/Sentiment-Analysis-Model-Tests

Sentiment Analysis on the IMDB dataset: testing different word-number encodings and models to maximize accuracy

Primary LanguageJupyter Notebook

Main:
Open In Colab

Transformers:
Open In Colab

Social Media Sentiment Analysis Project

Description: Python Project to classify social media comments as positive or negative
Status: In Progress

To Do List:

Features to add:
  • Add validation data from original dataset
  • Add more datasets: SST-2, Yelp-2, Amazon-2, DBPedia
  • Grid Search
  • Organize Code
  • Integrate with Django to create web application
Models to check:
  • BERT + XLNet
  • Use NEAT to tune hyperparameters of XLNet
  • ALBERT
  • LSTM/RNN Model Try
Lower Priority:
  • Try Keras Embedding Layer
  • Try GloVe embeddings
  • Write Vectorizer class to make code less repetitive
  • Complete state-of-the-art SMART-RoBERTa
  • Update IPYNB files with PY