Online Program Home
My Program

Abstract Details

Activity Number: 660 - Shrinkage Methods for Analyzing Complex Business Data
Type: Topic Contributed
Date/Time: Thursday, August 2, 2018 : 10:30 AM to 12:20 PM
Sponsor: Business and Economic Statistics Section
Abstract #328913
Title: Using Shrinkage to Detect Changes in Variance in Complex Business Data
Author(s): Rebecca Killick* and Jamie-Leigh Chapman and Idris Eckley
Companies: Lancaster University and Lancaster University and Lancaster University
Keywords: telematics; wavelet; locally stationary; changepoint

A nonparametric method of detecting changes in variance is developed for the case where assumptions of normality and independence are not appropriate. This may be the case in applications in complex business data such as financial returns, consumer data, and energy use. We use the Locally Stationary Wavelet (LSW) model to provide a local estimate of the variance of a time series. The wavelet transformation identifies changes in variance in a fully non-parametric setting. Shrinkage methods are used to determine the number of changepoints in a given data set. We will demonstrate the efficacy of our approach through simulations where we compare against the most commonly used approaches in a variety of settings. The proposed method performs particularly well in cases where the data contains outliers, heavy tails and/or autocorrelation. The wavelet methodology is very fast compared to other approaches where an ARMA model is used to account for the autocorrelation in a series. We demonstrate this on complex business data including telematics data from the insurance industry.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program