Basic idea: word2vec embedding + CNN, here's the paper
This model (trained for 3 epochs) yields AUC = 0.96823 (on test data).
Ensemble of three convolutional networks (having different number of convolutional layers and feature maps) gives AUC = 0.97310.
- Create a clean environment
conda create --name env python=3.9
conda activate env
- Install dependencies
pip3 install -r requirements.txt
- Download data
- Prepare data
python3 ./code/process.py
- Train
python3 ./code/train.py