recsys-challenge-2019

An Attentive RNN Model for Session-based and Context-aware Recommendations

7th place solution @ RecSys 2019 Data Challenge.

Setup instructions

1. Clone this repo:

   git clone https://github.com/hugoguh/saca_recsys

2. Copy/move/download the competion data files to ./data/

3. Install Anaconda (Python 3 version).

4. Install the environment:

   cd recsys-challenge-2019/
   conda env create --file environment_recsys-challenge-2019.yml

5. Activate the environment:

   conda activate recsys-challenge-2019

6. Run all features and model:

   cd scripts
   sh run_all_future.sh

File structure

```.
├── LICENSE
├── README.md
├── data
├── environment_recsys-challenge-2019.yml
├── features
├── helpers
│   ├── __init__.py
│   ├── feature_helpers.py
│   └── train_val_split_helpers.py
├── models
├── predictions
└── scripts
    ├── 001_Preprocess_Train_Test_split.py
    ├── 011_Features_Items.py
    ├── 012_Features_CTR.py
    ├── 013_Features_Dwell.py
    ├── 014_Features_General_01.py
    ├── 015_Features_General_02.py
    ├── 021_Run_Model.py
    ├── run_all_default.sh
    ├── run_all_future.sh
    ├── run_features_default.sh
    └── run_features_future.sh
```