Test Set Sizing Via Random Matrix Theory

Alexander Dubbs

This paper uses techniques from Random Matrix Theory to find the ideal training-testing data split for a simple linear regression with m data points, each an independent n-dimensio…