It does this by comparing the prediction faults of the two products in excess of a specific time period. The exam checks the null hypothesis the two models provide the exact same general performance on common, versus the alternative that they do not. If the take a look at statistic exceeds a critical worth, we reject the null speculation, indicatin