Baheta: Balanced and Unbalanced Dataset in Arabic Clickbait Detection Using a Deep Learning Model (LSTM)

The term “Clickbait” refers to content that has the express intention of grabbing the reader’s attention. It has become an annoyance for social media users because of the deception contained in the titles of Clickbait. Many studies detect Clickbait using deep learning (DL) and machine learning (ML) models. However, detecting Clickbait in Arabic titles was addressed by a few studies, all of which used ML techniques. This is where our research originated from. In our proposed work Baheta, which is an Arabic synonymous with lying, we suggest utilizing a deep learning model called long short-term memory (LSTM) to identify clickbait in Arabic headlines. In order to extract features from the text, we utilized Word2vec. In this study, we train the model on two Arabic datasets, the first is an unbalanced dataset and the second is a balanced dataset, which is about merging the unbalanced dataset with a fake news dataset. Word2vec provided the model with the best results, with a Macro-F value of 0.79 when applied to the unbalanced (raw) dataset. The LSTM model showed better performance with the unbalanced dataset, as it obtained a higher Macro-F value of 0.02 than that obtained by the LSTM with the balanced dataset.