L-Boost: Identifying Offensive Texts From Social Media Post in Bengali

Abstract

Due to the significant increase in Internet activity since the COVID-19 epidemic, many informal, unstructured, offensive, and even misspelled textual content has been used for online communication through various social media. The Bengali and Banglish(Bengali words written in English format) offensive texts have recently been widely used to harass and criticize people on various social media. Our deep excavation reveals that limited work has been done to identify offensive Bengali texts. In this study, we have engineered a detection mechanism using natural language processing to identify Bengali and Banglish offensive messages in social media that could abuse other people. First, different classifiers have been employed to classify the offensive text as baseline classifiers from real-life datasets. Then, we applied boosting algorithms based on baseline classifiers. AdaBoost is the most effective ensemble method called adaptive boosting, which enhances the outcomes of the classifiers. The long short-term memory (LSTM) model is used to eliminate long-term dependency problems when classifying text, but overfitting problems occur. AdaBoost has strong forecasting ability and overfitting problem does not occur easily. By considering these two powerful and diverse models, we propose L-Boost, the modified AdaBoost algorithm using bidirectional encoder representations from transformers (BERT) with LSTM models. We tested the L-Boost model on three separate datasets, including the BERT pretrained word-embedding vector model. We find our proposed L-Boost's efficacy better than all the baseline classification algorithms reaching an accuracy of 95.11%.INDEX TERMS Offensive text, social media harassment, natural language processing, ensemble learning, BERT model.

Deep-BERT: Transfer Learning for Classifying Multilingual Offensive Texts on Social Media

“…In Bengali text, they outlined the systematic technique of the BERT model for sentiment analysis. Recently, several NLP researchers have used BERT-based transformer models for text categorization [3], [19], [20]. A few researchers [3], [19], [20] used fine-tuning of BERT transformers, whereas Cohan et al [19] exploited pre-trained and fine-tuned datasets in sequential text classification and achieved better accuracy than other models.…”

“…The International Workshop on Semantic Evaluation (SemEval) has received proposals from over 100 groups for shared assignments [2]. Several attempts have been made to classify offensive languages, but only in monolingual contexts, focusing on English [2], Bengali [3], and other languages. Mridha et al [3] proposed a hybrid model for the Bengali offensive text classification.…”

“…Several attempts have been made to classify offensive languages, but only in monolingual contexts, focusing on English [2], Bengali [3], and other languages. Mridha et al [3] proposed a hybrid model for the Bengali offensive text classification. They proposed long short-term memory (LSTM)-BERT and Adaboost-BERT models separately and then combined these two models to create the final L-Boost model.…”

“…In their research, they analyzed offensive text in both Bengali and Banglish texts and compared the results with other existing models. This study was inspired by Mridha et al [3] to identify offensive text on multilingual-based datasets. In addition, numerous languages predominate as the medium of communication, resulting in the diversification of multilingual text classification, which can be effective in a variety of situations such as identifying offensive text and preventing intrusion.…”

AugFake-BERT: Handling Imbalance through Augmentation of Fake News Using BERT to Enhance the Performance of Fake News Classification

“…Health-related fake news affects people's lives [7]. During the COVID-19 pandemic, false information and offensive text were rapidly propagated, which misled people [8,9].…”