Implementasi Support Vector Machine (SVM) Untuk Klasifikasi Dokumen

Abstract

Document classification is the process of grouping documents into specific categories that have been defined previously. In the real world, the distribution of the data is generally non-linear, which means the distribution of the data did not separate properly. Therefore we need a method that can classify documents that are non-linear. Support vector machine can classify documents that are non-linear with increasing dimension of the distribution of documents using kernel trick. This study will use a linear kernel for the classification of text documents. Results greatest accuracy in this study was 76% of 150 test documents with epsilon value 0.01. Factors affecting the results of the classification of which is the value of epsilon and length of test documents.