Automatic channel selection for seizure detection.

This project aims to improve computational efficiency of the winning solution by Michael Hills (https://github.com/MichaelHills/seizure-detection) for UPenn and Mayo Clinic's Seizure Detection Challenge on Kaggle (https://www.kaggle.com/c/seizure-detection). By using our automatic channel selection engine, the run-time seizure detection is 2x faster than Hills approach.