Some Recent Developments in Change-Point Detection Using Integral Transforms

Abstract

Change-point inference has many real-world applications, including finance, medicine, algorithmic trading, and other domains. The integral transform method has recently gained popularity, particularly in the context of analyzing complex data structures. In this study, we introduce novel statistical tests for detecting change-points in the distribution of a sequence of independent observations of various types. These new tests are based on integral transforms and provide a practical and consistent way of identifying distributional changes. We focus on recommending tailored solutions for real-world use with the aim of making the methods practical and accessible. In addition, we discuss the possible limitations of the proposed method, especially in comparison to some other existing methodologies. This includes considerations of performance, applicability, and potential constraints in specific scenarios.