Quantization-based clustering algorithm

• This repository contains implementation of quantization based clustering algorithm proposed in this paper.
• Also contains experiments which compare it to another implementations of algorithms from K-Means family from sci-kit learn library.

Directory structure

qbca.py

• contains implementation of the proposed quantization-based clustering algorithm

max_heap.py

• simple implementation of max heap needed by the algorithm

test_gaussian.py

• compares QBCA with another algorithms using external indices on generated Gaussian data

test_gaussian_internal.py

• compares QBCA with another algorithms using internal indices on generated Gaussian data

explore.R

• exploring external indices using R

test_image.py

• compares QBCA with another algorithms using image segmentation on sample image "test_image.jpg"

test_iris.py

• compares QBCA with another algorithms using external indices on famous IRIS dataset

test_iris_external.py

• compares QBCA with another algorithms using internal indices on famous IRIS dataset

visualize_gaussian_data.py

• visualizing gaussian data that are used for experiments and generated

visualize_gaussian_results.py

• visualize external indices comparing QBCA to another algorithms

figures/

• contain figures that were generated for comparing algorithms and used in the report

results/

• contains CSV files containing results frm external indices comparing performance of different algorithms

screenshots/

• screenshots used in the paper when describing code optimalization