https://github.com/zenklinov/clustering_k-means_metrics_pca
Comparing Euclidean Distance, Manhattan Distance, Cosine Distance, with PCA in K-Means Clustering
https://github.com/zenklinov/clustering_k-means_metrics_pca
cluster kmeans-clustering metrics pca unsupervised-machine-learning
Last synced: 4 months ago
JSON representation
Comparing Euclidean Distance, Manhattan Distance, Cosine Distance, with PCA in K-Means Clustering
- Host: GitHub
- URL: https://github.com/zenklinov/clustering_k-means_metrics_pca
- Owner: zenklinov
- License: cc0-1.0
- Created: 2024-04-06T23:29:45.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-02-24T07:15:51.000Z (4 months ago)
- Last Synced: 2025-02-24T08:28:04.175Z (4 months ago)
- Topics: cluster, kmeans-clustering, metrics, pca, unsupervised-machine-learning
- Language: Jupyter Notebook
- Homepage:
- Size: 479 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# K-Means Clustering: Comparing Metrics with PCA
This repository contains a Jupyter Notebook (`K-Means-comparing-metrics-with-pca-2-unsupervised-machine-learning.ipynb`) that demonstrates the process of performing K-Means clustering on the Iris dataset using different distance metrics (Euclidean, Manhattan, Cosine) and evaluating the clustering results with Principal Component Analysis (PCA).
## Content
### 1. Importing Dataset (Iris)
The analysis starts with importing the Iris dataset using the `load_iris()` function from scikit-learn. Information about the dataset including feature names, target names, and data shape is displayed.
### 2. Performing K-Means Clustering
K-Means clustering is performed using scikit-learn's `KMeans` class with 3 clusters. The cluster centers and labels are obtained.
### 3. Computing Distances
Euclidean, Manhattan, and Cosine distances between data points and cluster centers are computed using scikit-learn's `euclidean_distances`, `manhattan_distances`, and `cosine_distances` functions.
### 4. Dimensionality Reduction with PCA
Principal Component Analysis (PCA) is applied to reduce the dimensionality of the dataset to 2 components.
### 5. Visualization of Clustering Results
Scatter plots are created to visualize the clustering results along with the distances using PCA. Each subplot represents a different distance metric (Euclidean, Manhattan, Cosine) with corresponding silhouette score displayed in the title.
### 6. Analysis of Inter-Data Distance
An analysis is provided based on the scatter plots and silhouette scores for each distance metric. The correlation between distance metrics and principal component 1 is discussed, along with insights into clustering quality.
### 7. Conclusion
The conclusion summarizes the findings of the analysis, including the assessment of clustering quality for each distance metric and recommendations for further analysis.
## Usage
To run the notebook locally, clone the repository and open the `K-Means-comparing-metrics-with-pca.ipynb` file in Jupyter Notebook or JupyterLab.
## Dependencies
- scikit-learn
- matplotlib
## Author
[zenklinov](https://github.com/zenklinov/)