A model was trained with the dataset created by combining the dataset consisting of Turkish tweets and slang words in Turkish. In this model, in which Overfitting was seen in the first study, Turkish offensive language detection is performed. In this study, which is under development, it is aimed to increase the accuracy by using pre-trained models and trying to increase the data.
Files | ||
---|---|---|
bad.txt | Turkish Slang words (+18)" |
|
Offensice_Language.ipynb | File with source code" |
- Pandas
- Numpy
- Tensorflow
- NLTK
- re
- Scikit-Learn
- datasets