sentiment-analysis-bert

Sentiment140 Twitter dataset sentiment analysis model with BERT (Bidirectional Encoder Representations from Transformers). The polarity of the tweets is classified in this project. 0 = negative, 4 = positive

See details about the dataset: http://help.sentiment140.com/for-students

Huggingface pytorch implementation for a lot of NLP tasks. In this project, BertForSequenceClassification is used from transformers library. Transformers library 4.3.2 is used. You can install this version of the library: pip install transformers==4.3.2

Colab is used for fine tuning the pretrained BERT model.

Read this paper about BERT -> BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

Results:

                  precision    recall  f1-score   support

        Negative      0.82      0.81      0.82      3749

        Positive      0.81      0.82      0.82      3752


        accuracy                           0.82      7501

       macro avg       0.82      0.82      0.82      7501

    weighted avg       0.82      0.82      0.82      7501