Constructed Analogs and Linear Regression

AbstractThe constructed analog procedure produces a statistical forecast that is a linear combination of past predictand values. The weights used to form the linear combination depend on the current predictor value and are chosen so that the linear combination of past predictor values approximates the current predictor value. The properties of the constructed analog method have previously been described as being distinct from those of linear regression. However, here the authors show that standard implementations of the constructed analog method give forecasts that are identical to linear regression forecasts. A consequence of this equivalence is that constructed analog forecasts based on many predictors tend to suffer from overfitting just as in linear regression. Differences between linear regression and constructed analog forecasts only result from implementation choices, especially ones related to the preparation and truncation of data. Two particular constructed analog implementations are shown to co...

[1]  P. Hansen.  Rank-Deficient and Discrete Ill-Posed Problems: Numerical Aspects of Linear Inversion , 1987 .

[2]  H. V. D. Dool,et al.  Searching for analogues, how long must we wait? , 1994 .

[3]  Thomas M. Smith,et al.  Improved Extended Reconstruction of SST (1854–1997) , 2004 .

[4]  E. Hawkins,et al.  Evaluating the potential for statistical decadal predictions of sea surface temperatures with a perfect model approach , 2011 .

[5]  Huug van den Dool,et al.  Performance and analysis of the constructed analogue method applied to U.S. soil moisture over 1981-2001 , 2003 .

[6]  Timothy DelSole,et al.  A Bayesian Framework for Multimodel Regression , 2007 .

[7]  A. Barnston,et al.  Skill of Real-Time Seasonal ENSO Model Predictions During 2002–11: Is Our Capability Increasing? , 2012 .

[8]  E. Maurer,et al.  Utility of daily vs. monthly large-scale climate data: an intercomparison of two statistical downscaling methods , 2007 .

[9]  Timothy DelSole,et al.  Predictable Component Analysis, Canonical Correlation Analysis, and Autoregressive Models , 2003 .

[10]  J. Storch,et al.  A Simple Empirical Model for Decadal Climate Prediction , 2011 .

[11]  Robert Tibshirani,et al.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction, 2nd Edition , 2001, Springer Series in Statistics.

[12]  Balaji Rajagopalan,et al.  Analyses of global sea surface temperature 1856–1991 , 1998 .

[13]  M. Tippett,et al.  Downscaling of Seasonal Rainfall over the Philippines: Dynamical versus Statistical Approaches , 2012 .

[14]  Cécile Penland,et al.  Prediction of Nino 3 sea surface temperatures using linear inverse modeling , 1993 .