1. To scrap twitter post by keyword, I used twint because i can't get Twitter API --> install twint https://github.com/twintproject/twint/wiki/Setup

   • if you can't install maybe upgrade pip or install visual C++ can fix it https://visualstudio.microsoft.com/visual-cpp-build-tools/

   

2. Train model, I used bi-LSTM train with SST2 dataset

1. input keyword

2. show comparison and top word

   • I'm not train much and because this is binary classification, model may bias on negative, next time we may try over-sampling and hyperparameter tuning for improve model

*bitcoin