dwaydwaydway / Pricing-Optimization

Pricing-Optimization

Environment

pip install -r requirements.txt

or

conda create --name <env_name> --file requirements.txt
conda activate <env_name>

Usage

Prepeocess Data

python src/preprocess.py -c config.yaml

Train a Model for Classification

python src/train_model.py -c config.yaml

Evaluate Pricing

python src/pricing.py -c config.yaml

or you can put all the congiuration files(all the experiments) in the exp2run/ folder, and run

bash src/run_exps.sh

Everything can be configured in config.yaml

To experiment with new solutions, add a new class in preprocessors.py, strategies.py, models.py or pricer.py

• A strategy class should define who the high-level of our algorithm runs
• A preprocessor class should define how to acquire the embeddings for the cold-starting users
• A model class should define how the predictions are made
• A pricer class should define how the to acquire the prices that optimize the revenue

Experiment

After running the script, a folder with the name of the exp_name you define in config.yaml will be automatically created inside exp/. A copy of the config.yaml you run with and the experiment result will be in it.

TODO

• Experiment with different model
• Start working on part 2
• Buy milk. I ran out of milk yesterday.