This project performs unsupervised clustering of time-series segments from the PulseDB dataset using divide-and-conquer algorithms.
It includes:

• Recursive clustering based on correlation similarity
• Closest-pair detection within each cluster
• Maximum subarray detection (Kadane’s Algorithm) to find active intervals

Files:

• main.py — main program containing all functions
• /out/ — folder where generated plots and CSV outputs are saved

Dataset Dataset used: VitalDB_AAMI_Test_Subset.mat (from the PulseDB Balanced Training and Testing dataset on Kaggle)

Requirements: Before running the program, make sure you install these libraries:

pip install numpy pip install matplotlib pip install h5py