CausalInferenceRecSys

Project Structure

• coat/ Folder containing the original COAT dataset.

• coat_sg/ Possibly a single-group variant or preprocessed version of COAT data.

• ml-100k/ MovieLens 100k dataset directory.

• check_point_coat/, check_point_100k/ Directories storing model checkpoints for the COAT and MovieLens-100k datasets, respectively.

• dataloader/ Custom dataloaders for processing COAT and MovieLens datasets.

• coatOg.py Script for training and evaluating the original (OG) deconfounded matrix factorization model on COAT.

• trainOg.py General training script for deconfounded MF across different datasets.

• ogModel.py Implementation of the original deconfounded MF model, including VPF (Variational Poisson Factorization).

• coatVae.py Training script for the VAE-based deconfounded recommendation model on COAT.

• trainVae.py General training script for VAE-based deconfounded MF.

• vaeModel.py Implementation of the VAE model for deconfounded recommendation.

• coatNeu.py Training script for NeuMF (Neural Matrix Factorization) on COAT.

• trainNeu.py General training script for NeuMF-based deconfounded recommendation.

• neuModel.py Implementation of the NeuMF model.

• __pycache__/ Python cache directory for compiled bytecode.