The Design And Implementation Of Creative Production Sharing Platform Based On Matching Recommendation

The popularity of short video has brought considerable user traffic to Internet companies that develop short video platforms.In order to enhance the company's business value,Internet companies need to consider monetizing some of the user's traffic,and placing short video ads from advertisers on short video platforms is one of the most effective means of monetizing traffic.In order to reduce the cost of ad serving,advertisers want to place high-return short video ads on short video platforms.To produce an excellent short video advertisement,three major problems need to be solved: "It is difficult to find ideas,it is difficult to create ideas,and it is difficult to match in both directions." At present,short video marketing centers for web applications such as Alimama Creative Center and Baidu Marketing Creative Center already exist on the market,which are intended to help advertisers and ad creators to solve the difficulties encountered in making short video ads.As a mobile application,Creative Production Sharing Platform can directly create short video ads on mobile devices and provide personalized recommendation capabilities to help advertisers and ad creators discover valuable short video ads and creative production experience articles.Creative Production Sharing Platform supports two major clients,IOS and Android,and the overall server is developed using a microservice architecture.The service is divided into multiple business independent microservice applications according to business division,and each service interacts through RPC remote call or message queue.Service instances are deployed through Docker,and service clusters can be scaled out through multi-instance deployment,thereby improving the throughput of service clusters.The business service uses the Hive data warehouse to store massive recommendation data and uses the Elastic Search search engine to achieve full-text search functions,and adopts a matching-based multi-channel recall recommendation strategy to recommend personalized short video ads and creative experience articles for users.This thesis describes the design and implementation of Creative Production Sharing Platform,introduces the background of the platform development and the development technology framework selected by the platform,describes the main requirements of the platform and the outline design of the system,and shares the detailed design and implementation of the article module,search module,comment sharing module,video recommendation module and article recommendation module.I have undertaken server-side development work in the project,mainly including the design and implementation of each functional module,recommendation algorithm improvement,system architecture upgrade and project reconstruction.At present,Creative Production Sharing Platform has been put into operation and operated stably,and has achieved good benefits in helping advertisers and advertising creators.