A model sufficiency test using permutation entropy

Abstract

Using the ordinal-pattern concept in permutation entropy, we propose a model sufficiency test to study a given model's point prediction accuracy. Compared with some classical model sufficiency tests, such as Broock et al.'s (1996) test, our proposal does not require a sufficient model to eliminate all structures exhibited in the estimated residuals. When the innovations in the investigated data's underlying dynamics show a certain structure, such as higher moment serial dependence, Broock et al.'s (1996) test can lead to erroneous conclusions about the sufficiency of point predictors. Due to the structured innovations, inconsistency between the model sufficiency tests and prediction accuracy criteria can occur. Our proposal fills in this incoherence between model and prediction evaluation approaches and remains valid when the underlying process has nonwhite additive innovation.

On the Genuine Relevance of the Data-Driven Signal Decomposition-Based Multiscale Permutation Entropy

“…Ordinal pattern-based approaches have recently gained attention because of their natural and efficient way to transform a time series into a sequence of symbols with a finite alphabet [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 ]. This ordinal pattern sequence is simply obtained by comparing the values of a finite number of neighboring samples and ranking them.…”