PROBLEMS OF NONNORMALITY AND MULTICOLLINEARITY FOR FORECASTING METHODS BASED ON LEAST SQUARES.

Abstract Many popular forecasting and time-series analysis methods assume that the variable to be forecast can be expressed as a linear function of a set of predictors. The predictors may include variables related in either a correlative or causal fashion to the response variable, lagged values of this variable, or known mathematical functions of time. The method of least squares is used almost exclusively to estimate the parameters in these models. This paper discusses two hazards in the indiscriminant use of least squares; nonnormality of the observations on the variable of interest and multicollinearity among the predictors. Robust estimation methods are suggested as alternatives to least squares for nonnormal data, and a robust version of exponential smoothing is developed. A small Monte Carlo study indicates that the robust procedure can be superior to ordinary exponential smoothing in many situations. The sources and effects of multicollinearity are discussed, and several diagnostic statistics are p...

[1]  David A. Belsley,et al.  Regression Analysis and its Application: A Data-Oriented Approach.@@@Applied Linear Regression.@@@Regression Diagnostics: Identifying Influential Data and Sources of Collinearity , 1981 .

[2]  Robert Fildes,et al.  Quantitative Forecasting — The State of the Art: Extrapolative Models , 1979 .

[3]  P. J. Huber.  Robust Estimation of a Location Parameter , 1964 .

[4]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1972 .

[5]  A. E. Hoerl,et al.  Ridge regression iterative estimation of the biasing parameter , 1976 .

[6]  D. F. Andrews,et al.  Robust Estimates of Location , 1972 .

[7]  R. R. Hocking,et al.  A Class of Biased Estimators in Linear Regression , 1976 .

[8]  K. O. Cogger.  The Optimality of General-Order Exponential Smoothing , 1974, Oper. Res..

[9]  Douglas C. Montgomery,et al.  Forecasting and time series analysis , 1976 .

[10]  Robert Goodell Brown,et al.  Smoothing, forecasting and prediction of discrete time series , 1964 .

[11]  D. W. Trigg,et al.  Exponential Smoothing with an Adaptive Response Rate , 1967 .

[12]  R. Hogg.  An Introduction to Robust Estimation , 1979 .

[13]  Paul W. Holland,et al.  Two Robust Alternatives to Least-Squares Regression , 1977 .

[14]  Stephen D. Roberts,et al.  The Development of a Self-Adaptive Forecasting Technique , 1969 .

[15]  D. Montgomery,et al.  Augmented Robust Estimators , 1980 .

[16]  Alan B. Forsythe,et al.  Robust Estimation of Straight Line Regression Coefficients by Minimizing pth Power Deviations , 1972 .

[17]  L. Denby,et al.  Robust regression estimators compared via monte carlo , 1977 .

[18]  A. E. Hoerl,et al.  Ridge regression:some simulations , 1975 .

[19]  Johannes Ledolter,et al.  A recursive approach to parameter estimation in regression and time series models , 1979 .

[20]  Donald W. Marquaridt.  Generalized Inverses, Ridge Regression, Biased Linear Estimation, and Nonlinear Estimation , 1970 .

[21]  R. Meyer,et al.  The Fundamental Theorem of Exponential Smoothing , 1961 .

[22]  J. Ramsay.  A Comparative Study of Several Robust Estimates of Slope, Intercept, and Scale in Linear Regression , 1977 .

[23]  R. Snee,et al.  Ridge Regression in Practice , 1975 .

[24]  P. J. Huber.  Robust Regression: Asymptotics, Conjectures and Monte Carlo , 1973 .

[25]  G. C. McDonald,et al.  A Monte Carlo Evaluation of Some Ridge-Type Estimators , 1975 .

[26]  Ed McKenzie.  An Analysis of General Exponential Smoothing , 1976, Oper. Res..

[27]  Norman R. Draper,et al.  Ridge Regression and James-Stein Estimation: Review and Comments , 1979 .

[28]  R. Hogg.  Statistical Robustness: One View of its use in Applications Today , 1979 .

[29]  M. L. Goodman,et al.  A New Look at Higher-Order Exponential Smoothing for Forecasting , 1974, Oper. Res..

[30]  A. E. Hoerl,et al.  Ridge Regression: Applications to Nonorthogonal Problems , 1970 .