Kohonen Self Organizing Feature Map

Description

Implementation of kohonen self organizing feature map to map 3-Dimensional date to 2-Dimensional data.

in this specific problem we implement a som to organize and map a 40 * 40 RGB color map to 2-D map.

We implement this problem in three-phase:

1. Phase 1: Considering Learning-Rate and Neighbouring-Radius as constant. (som.py)
2. Phase 2: Consider Learning-Rate as varying rate. (modified-lr-som.py)
3. Phase 3: Consider Neighbouring-Radius as varying radius. (modified-nr-som.py)

Getting Started

Dependencies

Install dependencies by running following command:
pip install -r requirements.txt

Executing

1. python som.py
2. python modified-lr-som.py
3. python modified-nr-som.py

Result

Result of som.py is something like this:
Left one showing initial color map and right one is organized color map.

Result of modified-lr-som.py is something like this:
Left one showing initial color map and right one is organized color map.

Result of modified-nr-som.py is something like this:
Left one showing initial color map and right one is organized color map.

Authors

Amir Rezaei @ameerezae