Predicting And Evaluating The Popularity Of Online News Based On R

Internet age is coming, the amount of information is rapidly growing, Online news has become the main carrier of the network information. People understand the people’s livelihood, current events through the online news. More and more people like to share news article, and the number that share news articles shows the popularity of news. In our work, part of the classification algorithm is applied to predict the popularity of online news, we want to explore the best model for online news popularity, and we want to help network news services predict the popularity of online news before publication,In our work, we forecast the popularity of online news according to the procedure of data analysis:before you can perform an analysis, you need to extract and pre-process the original data; and then, we apply feature selection to data set, our main feature selection method is based on the model, and that is recursive feature elimination of algorithms; Modeling analysis, we use a variety of different learning algorithm to fit data sets, such as adaptive boosting algorithm,random forest algorithm and support vector machine algorithm, then we analyze the model results. Finally, we found that random forests model is the best prediction model.Let’s see the article structure, In the first part of this article,we introduce the status,background significance and of online news research,In the second part, we review the model and methods used in the paper theoretically. As for the third part, we obtain the UCI data sets for data analysis, and the data sets come from the Mashable, which is a well-known online news sites, we do specific data analysis according to the procedure of data analysis. In the final part, we have obtained a reference conclusions,, analyze shortcomings and prospects of the paper.