Usage

1. Create conda environment using provided env file

   $ conda create -f ./conda-environment.yml
   $ source activate rtb
   

2. Download and extract IPinYou data. You may use logs from training2 or training3. Those folders contain extended data compared to training1

   $ cd ./ipinyou_data_dir/training2 && bunzip2 -k clk* && bunzip2 -k imp*
   $ mv *.txt rtb_project_dir/data
   

3. Preprocess data. Script needs to be run from project root.

   $ python preprocess.py -h
   $ python preprocess.py -i './data/imp*' -c './data/clk*' -v
   

4. Evaluate bidding strategies

   $ python ctr_model.py -h
   $ python ctr_model.py