Text Summarization terhadap Berita Bahasa Indonesia menggunakan Dual Encoding

Abstract

Text summarization or automatic text summarization can make readers receive information quickly without having to read the entire news text, so readers can get more time in reading other news texts. Making text summarization can use two techniques, namely, extractive and abstractive techniques. Abstractive techniques have the aim of producing summary sentences with concepts as humans take the essence of a document that is read. In this study, the author builds an abstractive summarization model using the Dual Encoding method consisting of GRU. The evaluation was carried out using K- Fold Cross Validation, the number of folds used was 5. By using K-Fold Cross Validation, the ROUGE- 1, ROUGE-2, and ROUGE-L values were 0.2127749, 0.119851, dan 0.1880595, respectively. For testing when using new data ROUGE-1, ROUGE-2, and ROUGE-L values were 0.3387776, 0.2395176, dan 0.3077376, respectively.