Research And Implementation Of Movie Recommendation System Based On Hadoop

The traditional film recommendation system simply takes the user's interest degree as the main reference standard for recommendation.Although the recommended film meets the user's interest,due to the low quality of the recommended film,users often give up the use halfway.In order to enable users to find the types of films they are interested in in a timely manner,and at the same time ensure that the recommended films have high quality standards,and accurately recommend the films they are interested in to users.This topic is a combination of Hadoop architecture technology and collaborative recommendation algorithm,combined with the movie ratings data at the same time.The system recommended by the movie not only can fit their taste,but also with the characteristics of high grade,good reputation,to enhance the user recommendation system using viscosity,thus can continue to recommend for users of the system in accordance with the user type of movie for a change,to achieve a high quality movie recommendation at the same time the movie recommendation system conforms to the user's personal preference.This system will use collaborative recommendation algorithm to provide theoretical basis for film recommendation,guide the construction of recommendation process,and use Hadoop technology to provide basic architecture support for the system,so that a large amount of film resource data can be simultaneously processed and stored at a high speed.At the same time,this system conducts a survey on the practical requirements of film recommendation,and understands the relevant application algorithms of film recommendation,as well as the basic logical process of recommendation.The rating data in the film platform were used to provide references for the rating in the system.Based on the rating data,a user rating matrix was constructed to calculate the similarity between users.The system will be divided into the background management end and user android end,the background management end mainly carries out centralized processing of movie data,applies the recommendation algorithm,and extracts the recommendation objects.Android users mainly browse all movie resources and movies recommended for users.Analyze the requirements of Hadoop based film recommendation system developed in this project,mainly analyze the functional and non-functional requirements,investigate the realistic requirements of film recommendation,understand the relevant application algorithms of film recommendation,and the basic logical process of recommendation,etc.Put forward the functional business requirements of film recommendation system to meet the interests of viewers,and designed the structure system and general functional framework of the film recommendation system based on Hadoop.The system functions are divided into five functional modules: user information management,film information management,data processing module,feature analysis,and film recommendation.Using Android,C# programming language and MySQL database technology to form the technical basis,build the user interest model of the system,and finally complete the detailed design of the system's functional modules and database design.The system was tested,the system test steps were described,and then the system was tested for performance and function.In this paper,Apache Hadoop was selected as the distributed computing platform of the application system of this paper.Combined with the collaborative recommendation algorithm,this paper innovatively utilized the scoring data information in the movie viewing platform to conduct centralized processing of movie data,applied the recommendation algorithm,and extracted the recommendation objects.A powerful film recommendation system can be implemented to accurately recommend the types of films that meet users' interest.Moreover,the recommended films have the characteristics of high quality and high score,making them different from traditional film recommendation systems.