Design And Implementation Of TikTok Live Monitoring And Advertising Recommendation System

With the rise of short video and live broadcast,the channels for young consumers to obtain product information are undergoing tremendous changes.Content e-commerce based on short video platform is reconstructing the connection between consumers and brand merchants,and traditional e-commerce is facing unprecedented challenges.TikTok live broadcast business is developing rapidly,and it has become a new trend for users to watch live broadcast and buy goods in the live broadcast.But on the TikTok advertising needs to be done through domestic agents,to make goods on users,choose what kind of the host to take the goods is very difficult and there is no a good product can be checking and analysis of a televised commodity transaction data,popular master data as well as selling category data,Due to the large user base of short video platform,problems such as low efficiency of advertisement recommendation and low advertisement conversion rate are easy to occur in the effect of advertisement recommendation.Therefore,it is necessary to combine intelligent recommendation algorithm and find user preference according to user history information to realize intelligent recommendation of advertisement.Based on this situation,this paper takes TikTok live broadcast business as the background to design and implement TikTok live broadcast monitoring and advertising recommendation system.This paper first introduced the birth background of the project,compared with similar industries at home and abroad,the development status of similar systems,and clarified the focus and necessity of the project.Then it describes the theoretical technology and system framework involved in the project,and makes a detailed demand analysis,outline design and detailed design of the system.The whole system is divided into user management function,live broadcast monitoring function,background management function and intelligent recommendation function.In the process of research and development,this project adopts the client-load balancer-middle server-application server-database framework,using the remote call Thrift framework to implement a distributed system,Kafka as message middleware to realize producer consumer model,using My SQL as a storage medium,Redis as a cache to increase system response data rate.After the completion of development,test cases were designed for each function,and the system performance was tested with the help of pressure testing tools to ensure that the system could run stably under actual user conditions.The TikTok live monitoring and advertising recommendation system designed in this paper has been running steadily since it was put into use,basically meeting business needs and providing a platform for users to open accounts and register,replay live broadcast data and recommend advertisements.