This repository contains my project submission for the Visual Perception class, focusing on using Self-Organizing Maps (SOM) to classify patients and healthy subjects based on marker data. This project is structured into three parts:
- Part I: Construction of a Kohonen network for vector classification.
- Part II: Training the network with provided healthy and patient datasets.
- Part III: Testing the network with a new set of data for identification.
Sevillano_Colina_Kimberly_Grace.ipynb: Jupyter Notebook with detailed solutions for all three parts.Pdf_version_notebook.pdf: PDF version of the Jupyter Notebook for better readability.colina.mat,healthy.mat,patient.mat: Dataset files used in the project.healthy.csv,patient.csv,colina.csv: CSV versions of the datasets for easier access and manipulation.SOM.pdf: Additional notes and instructions related to the Self-Organizing Maps used in the project.
The project aims to demonstrate the application of SOMs in distinguishing between patients and healthy subjects using data from markers placed on the subjects. Through this project, I constructed and trained a Kohonen network, adjusting it to achieve optimal classification performance, and tested it with unseen data to verify its predictive capability.
- Ensure you have Jupyter Notebook or JupyterLab installed.
- Clone this repository to your local machine.
- Navigate to the repository directory and launch Jupyter Notebook or Lab.
- Open
Sevillano_Colina_Kimberly_Grace.ipynbto view the project notebook.
- The importance of feature extraction and dimensionality reduction in machine learning.
- Practical application and tuning of Self-Organizing Maps.
- The process of training a machine learning model and testing its performance on unseen data.
I would like to express my gratitude to Elizabeth Thomas for her guidance throughout this project, and the Université de Bourgogne for providing the resources and support necessary to complete this work.