Analysing user ratings for classifying online movie data using various classifiers to generate recommendations

With the tremendous growth of information available on the Web, there is a dire need of classifying it for the ease of use and for the brisk accessibility. The classified data can be used for making recommendations. This paper discusses and compares three classifiers applied on movie data-set using WEKA 3.7 data mining tool. The data is classified into five different classes namely: bad, ok, average, good and excellent. A discussion about true positive rate, false positive rate, precision, and recall based on confusion matrix for each class is carried out. Subsequently the boundary visualization of data is captured in form of a graph, and a meaningful comparison between Zero R rule, Naïve Bayes classifier and J48 tree is done by experimentation and analysis. In this paper, an attempt has been made to analyze the best classifier for movie data based on users´ ratings and then the classification is used for making the recommendations for users.