Klasifikasi Dokumen Bahasa Indonesia Menggunakan Semantic Smoothing dengan Ekstraksi Ciri Chi-square

Abstract

One of supervised learning methods for document classification is Naive Bayes classifier. A common problem that often occurs on simple method like naive bayes is data sparsity. This problem especially occurs when the size of training and testing data is too small. Smoothing technique is a method for handling the sparsity problem; one method of smoothing technique is semantic smoothing. This research is intended to implement chi-square term extraction on document classification using semantic smoothing and to compare the classification accuracy rate with previous research. Chisquare term extraction was used to make the classifier work efficiently and to increase classification accuracy. Agriculture Research Journal Document of holticulture domain are used for this research, consisting of three classes. The average for accuracy of document classification on semantic smoothing with chi-square term extraction is 96%. The results of the classification using semantic smoothing with chi-square Term extraction have been able to classify Agriculture Research Journal Documents in the holticultural domain.