Peringkasan Dokumen Teks Bilingual sebagai Reduksi Fitur untuk Klasifikasi Menggunakan Algoritma K-NN

Abstract

Summarizing text is a step to extract the essence of a text document with no more than half. Summarizing text has an important role in extracting the core information from a document in a more concise form. Summarizing text documents can be used as feature reduction in classifying text documents because it can reduce features that are considered irrelevant. Text documents are summarized using the Term Frequency-Inverse Document Frequency (TF-IDF) method, then classified using the K-Nearest Neighbor (K-NN) algorithm. One of the disadvantages of the K-NN algorithm is that it is not optimal in classification if the k value is not appropriate, as well as the selection of an inappropriate distance calculation method. By testing various k values ​​and using the Euclidean Distance distance measurement method, you can increase the accuracy of text document classification. Text document summarization using the proposed TF-IDF method is proven to increase when classification is carried out with K-NN. From the research results, it was found that the classification accuracy at the compression rate increased by 50% with a k value of 6 to 8 of 95.33%. This shows that text document summarization as feature reduction has a positive role in the classification process using the K-NN algorithm.