How to evaluate models:: Observed vs. predicted or predicted vs. observed?

被引:669
作者
Pineiro, Gervasio [1 ]
Perelman, Susana
Guerschman, Juan P.
Paruelo, Jose M. [1 ]
机构
[1] Univ Buenos Aires, CONICET, Fac Agron, IFEVA,Catedra Ecol,Lab Anal Reg & Teledetecc, RA-4453 San Martin, Capital Federal, Argentina
关键词
measured values; simulated values; regression; slope; intercept; linear models; regression coefficient; goodness-of-fit; 1 : 1 line;
D O I
10.1016/j.ecolmodel.2008.05.006
中图分类号
Q14 [生态学(生物生态学)];
学科分类号
071012 ; 0713 ;
摘要
A common and simple approach to evaluate models is to regress predicted vs. observed values (or vice versa) and compare slope and intercept parameters against the 1:1 line. However, based on a review of the literature it seems to be no consensus on which variable (predicted or observed) should be placed in each axis. Although some researchers think that it is identical, probably because r(2) is the same for both regressions, the intercept and the slope of each regression differ and, in turn, may change the result of the model evaluation. We present mathematical evidence showing that the regression of predicted (in the y-axis) vs. observed data (in the x-axis) (PO) to evaluate models is incorrect and should lead to an erroneous estimate of the slope and intercept. In other words, a spurious effect is added to the regression parameters when regressing PO values and comparing them against the 1:1 line. observed (in the y-axis) vs, predicted (in the x-axis) (OP) regressions should be used instead. We also show in an example from the literature that both approaches produce significantly different results that may change the conclusions of the model evaluation. (C) 2008 Elsevier B.V. All rights reserved.
引用
收藏
页码:316 / 322
页数:7
相关论文
共 17 条
[1]   Application of MODIS derived parameters for regional crop yield assessment [J].
Doraiswamy, PC ;
Sinclair, TR ;
Hollinger, S ;
Akhmedov, B ;
Stern, A ;
Prueger, J .
REMOTE SENSING OF ENVIRONMENT, 2005, 97 (02) :192-202
[2]   Model evaluation by comparison of model-based predictions and measured values [J].
Gauch, HG ;
Hwang, JTG ;
Fick, GW .
AGRONOMY JOURNAL, 2003, 95 (06) :1442-1446
[3]   Evaluating predictive errors of a complex environmental model using a general linear model and least square means [J].
Knightes, CD ;
Cyterski, M .
ECOLOGICAL MODELLING, 2005, 186 (03) :366-374
[4]   Comparing simulated and measured values using mean squared deviation and its components [J].
Kobayashi, K ;
Salam, MU .
AGRONOMY JOURNAL, 2000, 92 (02) :345-352
[5]  
Lauenroth W., 2003, MODELS ECOSYSTEM SCI
[6]  
Lobell DB, 2005, AGRON J, V97, P241
[7]   A hypothesis testing framework for evaluating ecosystem model performance [J].
Loehle, C .
ECOLOGICAL MODELLING, 1997, 97 (03) :153-165
[8]   Evaluation of simple statistical criteria to qualify a simulation [J].
Mesple, F ;
Troussellier, M ;
Casellas, C ;
Legendre, P .
ECOLOGICAL MODELLING, 1996, 88 (1-3) :9-18
[9]   Misuse of regression for empirical validation of models [J].
Mitchell, PL .
AGRICULTURAL SYSTEMS, 1997, 54 (03) :313-326
[10]   Uncertainty analysis and validation of environmental models: The empirically based uncertainty analysis [J].
Monte, L ;
Hakanson, L ;
Bergstrom, U ;
Brittain, J ;
Heling, R .
ECOLOGICAL MODELLING, 1996, 91 (1-3) :139-152