Design And Implementation Of Online Evaluation And Recommendation System

In recent years,with the development of Internet technology,personalized and intelligent online evaluation has gradually become an important way for the public to conduct self-learning test.Because the knowledge evaluation method of traditional classroom education model is difficult to meet the needs of people's self-evaluation.Therefore,online evaluation technology based on Internet has been widely concerned.At the same time,contact the recommendation system,from the perspective of the user's problem record,recommend the relevant test questions for the user according to the user's completed homework,in order to liberate the user from the problem sea tactics,so as to improve the user's learning efficiency and learning experience.This thesis mainly studies online evaluation and test recommendation system.Design and implement a web-based online evaluation system for operation test questions,which can download the operation questions by users themselves,complete the local answer and upload the platform,and automatically score and feed back to users by the background.And relying on this platform,we can obtain user behavior data,accurately recommend test questions for users,solve the problem of "overload" of test resources,and meet the user's personalized learning and testing needs.The main contents of this thesis include the following aspects:1.Based on the data of user's online test behavior,design the method of test recommendation.Based on the study of the classic recommendation algorithm,according to the application scenarios with many users and rich test resources,the LFM recommendation system model is selected.Put forward a strategy of recommendation: There are many aspects of users' needs,sometimes they want to recommend difficult test questions first,sometimes they want to recommend test questions whose mastery level has not reached a certain threshold.Therefore,two methods of selecting top-N test questions with different ranking directions are proposed to meet the personalized needs of users.For new users,according to the user registration information,it is recommended that other similar users have done the test questions to solve the cold start problem of the system.3.Put forward the method to improve the accuracy of test recommendation by using test label.This thesis uses the technology of Jieba to segment the test text,and assists the domain experts(teachers)to set the knowledge point label for the test.The k-means algorithm is used to cluster the test questions,filter the similar test questions that users have mastered,and optimize the recommended test set.4.Realize online evaluation and recommendation system.Using the Django framework of python,the prototype system of MS Office Online Evaluation Based on web is designed and implemented.The system can download office operation questions by users themselves,complete the answers locally and upload them to the platform.The website background automatically scores and feeds back to users in time.And rely on this platform,recommend test questions for users accurately.The main functions of the system are tested to achieve the design goal.