Clustering Algorithms Comparison

This repository provides a framework for clustering analysis using K-Means and DBSCAN algorithms. The project demonstrates the effect of data standardization on clustering performance and includes visualizations for deeper insights into the results.

---

Project Overview

This project evaluates the performance of clustering algorithms on standardized and non-standardized datasets. Key objectives include:

1. Clustering Analysis: Compare the performance of K-Means and DBSCAN.
2. Impact of Standardization: Highlight how standardizing the data affects clustering results.
3. Visualizations: Generate plots to visualize clusters and performance metrics.

---

Datasets

1. pricerunner_aggregate 2.csv

• Contains product and merchant information used for clustering.
• Key Columns:
  • Product ID
  • Merchant ID

2. Clustered Data Outputs

• non-standardize_k-means_clustered_data.csv: Results of K-Means clustering without standardization.
• standardize_k-means_clustered_data.csv: Results of K-Means clustering with standardization.
• dbscan_clustered_data.csv: Results of DBSCAN clustering.

---

Python Scripts

1. non-standardize_k-means.py

• Performs K-Means clustering without standardizing the data.
• Outputs clustering metrics and visualizations.

2. standardize_k-means.py

• Performs K-Means clustering on standardized data.
• Highlights the impact of standardization on clustering.

3. dbscan.py

• Implements DBSCAN clustering algorithm.
• Includes clustering metrics and visualizations for non-standardized data.

---

Performance Metrics

1. Silhouette Score

• Measures the similarity of points within the same cluster.
• Range: -1 (poor clustering) to 1 (ideal clustering).

2. Davies-Bouldin Index

• Evaluates the compactness and separation of clusters.
• Lower values indicate better clustering performance.

---

Visualizations

Key Visualizations:

1. K-Means vs DBSCAN Performance (Non-Standardized):

2. Impact of Standardization on K-Means:

3. K-Means vs DBSCAN Performance (Standardized):

---

Installation

Step 1: Clone the Repository

git clone https://github.com/<your_username>/<repository_name>.git
cd <repository_name>

Step 2: Install Dependencies

Install the required Python libraries:

pip install -r requirements.txt

---

How to Run

1. Running Non-Standardized K-Means

• Execute the script:

  python non-standardize_k-means.py

• Ensure the input file path and output file path are correctly set within the script:

  file_path = 'path_to_your_file'  # Replace with the path to your input dataset
  output_file_path = 'path_to_output_file'  # Replace with the path where the results will be saved

2. Running Standardized K-Means

• Execute the script:

  python standardize_k-means.py

• Ensure the input file path and output file path are correctly set within the script:

  file_path = 'path_to_your_file'  # Replace with the path to your input dataset
  output_file_path = 'path_to_output_file'  # Replace with the path where the results will be saved

3. Running DBSCAN

• Execute the script:

  python dbscan.py

• Ensure the input file path and output file path are correctly set within the script:

  file_path = 'path_to_your_file'  # Replace with the path to your input dataset
  output_file_path = 'path_to_output_file'  # Replace with the path where the results will be saved

---

Project Structure

project_directory/
├── images/
│   ├── non-standardize_k-means_vs_dbscan.png
│   ├── non-standardize_vs_standardize.png
│   ├── standardize_k-means_vs_dbscan.png
│   ├── dbscan.png
│   ├── k-means_non-standardized.png
│   ├── k-means_standardize.png
├── data/
│   ├── pricerunner_aggregate 2.csv
│   ├── non-standardize_k-means_clustered_data.csv
│   ├── standardize_k-means_clustered_data.csv
│   ├── dbscan_clustered_data.csv
├── non-standardize_k-means.py
├── standardize_k-means.py
├── dbscan.py
├── requirements.txt
└── README.md

---

Key Notes

1. The repository contains a requirements.txt file listing all necessary Python libraries. Make sure to install the dependencies before running any scripts.
2. Visualizations generated by the scripts are automatically saved in the images/ folder.
3. Replace all instances of 'path_to_your_file' and 'path_to_output_file' in the scripts with the correct file paths for your dataset and desired output.

---