project-7: Unsupervised Learning (K-Means & Hierarchical Clustering)

Overview

This project focuses on clustering techniques (K-Means and Hierarchical Clustering) applied to stock market data. By grouping stocks exhibiting similar characteristics, one can make more informed decisions about allocating investments and diversifying portfolios.

Project Structure

• USL_Project_LearnerNotebook_FullCode-1.ipynb: Jupyter Notebook with detailed exploratory data analysis, data preprocessing, and clustering implementation steps.
• stock_data.csv: Dataset containing stock data (current price, volatility, financial ratios, etc.) from several NYSE-listed companies.
• README.md: Documentation on project purpose, requirements, and usage guidelines.

Getting Started

1. Install dependencies:

   • Python 3.x
   • pandas, numpy, matplotlib, seaborn, scikit-learn, scipy
   • yellowbrick

2. Set up environment (example with pip):

   pip install pandas numpy matplotlib seaborn scikit-learn scipy yellowbrick

3. Run the Notebook:

   1. Open the Jupyter Notebook (USL_Project_LearnerNotebook_FullCode-1.ipynb).
   2. Restart the kernel and run all cells sequentially.

Methodology

1. Data Loading: Reads and cleans the stock dataset.
2. Exploratory Analysis: Uses summary statistics and plots to explore data distribution.
3. Preprocessing: Standardizes and prepares data for clustering.
4. K-Means Clustering: Determines the optimal number of clusters using the elbow method and silhouette analysis.
5. Hierarchical Clustering: Explores dendrograms to select the best linkage and clusters.

Results & Insights

• Stocks are grouped into clusters based on similarities in financial metrics, which aids in creating a balanced and diversified portfolio.
• K-Means provides direct cluster assignments, while Hierarchical Clustering offers a visual representation (dendrogram) of how stocks group together across different levels of similarity.

Contributing

Contributions in the form of issue reports or suggestions are welcome.