This project is a clickbait news headline classification project using three different deep learning algorithms namely LSTM, Bi-LSTM, and GRU. The three algorithms are compared based on accuracy, precision, recall, and F1-score metrics. The performance comparison of the three algorithms is also based on the amount of time required for the training process. This project is also an experiment that underlies the creation of a journal entitled "Perbandingan Kinerja LSTM, Bi-LSTM, dan GRU pada Klasifikasi Judul Berita Clickbait" which in English means "Comparison of LSTM, Bi-LSTM, and GRU Performance on Clickbait News Headline Classification".
Clickbait Dataset is dataset contains headlines from various news sites such as ‘WikiNews’, ’New York Times’, ‘The Guardian’, ‘The Hindu’, ‘BuzzFeed’, ‘Upworthy’, ‘ViralNova’, ‘Thatscoop’, ‘Scoopwhoop’ and ‘ViralStories’. It has two columns first one contains headlines and the second one has numerical labels of clickbait in which 1 represents that it is clickbait and 0 represents that it is non-clickbait headline. The dataset contains total 32000 rows of which 50% are clickbait and other 50% are non-clickbait.