python-kkmeans 0.1.0

An implementation of the Kernel k-means algorithm for a bachelorthesis.

kkmeans

Repository: https://git01lab.cs.univie.ac.at/osterj97/oster-bachelor-thesis

Implementation of the Kernel k-means clustering algorithm for multiple dimensions (including as well one-dimensional data) in Python. The results are visualized in plots and measured with quality metrics (e.g. Silhouette Coefficient). The algorithm can be used as a framework.

https://sites.google.com/site/dataclusteringalgorithms/kernel-k-means-clustering-algorithm

The related bachelor thesis can be found here: https://www.overleaf.com/read/shnzrthjgqtw (read only) or as an exported .pdf here Bachelor_Thesis____Kernel_k_means_clustering_framework_in_Python.pdf

Installation

pip install python-kkmeans

The package can also be installed directly from this repository using pip install git+ssh://[e-mail]/osterj97/oster-bachelor-thesis.git.

Quickstart

from sklearn.datasets import make_blobs
from sklearn.preprocessing import StandardScaler
import matplotlib.pyplot as plt

from kkmeans import kkmeans

n_clusters = 3
X, _ = make_blobs(n_samples=100, n_features=3, centers=n_clusters)
X_scaled = StandardScaler().fit_transform(X)

cluster_assignments = kkmeans(X, n_clusters=n_clusters)
plt.scatter(X_scaled[:,0], X_scaled[:,1], c=cluster_assignments)

Examples

Take a look at the examples folder to see benchmarking and plotted examples of the algorithm in action. The playground notebook should give you a comprehensive overview of the evaluation of the algorithm. It can be run after installing jupyter and running jupyter notebook from this folder.

Local Development

The following commands assume you run on a Linux or Mac, for Windows instructions take a look at the official documentations.

• (optional) Create a venv python3 -m venv .venv
• (optional) Active venv source .venv/bin/activate
• (optional) Update pip pip install --upgrade pip
• Install all required packages pip install -r requirements.txt

Test

Testing is set up using pytest. Run all tests running the command pytest in the root directory. For detailed description on pytest see: Full pytest documentation