hashihab / mdcgenpy

mdcgenpy

mdcgenpy is a Multidimensional Dataset for Clustering Generator. This tool is aimed at researchers looking for synthetic datasets, in particular for testing clustering algorithms. A variety of customization options are available, in order to allow for a wide range of use cases.

Using the generator is simple, and can even be used without parameters:

import mdcgenpy

# Initialize cluster generator (all parameters are optional)
cluster_gen = mdcgenpy.clusters.ClusterGenerator()

# Get tuple with a numpy array with samples and another with labels
data = cluster_gen.generate_data()

Installation

The package can be installed with pip:

$ pip install git+https://github.com/CN-TU/mdcgenpy

Generating data outside Python

It is also possible to use mdcgenpy without knowing python.

To do this, you just need to give as input a JSON file (check specification details here). Using the mdcgenpy.py script, the output will be sent in CSV format to stdout.

Example:

$ ./mdcgenpy.py input_parameters.json > output.csv

Features

• Efficient code, compatible with Python 2 and Python 3.
• Various possible distributions for the clusters are available out-of-the-box, and custom distributions are also allowed.
• Parameters allow for control over the overlap of the clusters, outliers, noise, correlation inside each cluster, etc.

Documentation

Documentation for this project can be found at https://mdcgenpy.readthedocs.io.