Premium
Posterior singular spectrum analysis
Author(s) -
Holmström Lasse,
Laun Ilkka
Publication year - 2013
Publication title -
statistical analysis and data mining: the asa data science journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.381
H-Index - 33
eISSN - 1932-1872
pISSN - 1932-1864
DOI - 10.1002/sam.11195
Subject(s) - singular spectrum analysis , computer science , mathematics , pattern recognition (psychology) , artificial intelligence , singular value decomposition
A method is proposed for finding interesting underlying features of a time series, such as trends, maxima, minima, and oscillations. A combination of singular spectrum analysis (SSA) and Bayesian modeling is used where the credibility of SSA signal components is analyzed via posterior simulation. The potential of the technique is demonstrated using artificial and real data examples. Our analysis of a Bayesian reconstruction of post‐Ice Age temperature variation lends support for the presence of oscillations detected in previous studies of the paleoclimate. © 2013 Wiley Periodicals, Inc. Statistical Analysis and Data Mining, 2013