Research On The Application Of Data Mining In Cloud Classroom Operation

With the rapid development of the modern Internet,Internet technology has been applied to all fields of life,which not only changes our lifestyle,improves the quality of life,but also changes the previous education mode and learning mode,at the same time the online learning mode has also been greatly developed and improved,such as through multimedia technology video and other forms of teaching resources reasonable and effective use into the online teaching practice process,which makes online learning an important way for people to learn daily.However,in the face of the rich course information provided by many online course platforms,resulting in the problem of "information overload",learners will inevitably fall into the "information trek" of course selection.With the development of the times,the competition of cloud classroom is becoming increasingly fierce,and in order to achieve long-term sustainable development,it is necessary to adopt a reasonable operation model and strategy.To this end,more and more Internet companies are actively using data mining technology,using technology to assist manual operations,and providing high-quality marketing strategies for enterprises,thereby promoting the healthy development of the Internet.This paper relies on the operation data of the cloud classroom and uses cloud model,collaborative filtering,deep learning and other technologies to help the cloud classroom provide users with better services,thereby increasing the number of users of the platform.In view of the current situation of cloud classroom operation,this paper studies the following three aspects:(1)Research how to target the target population of the cloud classroom.By analyzing the registered users of the entire platform,it is found that only a small part of a large number of registered users are really learning in the cloud classroom,hoping to screen out the registered users who are interested in course learning,then for this unbalanced data set,how can the system screen them out and formulate a certain message push and marketing strategy,this paper uses the cloud model to solve the imbalanced data set,and then through SVM classification and screening,to lock the target users of the cloud classroom.After the cloud model is experimentally processed for data,the classification accuracy,recall rate and F1 index of SVM are improved compared with direct SVM,and the cloud model has universal applicability in machine learning algorithms and can be applied to common machine learning algorithms,such as KNN,decision trees,random forests and na(?)ve Bayes.(2)Research how to recommend suitable courses for users.In response to this problem,this paper proposes an improved collaborative filtering algorithm based on time factors,aiming at accurately recommending courses for users,even for new users,but also according to some basic attributes of users,give some correct guidance,so that users can quickly find courses suitable for themselves on the platform and give users a good experience.in order to enhance the reputation of the platform.Through experiments,it is found that the improved collaborative filtering algorithm,compared with the traditional user-and project-based collaborative filtering,the recommendation accuracy will be higher,and it also plays a certain role in the cold start of new users.(3)Research how to generate comprehensive ability scores for users.Based on deep learning,a comprehensive ability score is automatically generated for users based on their historical learning records.Let users quickly have a clearer understanding of their own knowledge,which is convenient for the arrangement of subsequent learning plans.