Hate-Speech-Detection-in-Tamil

Hate Speech Detection and Sentiment Analysis in Tamil Language.

Datasets Used

1. HASOC - Task1, Task2, HL & Tanglish
2. EACL 2021
3. Dravidian 2021

Methods used for Pre-processing

1. Stop word removal
2. Stemming

Models used for Embedding

1. IndicBert
2. Bert Base Uncased (MaskedLM)
3. XLM-RoBERTa (Base)
4. TaMillion
5. MuRIL
6. mBERT

TF-IDF used as vectorizer for baseline model