Abstract

Some of the basic issues affecting the comparison of population based statistical models and data-centric artificial neural network machine learning models are reviewed in moderate sized data samples. Comparisons of artificial neural networks and population-based models should consider and reflect both the data-centric and probability based nature of the models being compared. This is examined in a series of examples. Some guidelines for developing useful comparative settings are given. Improving the understandability of machine learning methods is an important goal.

This work is licensed under a Creative Commons Attribution 4.0 License.