Design And Implementation Of Video Search Recommendation System

Nowadays,the Internet is developing rapidly.With the increase of its influence,the amount of information in the Internet is also increasing explosively.The system that can help users screen out the information they need from the mass of information comes into being,among which search engine and recommendation system are both solutions.The application scenario of search engine is that users need to accurately query the data they want to obtain,but the information explosion leads to many search results,which still needs users to filter by themselves.In addition,the user's search may not be able to accurately describe the content to be searched,and fuzzy search terms will also bring trouble to search engines to get the results users want.Therefore,thesis hopes to combine search and recommendation,use search engines to provide data sources,and use the advantages of recommendation system to solve the shortcomings of search engines.Centering on the above themes and purposes,thesis mainly studies the design and implementation of a video-oriented search and recommendation system using big data technology and tools.From the system function module,the system is mainly divided into data processing module,search module,recommendation module and web module.From the specific technical implementation of the system,this system focuses on the following three aspects.The data storage system is composed of a combination of traditional relational database and NOSQL database,which ensures the high-speed access of various types of data and the integration of high-capacity,low-cost,and high-speed storage system.Kafka and Spark are mainly used for data processing and calculation,which not only realizes the stability and reliability of the system data processing process,but also realizes the efficient processing of large quantities of streaming data and off-line calculation of timed processing data,which ensures the overall efficiency of the system.The result display uses Java related technologies to build a front-end and back-end separated web system,and uses the browser to realize the interactive use of the system and the display of the results.The main function of this system and action have the following points.The personalized recommendation function is introduced in video search to improve the user's search efficiency.Various data are used to train the recommendation model,and different recommendation algorithms are configured to provide diversified recommendation functions.In the end,the system not only achieves the established functional goals,but also ensures the performance of the system under the guidance of hierarchical design and distributed cluster computing,which is finally verified by the system test.