Families of discrete circular distributions with some novel applications

TL;DR

This paper introduces new classes of discrete circular distributions motivated by modern circular data, providing four construction methods, analyzing their properties, and demonstrating applications in roulette, smart home data, and changepoint detection.

Contribution

The paper develops four novel methods for constructing discrete circular distributions and explores their properties, especially focusing on marginalized and conditionalized approaches, with practical applications.

Findings

Inherited properties from continuous distributions like von Mises and wrapped Cauchy.

Effective methods for changepoint detection and mixture modeling in circular data.

Applications to roulette, smart home data, and testing for uniformity and serial correlation.

Abstract

Motivated by some cutting edge circular data such as from Smart Home technologies and roulette spins from online and casino, we construct some new rich classes of discrete distributions on the circle. We give four new general methods of construction, namely (i) maximum entropy, (ii) centered wrapping, (iii) marginalized and (iv) conditionalized methods. We motivate these methods on the line and then work on the circular case and provide some properties to gain insight into these constructions. We mainly focus on the last two methods (iii) and (iv) in the context of circular location families, as they are amenable to general methodology. We show that the marginalized and conditionalized discrete circular location families inherit important properties from their parent continuous families. In particular, for the von Mises and wrapped Cauchy as the parent distribution, we examine their properties including the maximum likelihood estimators, the hypothesis test for uniformity and give a test of serial independence. Using our discrete circular distributions, we demonstrate how to determine changepoint when the data arise in a sequence and how to fit mixtures of this distribution. Illustrative examples are given which triggered the work. For example, for roulette data, we test for uniformity (unbiasedness) , test for serial correlation, detect changepoint in streaming roulette-spins data, and fit mixtures. We analyse a smart home data using our mixtures. We examine the effect of ignoring discreteness of the underlying population, and discuss marginalized versus conditionalized approaches. We give various extensions of the families with skewness and kurtosis, to those supported on an irregular lattice, and discuss potential extension to general manifolds by showing a construction on the torus