Stock Analysis - Find key levels of pricing using Clustering Models

Author: Poongodi P

Description:

This project aims at finding the key levels of pricing that group with each other for all the available tickers individually.

Input:

CSV file with approx 20K rows

Models used:

1. KMeans clustering (Centroid based)
2. DBSCAN clustering (Density Based)
3. Agglomerative Clustering (Hierarchial)

Evaluation metrics:

Silhouette score

Elbow method to find optimum number of clusters