Ordinal Pareto seasonality in Walsh coefficients

Abstract We present a statistical methodology for investigating oscillations in the underlying number of causes contributing to a phenomenon. We utilize the Walsh transform of a ternary signal derived from the unbiased MLEs of the shape parameters in an ordinal indexed series of Pareto distributions. To this end, we outline the details of our methodology, discuss a simulation study, and apply our method to real world data (NHTSA recall data). Our results indicate that the proposed methodology is able to reliably detect oscillations in the shape parameter.