pyDPP

A python package for sampling from determinantal point processes. Below are instances of sampling from a bicluster and from a random set of points using pyDPP. Refer to examples and references for more information.

Usage

Usage example:

>>> from pydpp.dpp import DPP
>>> import numpy as np
>>> X = np.random.random((10,10))
>>> dpp = DPP(X)
>>> dpp.compute_kernel(kernel_type = 'rbf', sigma= 0.4)               # use 'cos-sim' for cosine similarity
>>> samples = dpp.samples()                   # samples := [1,7,2,5]
>>> ksamlpes = dpp.sample_k(3)                # ksamples := [5,8,0]

Installation

Stable

The stable release can directly be installed by pip:

$ pip install pydpp

Dev

To get the project's source code, clone the github repository:

$ git clone https://github.com/satwik77/pyDPP.git
$ cd pyDPP

Create a virtual environment and activate it. [optional]

$ [sudo] pip install virtualenv
$ virtualenv -p python3 venv
$ source venv/bin/activate
(venv)$

Next, install all the dependencies in the environment.

(venv)$ pip install -r requirements.txt

Install the package into the virtual environment.

(venv)$ python setup.py install

Requirements

• Numpy
• Scipy

To run the example jupyter notebook you need install jupyter notebook, sklearn and matplotlib.

Compatibility

The package has been test with python 2.7 and python 3.5.2

Reference

• Kulesza, A. and Taskar, B., 2011. k-DPPs: Fixed-size determinantal point processes. In Proceedings of the 28th International Conference on Machine Learning (ICML-11) (pp. 1193-1200). [paper]
• Kulesza, A. and Taskar, B., 2012. Determinantal point processes for machine learning. Foundations and Trends® in Machine Learning, 5(2–3), pp.123-286. [paper]