SSN_NLP_MLRG at SemEval-2020 Task 12: Offensive Language Identification in English, Danish, Greek Using BERT and Machine Learning Approach

Proceedings of the 14th International Workshop on Semantic Evaluation, pages 2161–2170 Barcelona, Spain (Online), December 12, 2020. 2161 SSN_NLP_MLRG at SemEval-2020 Task 12: Offensive Language Identification in English, Danish, Greek using BERT and Machine Learning Approach Kalaivani A Department of CSE SSN College of Engineering, INDIA kalaiwind@gmail.com Thenmozhi D Department of CSE SSN College of Engineering, INDIA theni_d@ssn.edu.in Abstract Offensive language identification is to detect the hurtful tweets, derogatory comments, swear words on social media. As an emerging growth of social media communication, offensive lan- guage detection has received more attention in the last years. We focus to perform the task on English, Danish and Greek languages. We have investigated which can be effect more on pre- trained models BERT (Bidirectional Encoder Representation from Transformer) and Machine Learning Approaches. Our investigation shows the performance between the three languages and to identify the best performance is evaluated by the classification algorithms. In the shared task SemEval-2020, our team SSN_NLP_MLRG submitted for three languages that are Sub- tasks A, B, C in English, Subtask A in Danish and Subtask A in Greek. Our team SSN_NLP_MLRG obtained the F1 Scores as 0.90, 0.61, 0.52 for the Subtasks A, B, and C in English, 0.56 for the Subtask A in Danish and 0.67 for the Subtask A in Greek respectively. 1 Introduction Offensive language detection is the process of identifying and detecting the user generated offensive content or comments that are insult, hurt, profanity, and racism and targeted the individual or group in the massive social media (Marcos Zampieri et al., 2019). As an immense growth of social media, user generated content increasingly occurs on the social media platforms and much more attention to deal with the offensive content. Detecting the offensive content is helpful for the field of sentimental analy- sis (kalaivani A and Thenmozhi D, 2019), abusive identification (Kenneth Steimel et al., 2019), aggre- gation and cyber bullying. Several research works have been performed to identify the offensive lan- guage in social media. Research workshop has reported in NLP communities that are GermEval 2018 Shared Task on the Identification of Offensive Language (Michael Wiegand et al., 2018). OffensE- val@SemEval2019 (Zampieri et al., 2019b) shared task focuses on identification and categorization of offensive content in social media. OffensEval@SemEval2020 (Zampieri et al., 2020) shared task fo- cuses on multilingual identification and categorization of offensive content for five different languages in social media. Recent research area has reported to identifying the offensive content (Zeses Pitenis et al., 2020; Gudbjartur Ingi Sigurbergsson and Leon Derczynski, 2019) and abusive comment (Kenneth Steimel et al., 2019) in different languages. Zeses Pitenis (2020) used machine learning and deep learning techniques for the identification of offensive content in Greek annotated Tweet dataset. (Gud- bjartur Ingi Sigurbergsson and Leon Derczynski, 2019) used Logistic regression and more different attention in BiLSTM models to detect and categorize the offensive language in the English and Danish annotated Facebook, reddit dataset. (Kenneth Steimel et al., 2019) used machine learning approaches and sampling method to detect …

Content truncated to 500 words