Unsupervise_Learning_Method

======GEOL0069 AI4EO======

Introduction

This repository contains resources and examples on Unsupervised Learning Methods. It aims to introduce and apply unsupervised learning techniques to Earth observation data.

Learning Objectives

• Understand the basics of unsupervised learning methods.
• Learn to apply K-means Clustering and Gaussian Mixture Models to classify Earth observation data.
• Classify the echoes in leads and sea ice and produce an average echo shape as well as standard deviation for these two classes.

Key Concepts

• Unsupervised Learning: Techniques used when training data is not labeled.
• K-means Clustering: A method to partition data into K distinct clusters.
• Gaussian Mixture Models: A probabilistic model for representing normally distributed subpopulations within an overall population.
• Plotting pictures: Using code to add legends and titles etc, change colors, specify the line displayed.

Preperation

• Python 3.x
• Jupyter Notebook
• Required Python libraries (numpy, matplotlib, scikit-learn, etc.)

Sections

• Unsupervised Learning: An introductory section that outlines the basics of unsupervised learning, including its significance and applications.

• K-means Clustering: Discusses the K-means algorithm, its key components, and the iterative process. Highlights the advantages and provides a basic code implementation for understanding and application.

• Gaussian Mixture Models (GMM): Introduces GMM, detailing why it's used for clustering, its key components, and the Expectation-Maximization (EM) algorithm. Explains the advantages of GMM and includes basic code implementation.

• Application to Image and Altimetry Classification: Demonstrates the application of K-means and GMM in image classification and altimetry data classification, including necessary code and functions.

How to Use

This section guides users through setting up their environment, downloading datasets, and running example scripts.

Website/software using

• Google Dirve
• Google Colab
• Copernicus Data Space Ecosystem
• Google cloud

Citation

@misc{unsupervised_learning_chapter1, title={Chapter 1: Unsupervised Learning Methods}, author={Lining Chen}, year={2024}, howpublished={Unsupervise_Learning_Method}, url={https://github.com/LiningChen/Unsupervise_Learning_Method.git)https://github.com/LiningChen/Unsupervise_Learning_Method.git} }