SELECTION OF SIMPLIFIED MODELS: I. ANALYSIS OF MODEL-SELECTION CRITERIA USING MEAN-SQUARED ERROR

被引：22

作者：

Wu, Shaohua ^{[2
]}

McAuley, K. B. ^{[1
]}

Harris, T. J. ^{[1
]}

机构：

[1] Queens Univ, Dept Chem Engn, Kingston, ON K7L 3N6, Canada

[2] Honeywell Canada, Mississauga, ON L5L 3S6, Canada

来源：

CANADIAN JOURNAL OF CHEMICAL ENGINEERING | 2011年 / 89卷 / 01期

基金：

加拿大自然科学与工程研究理事会;

关键词：

mean-squared prediction error; model-selection criteria; noncentral F distribution; simplified models; CHEMICAL-PROCESSES; DYNAMIC-MODEL; BED REACTOR; REGRESSION; SYSTEMS; IDENTIFICATION; EXTRACTION; EXCHANGER; VARIABLES; KINETICS;

D O I：

10.1002/cjce.20406

中图分类号：

TQ [化学工业];

学科分类号：

0817 ;

摘要：

Mean-squared error (MSE) is used to analyse nine commonly used model-selection criteria (MSC) for their performance when selecting simplified models (SMs). Expressions are derived to enable exact calculations of the probability that a particular MSC will select a SM. For several common MSC, the relative propensities to select SMs are independent of model structure and data. It is shown that MSC that are effective in preventing overfitting are prone to underfitting when information content of the data is low. In a subsequent article, results are extended to develop a new MSE-based MSC for selecting nonlinear multi-response SMs.

引用

页码：148 / 158

页数：11

共 54 条

[1] BAYESIAN-ANALYSIS OF MINIMUM AIC PROCEDURE [J].

AKAIKE, H .

ANNALS OF THE INSTITUTE OF STATISTICAL MATHEMATICS, 1978, 30 (01) :9-14

[2]

Akaike H., 1973, 2 INTERNAT SYMPOS IN, P267, DOI [DOI 10.1007/978-1-4612-1694-0_15, 10.1007/978-1-4612-1694-0, 10.1007/978-1-4612-0919-5_38]

[3]

[Anonymous], 2006, Introduction to Linear Regression Analysis

[4]

[Anonymous], 2002, Model selection and multimodel inference: a practical informationtheoretic approach

[5] Data reconciliation in gas pipeline systems [J].

Bagajewicz, MJ ;

Cabrera, E .

INDUSTRIAL & ENGINEERING CHEMISTRY RESEARCH, 2003, 42 (22) :5596-5606

[6]

Beck J.V., 1977, Parameter estimation in engineering and science

[7] HOW MANY VARIABLES SHOULD BE ENTERED IN A REGRESSION EQUATION [J].

BREIMAN, L ;

FREEDMAN, D .

JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1983, 78 (381) :131-136

[8] A simplified model for trace organics removal by continuous flow PAC adsorption/submerged membrane processes [J].

Chang, S ;

Waite, TD ;

Fane, AG .

JOURNAL OF MEMBRANE SCIENCE, 2005, 253 (1-2) :81-87

[9] Modeling and prediction of the dynamic behavior in a reactor-exchanger using NARMAX neural structure [J].

Chetouani, Yahya .

CHEMICAL ENGINEERING COMMUNICATIONS, 2007, 194 (05) :691-705

[10] Dynamic model-based batch process monitoring [J].

Choi, Sang Wook ;

Morris, Julian ;

Lee, In-Beum .

CHEMICAL ENGINEERING SCIENCE, 2008, 63 (03) :622-636

← 1 2 3 4 5 6 →