Popular Content Prediction Based on Web Visitor Data With Data Mining Approach

Abstract

A quality website has five parameters that must be considered are: information, security, convenience, comfort, quality of service. But of course, the fifth parameter does not always guarantee the amount through its Web page will increase, from that problem. So research is conducted to predict website content based on visitor data with a data mining approach, this research aims to improve the quality of content on target according to the interest of website visitors. Evaluation of random forest algorithm has the value the accuracy of classification of 71 percent by value of Kappa 0.712 whereas the k-NN algorithm has higher accuracy values of random Forest algorithm i.e. worth 84.88 percent and kappa values of the mean 0.847 the k-NN algorithm performs data processing process predictions against data of web content more effectively than the random forest

Web Based Online Hybrid Teaching Method of Network Music Course

“…The hierarchical recommendation algorithm will not only consider the interests and hobbies of the target users, but also combine the knowledge characteristics of the target users on the target of the resource base, including the theoretical knowledge level, the actual ability level and the local characteristics of the target, and take the above characteristics as the main recommendation. The advantage of this recommendation algorithm is that it considers the interests and abilities of the target users, the teaching level and methods of the course teachers, and the music course itself has different requirements for the theoretical level and actual performance level of the target users, so as to form a more accurate and three-dimensional teaching recommendation with a deeper degree of fit [18].…”