Extractive Text Summerization Pada Berita Berbahasa Indonesia Menggunakan Algoritma Support Vector Machine

Abstract

According to the Program for International Student Assessment (PISA) for the 2018 survey of 61 countries that participated in PISA, the reading interest of the Indonesian people still received a low score of 358 out of an overall average score of 472. One of the consequences of low reading is the difficulty of understanding the content of reading, especially for long and many texts, so it will be easier to read the summary. With advances in text summarization technology can be done using text mining methods. text mining will retrieve information on big data from text-based documents, the summary process will take the main points of news or important sentences without changing the content of the reading or also called extraction techniques. To get maximum results, the weighting is done by extracting sentence features based on numerical data, quotations, sentence length, sentence position in paragraphs, and overall sentence position. The research methodology uses knowledge discovery in database (KDD) and modeling using support vector machine algorithms. Testing or evaluation using recall, precision and F-measure. The best research result is the scenario of comparison of test data and training data 7:3, using the Linear kernel, with accuracy 72,4%, precision 63,4%, recall 51,9%, and F-measure 57,1%.