Unsupervised Learning for Clustering Car and Truck Images

Overview

This GitHub repository contains the code and documentation for a project applying unsupervised learning techniques, particularly clustering, on a Kaggle dataset containing images of cars and trucks. The primary goal is to explore and implement clustering algorithms to group similar images without using labeled examples. This project serves as the author's first venture into unsupervised learning, with a focus on clustering as an introductory task.

Keywords

• Clustering
• Unsupervised Learning
• AI
• Computer Vision

Table of Contents

1. Introduction

   • Brief overview of unsupervised learning and its applications.
   • Related studies highlighting the importance of clustering in various domains.

2. State of the Art

   • Overview of current advancements in machine learning, with a focus on agglomerative clustering.
   • Detailed explanation of agglomerative clustering, including key aspects and considerations.
   • Introduction to Random Forests and Isolation Forests, discussing their applications in regression, classification, and anomaly detection.

3. Experiment

   • Motivation for the theme, choice of programming language, and libraries used.
   • Detailed information on the dataset, data preprocessing steps, and initial data analysis.
   • Implementation details for agglomerative clustering and isolation forests.
   • Visualization of the clustering results and anomaly detection.

4. Conclusions

   • Summary of findings.
   • Challenges encountered during the experiment.

5. Future Work

   • Proposed ideas for future experiments and improvements.

6. References

   • Citations for relevant studies and tools used in the project.