Clustering of Time Series

• Activity Number: 131
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Business and Economic Statistics Section
• Author(s): Raja Velu*+
• Companies: Syracuse University
• Keywords: clustering ; time series

Abstract

Methods for clustering static data have been extensively studied in the literature, but interest in clustering chronological data is of recent origin. The key problem appears to be in developing a distance measure that captures various features of time series: level, spread and dependence. We begin with an application in finance followed by another in sponsored search to illustrate where clustering is of interest. We develop some intuitive algorithms and illustrate them with exchange rate data. Some tentative suggestions for applications in large scale are proposed; similarities with co-integration concepts in economics are noted.