Due to storage issue, I did NOT put the data file into this repository.
Please download from https://snap.stanford.edu/data/bigdata/amazon/amazon-meta.txt.gz
Then upzip it, put in the 'data/' folder as 'data/amazon_meta.txt'.
Some python packages need to be installed, including numpy, pandas, scipy, multiprocessing and cvxopt.
$ bash manifest.sh
├── amazon_data_util.py some util files
├── baseline_random.py random select k products to bundle
├── data
│ ├── amazon-meta.txt Amazon product co-purchasing dataset (from SNAP, Stanford)
│ └── amazon_price.json prices of some products on Amazon (crawled from thetracktor.com)
├── determine_f.py determine the parameter a in f(.)
├── exp_setting.py bundle size settings
├── kbundle.py Our bundling algorithm (Algorithm 3)
├── manifest.sh !!! run all the experiments
├── n01_products_raw.py
├── n02_products_copurchase_select.py
├── n03_copurchase_prob.py
├── n04_empirical_mean_std.py
├── n05_empirical_cor.py
├── n06_smf.py
├── n07_accuracy.py
├── n08_run_kbundle.py
├── n09_approximation_ratio.py
├── pi.py $\pi(.)$ function
├── profit.py calculate the profit of a bundling strategy
└── README