Design And Implementation Of Encyclopedia Knowledge Recommendation System Based On Deep Learning

With the popularity of the Internet,more and more people can participate in the world of the Internet,and the content of the Internet world is becoming more and more abundant;With the development of society and the advancement of technology,mobile phones are becoming more and more popular nowadays,and more and more people can use mobile phones to experience the charm of the Internet.The function of the mobile phone is now more than just a tool for communication,but more of an interface that we interact with the Internet world.People use the mobile phone to access the Internet to query data and get information to understand the world.However,with the exponential growth of the amount of information on the Internet,the question people are now facing is how to find content that is of interest to their needs in the ocean of such information.The recommendation system came into being in the context of this information overload.The recommendation system to solve is how to accurately push the huge amount of information on the Internet to users interested in it.This thesis designs and implements an encyclopedic knowledge recommendation system based on the mobile APP application scenario.The main function of the system is to recommend various encyclopedic knowledge that user will be interested in.The background of the knowledge recommendation system is described in the paper.The model used in the recommendation algorithm and the evaluation method of the model are briefly introduced.Through the analysis of business processes,five system modules such as data collection and preprocessing,model construction,recommended item recall,recommended item sorting and recommendation result evaluation are determined.The system uses multi-channel recall technology in the recall phase,and adopts the CTR estimation model in the sorting phase to estimate and rank the click rate.In order to effectively evaluate each model used in the recommendation system online,compare the effects of each model in the actual application scenario,and also perform AB test on the system,and compare the effect of the model through the click rate in the actual application scenario.The personalized encyclopedic knowledge recommendation system described in this article has been officially put into online operation.The personalized encyclopedia knowledge recommendation function provided by the system can help users to discover the encyclopedia knowledge they are interested in.For the user,the system can greatly help them explore their interests and enhance the user experience when using software products;For software products,user retention is increased and new users are attracted.At the same time,in the process of system implementation,some engineering problems that will be generated when applying the deep learning model to practice are solved,which provides a valuable reference for solving such problems in the future.