information overload.As recommendation engine can help users filter informationintelligently.Collaborative filtering to recommend the technology and content filteringrecommendation is popular recommendation technology, they can filter information forthe user.But in the actual application process is still facing problems such as the lowquality of recommended, degree of automation, cold start, real-time response.By analyzingthe advantages and disadvantages of a single collaborative filtering and a single contentfiltering technology, based on collaborative filtering technology, integration of contentfiltering technology, put forward the idea of a hybrid recommendation algorithm.Usingproject’s content to compensate for the lack of ratings of users on the project, which usethe content filtering for the user to find similar neighbors, then you can use collaborativefiltering technology to produce the recommended results.KNN algorithm-based contentfiltering technology to supplement the sparse matrix can improve the quality ofcollaborative filtering recommendation, thereby enhancing the final recommendationquality.Based on the above hybrid recommendation algorithm ideas and theory and designonline, offline module,we can make mixed recommendation algorithm, by the measure ofthe average absolute deviation of the results of assessment, it verify the superiority of thehybrid recommendation algorithm.Finally, design a movie recommendation system based on a hybrid recommendationalgorithm. The system is based on the hybrid recommendation algorithm explored,provide personalized recommendation services, and comprehensive record of userbehavior, from different perspectives to collect users’ interests and hobbies, to get user information to enhance the user experience.