K-Means Clustering with Random Data Generation

---

1. Project Title

K-Means Clustering with Random Data Generation

---

2. Problem Statement and Goal of Project

The goal of this project is to demonstrate the K-Means clustering algorithm by applying it to synthetically generated data. The dataset is created randomly with multiple centers, and the project aims to:

• Illustrate the effectiveness of K-Means in clustering random data.
• Visualize the results and understand how the algorithm clusters the data.
• Experiment with the impact of different configurations of the K-Means algorithm.

The focus is on understanding the clustering mechanism in a controlled setting using artificial data, allowing for easy visualization and evaluation.

---

3. Solution Approach

The notebook follows an end-to-end workflow:

• Data Generation: Using the make_blobs function from sklearn.datasets, random data with multiple centers is created.
• Clustering: K-Means clustering is applied to the generated data with various numbers of clusters.
• Visualization: The data and the results of clustering are visualized using matplotlib.
• Evaluation: Although no explicit evaluation metric is mentioned, the visual separation of clusters provides insight into the performance of the algorithm.

---

4. Technologies & Libraries

• Python – Programming language used
• NumPy – Numerical computation
• Matplotlib – Data visualization
• Scikit-learn – Machine learning library (KMeans, make_blobs)
• Random – Python standard library for generating random numbers

---

5. Description about Dataset

• Source: Generated using make_blobs from scikit-learn.

• Data Structure:

  • n_samples=5000: Total number of data points.
  • n_features=2: Each data point has 2 features (2D).
  • centers: Randomly placed at multiple centers: [[2,3],[5,6],[6,-2],[-3,6],[8,9]].
  • cluster_std=1: Standard deviation for the clusters.
  • random_state=101: Ensures reproducibility of the random data generation.

---

6. Installation & Execution Guide

Prerequisites:

• Python 3.x
• Jupyter Notebook or JupyterLab

Clone the repository:

git clone <repository_url>
cd <repository_directory>

Install required dependencies:

pip install numpy matplotlib scikit-learn

Run the notebook:

jupyter notebook K-Means_random.ipynb

---

7. Key Results / Performance

• The project explores how the K-Means algorithm clusters randomly generated data into distinct groups.
• The results can be visualized through scatter plots where each cluster is color-coded.
• The performance of the algorithm is primarily assessed through visual inspection, highlighting the ability of K-Means to identify and separate clusters based on the data.

---

8. Screenshots / Sample Outputs

• Cluster Visualization: Scatter plots showing the data points and their corresponding cluster assignments.
• Centroid Markers: The centers of each cluster are marked using a red "X" to visualize the clustering result.

💡 Some interactive outputs (e.g., plots, widgets) may not display correctly on GitHub. If so, please view this notebook via nbviewer.org for full rendering.

---

9. Additional Learnings / Reflections

• Experimented with the K-Means algorithm using random data, gaining an understanding of its behavior with different numbers of clusters.
• The visualization approach is key for assessing clustering performance when the data is generated synthetically.
• The project also highlights the importance of tuning K-Means parameters like n_init, which controls the number of initializations for finding the best clustering solution.

---

👤 Author

Mehran Asgari Email: imehranasgari@gmail.com GitHub: https://github.com/imehranasgari

---

📄 License

This project is licensed under the Apache 2.0 License – see the LICENSE file for details.

---

💡 Some interactive outputs (e.g., plots, widgets) may not display correctly on GitHub. If so, please view this notebook via nbviewer.org for full rendering.

---