Table 10: Comparison of MACD

However, for other languages, presence of large-scale datasets is still lacking. Anonymized comments: To further address privacy concerns, we anonymize our dataset. User mentions and email addresses contain "@" and we use regular expressions to search for matching Raw data: We do not store the raw data used for this study. We combine the hate and offensive categories in these datasets for training a binary classification model. In this section, we discuss errors across different languages.