convpy

Library for lagged conversion rate estimation.

Based on the paper Chapelle: "Modeling Delayed Feedback in Display Advertising".

This is a Python 3 library.

Setup

Create conda environment:

$ conda create --name convpy python=3 -y

Activate the conda environment:

$ source activate convpy

Clone convpy repository to current directory:

$ git clone git@github.com:markovianhq/convpy.git

Install all requirements:

$ conda install --file requirements.txt --yes
$ pip install -r requirements_pip.txt

Now install the convpy package:

$ python setup.py develop

Testing

Install extra packages required for testing:

$ pip install -r requirements_pip_test.txt --yes

Run tests:

$ pytest convpy --flake8