Compares the accuracy and hypothetical returns of multiple machine-learning model types and time periods for a chosen crypto pair to determine which will yield the best results, with the intent of informing and optimizing crypto trading algorithms