PENERAPAN PEMBOBOTAN ATRIBUT PADA ALGORITMA NAIVE BAYES UNTUK ANALISIS SENTIMEN REVIEW APLIKASI ANDROID DARI GOOGLE PLAY

Aris Tri Jaka Harjanta
Abdul Syukur
Catur Supriyanto

Jurnal Teknologi Informasi Cyberku
Vol 11 No 1 (2015): Jurnal Teknologi Informasi CyberKU Vol.11 no 1

Fast growing technology corelate with the demand of faster information access. Recently, the information technology dominate by android (open source) based smartphone, it makes many application developer build the application base on this operating system. With so many existing applications, users need a reference to see the application in general, although it has been provided a facility user review for this application, large number of users review are make the user difficult to be able read one by one. Thus it is necessary to know how the sentiment classification of users on the application. In this experiment, algorithms naïve bayes classifications applied are shown to have good performance on large data and have proven reliable in a variety of domains. As well as adding a attribute weighting use algoritm of weight by correlation, weight by chi squered statistical and weight by SVM on the data, so expect a good accuracy of the sentiment analysis android application to use in Indonesian sentiment.