PENERAPAN ADABOOST UNTUK MENINGKATKAN AKURASI NAIVE BAYES PADA PREDIKSI PENDAPATAN PENJUALAN FILM

Abstract

For economists and financial experts predicting the success of doing business is very interesting. With the data analytics the prediction process has been facilitated by the past data stored to find out what will happen in the future. This research was conducted to facilitate the film industry players in considering the factors that can influence the income of the film to be produced. The naive bayes method is a popular machine learning technique for classification because it is very simple, efficient, and has good performance on many domains. But naive bayes has a disadvantage that is very sensitive to too many features, thus making the accuracy to be low, in this case the adaboost method to reduce bias so that it can and improve accuracy from naive bayes. Validation is done by using 10 fold cross validation while measuring accuracy using confusion matrix and kappa. The results showed an increase in the accuracy of Naive Bayes from 83.22% to 84.44% and the kappa value from 0.706 to 0.731. So that it can be concluded that the application of adaboost on 2014 & 2015 CSM film data is able to improve the accuracy of the Naive Bayes algorithm