Coupon acceptance prediction

Introduction

The main goal of this code is to make a predictive model for predicting whether someone will or will not accept coupons. The data used is a tabular data containing different factors during coupon offering, and whether the coupons were accepted or not.

Requirements

The required packages can be found in config/env_files/coupon_env.yml. Dependencies could be installed by running:

conda env create -f config/env_files/coupon_env.yml

Configuration

The experiments are run according to configurations. The config files for those can be found in config/config_files. Configurations can be based on each other. This way the code will use the parameters of the specified base config and only the newly specified parameters will be overwritten.

The base config file is base.yaml. A hpo example can be found in base_hpo.yaml which is based on base.yaml and does hyperparameter optimization only on the specified parameters. An example for test based on base.yaml can be found in test.yaml.

Arguments

The code should be run with arguments:

--id_tag specifies the name under the config where the results will be saved
--config specifies the config name to use (eg. config "base" for config/config_files/base.yaml)
--mode can be 'train', 'val', 'test' or 'hpo'

Required data

The required data's path should be specified inside the config file like:

data:
  params:
  dataset_path: 'dataset'
  filename: 'coupon_data' \

During train and hpo the raw data should be under {dataset_path}/{filename}, or there should be a train and val csv under dataset_path
During test the file should be under {dataset_path}/{filename}.

Saving and loading experiment

The save folder for the experiment outputs can be set in the config file like:

id: "base"
env:
  result_dir: 'results'

All the experiment will be saved under the given results dir: {result_dir}/{config_id}/{id_tag arg}

1. (feature importance png if decision trees are used)
2. accuracy txt
3. the best model
4. (val/test preds csv if it is set to be saved)

If the result dir already exists and contains a model file then the experiment will automatically resume (either resume the training or use the trained model for inference.)

Usage

Training

To train the model use. During training besides the best model, the feature inmortances are saved (id the used model is a decision tree):

python run.py --config base --mode train

Eval

For eval the results dir ({result_dir}/{config_id}/{id_tag arg}) should contain a model as model_best.pth.tar or model.pkl. During eval the validation file will be inferenced and the accuracy will be calculated.

python run.py --config base --mode val

Test

For test the results dir ({result_dir}/{config_id}/{id_tag arg}) should contain a model as model_best.pth.tar or model.pkl. During test the predictions will be saved.

python run.py --config test --mode test

HPO

For hpo use:

python run.py --config base_hpo --mode hpo

Results:

random forest

HPO best result: 73.3%

Feature importance:

 

xgboost

HPO best result: 75.0%

Feature importance:

 

feedforward network with category embedding

HPO best result: 68.1%

 

TabNet

HPO best result: 66.2%