Implementasi Kombinasi Algoritma Convolutional Neural Network dan Long Short Term Memory (CNN – LSTM) untuk Mengidentifikasi Judul Berita Clickbait Bahasa Indonesia

Abstract

The development of communication and information technology can now be disseminated quickly through online news portals. Based on the Survey of the Indonesian Internet Service Providers Association (APJII), the penetration of internet users in Indonesia in 2023 has reached 78.19 percent or 215.62 million out of a total population of 275.77 million. These online news providers benefit from advertising on the online news sites they create to gain trust from advertising to achieve this must have high traffic. Clickbait is a headline on a piece of content designed to manipulate or provoke readers' curiosity to click on a news link, often in a way that is misleading or not entirely in line with the content of the article. This is what makes clickbait have a negative impact on readers, due to the low level of public literacy. This research was conducted with the aim of automatically identifying clickbait and non-clickbait news titles using the Convolutional Neural Network Algorithm combined with the Long Short Term Memory Algortima. The research was conducted by involving 8,613 Indonesian language news headline data, through the preprocessing stage, followed by the word embedding stage using the fastText library. The accuracy rate obtained from the model built in this study is 0.92 and the loss is 0.33. After several evaluations, it can be concluded that the combination of algorithms used by researchers can identify clickbait and non-clickbait headlines with good performance.