EKONOMETRİK MODEL SEÇİM KRİTERLERİ ÜZERİNE KISA BIR İNCELEME

Makalenin İngilizce İsmi: 
A Brief Survey Of Econometrics Model Selection Criteria
Makale İçerik Bilgileri
Makale Dili: 
Türkçe
Anahtar Kelimeler: 
Bilgi Teorisi
Model Seçim Kriterleri
Bootstrap
Çapraz Geçerlilik
Türkçe Özet: 

“Hangi değişkenler önemli? , Bir model nasıl seçilir? gibi sorular modellemede çok
önemlidir. İyi araştırma (ekonometrik) teknikleri altında iyi bir model kesinlikle verileri
uygun tahmin eder. Birden fazla uygun model tanımlamasının bulunduğu durumlarda
ekonometrisyen ve istatistikçi mevcut olan veri setinden uygun modeli seçmek ister. Model
seçim kriterleride en uygun model kararının verilmesi için bir yoldur.
Bu çalışma farklı model seçim kriterlerinin kısa incelemesini ve birbirleriyle
karşılaştırmasını içermektedir. Analiz edilen model seçim kriterleri geleneksel hipotez
testine bağlı metodlardan farklı olarak bilgi teorisine dayanmaktadır. Kullback-Leibler
uyumsuzluğuna dayanan Akaike Bilgi Kriteri ile bilgi teorisi yaklaşımı 1970’lerde
popülerdi. Daha sonraları bu yaklaşım Bayes Bilgi Kriteri(BIC), Schwartz Bilgi Kriteri
(SIC), Mallow’un Cp kriteri gibi örneklerle çeşitlenerek gelişmiştir.
Bu çalışmada ayrıca yeniden örnekleme methodlarından bootstrap ve çapraz-geçerlilik.te
model seçim kriterleri içinde anlatılmıştır.

Key Words: 
Information Theory
Model Selection Criteria
Bootstrap
Cross-Validation
İngilizce Özet: 

“Which variables are important?, How to select a model?” kind of questions are very
important for the modeling. A good model certainly fits the well in to the data under
investigation (econometrics). The econometrician and statistician would like to select most
appropriate model from data sets, where there may be more than one definition of
“appropriate”. Model selection criteria are one way to decide on the most appropriate
model.
This paper surveys briefly the different model selection criteria and compares them with
each other. The analyzed model selection criteria are based on the information theory and
are quite different from the usual methods based on null hypothesis testing. Information
theory approaches were popular in the 1970s with the land mark Akaike Information
Criteria based on the Kullback-Leibler discrepancy. Later, those approaches were
diversified and such criteria as Bayes Information Criterion (BIC), Schwartz Information
Criterion (SCI), and Mallow’s Cp were developed. In the paper, the resample methods (bootstrap and cross validation) were also explained in
the contents of model selection criteria.

Yazar Bilgileri
1. Yazar
Yazar Adı: 
Meltem Şengün UCAL
Yazar Ünvanı: 
Doktor
Yazar Üniversitesi: 
Kadir Has Üniversitesi
Yazar Fakültesi: 
İktisadi İdari Bilimler Fakültesi
Yazar Anabilim Dalı: 
İktisat Anabilim Dalı
Makale Künye Bilgisi
Makalenin Yayımlandığı Dergi: 
Cumhuriyet Üniversitesi İktisadi ve İdari Bilimler Dergisi
Makale Yayın Yılı: 
2006
Cilt/Sayı: 
7
Sayı: 
2
Sayfa Aralığı: 
41-57
PDF Dosyası: 
application/pdf iconarastirmax_1776_pp_41-57.pdf
Makalenin Yayınlandığı Web Sitesindeki Linki: 
Makaleyi Dergi Web Sitesinden İndirin
Referanslar: 

1. Akaike, H. (1970),”Statistical Predictor Identification”, Annals of the Institute of
Statistical Mathematics, 22, pp. 203-217.
2. ------------ (1973), Information Theory and an Extension of the Maximum
Likelihood Principle. InB.N. Petro and F. Csaki ed. 2nd International Symposium
on Information Theory, pp. 267-281.
3. ------------ (1978), “A Bayesian analysis of the Minimum AIC Procedure,” Annals
of the Institute of Statistical Mathematics, 20, pp. 9-14.
4. Banke, O. and Drage, B.(1984), “Bootstrap and Cross-Validation Estimates of the
Prediction Error for Linear Regression Models”, The Annals of Statistics, 12, pp.
1400-1424.
5. Bhansalli, R. J. Ve Downham, D. Y. (1977), “Some Properties of the Order of an
Autoregressive Model Seected by a Generalization of Akaike’s EPF Criterion,”
Biometrika, 64, pp. 547-551.
6. Box, G.E.P. ve Jenkins, G.M. (1970), Time Series Analysis: Forcasting and
Control, Holden-Day, London.
7. Burham, K.P. ve Anderson, D.R.(1998), Model Selection an Inference: A Practical
Information Theoric Approach, Springer-Verlag, New York.
8. Mallows, C.L.(1973), ”Some Comments on Cp”, Technometrics, 15, pp. 661-675.
9. ---------------- (1995), “More Comments on Cp,” Technometrics, 37, pp. 362-372.
10. Davidson, A. C. ve Hinkley, D.V. (1997), Bootstrap Methods and Their
Application, New York, Cambridge University Press.
11. Draper, N. R., Applied Regression Analysis, John Wıley & Sons, New York, 1998.
Ebbler, D. H.(1975), “On the Probability of Correct Model Selecting Using the
12. Maximum R-Square Choıce Criterion,” International Economic Review, 16-2, pp.
516-520.
13. Efron, B. (1979), “Bootstrap methods: Another look at the jackknife”, Ann. Statist.,
1, pp.446- 453.
14. ----------- (1982), The Jacknife, the Bootstrap and other Resampling Plans,
Philadelphia, Society for Industrial and Applied Mathematics.
15. ----------- (1983), “Estimating the Error Rate of Prediction Rule Improvement on
Cross-Validation”, Journal of the American Statistical Association, 78, 316-331.
16. Efron, B. ve Tibshirani, R. J. (1993), An Introduction to the Bootstrap, New York:
Champman & Hall.
17. Geweke, J. Ve Meese, R. (1981), “Estimating Regression Models of Finite but
Unknown Order,” International Economic Review, 22, pp. 55-70.
18. Gujarati, D. (2003), Basic Econometrics, McGraw Hill, Singapure.
Hanmam E. J. ve Quinn, B. G. (1979), “The Determination of the Order of an
Auturegression,” Jounal of the Royal Statistical Society, 41, pp. 190-195.
19. Harvey, A. (1990), The Econometric Analysis of Time Series, The MIT Press, 2.b.,
Cambridge Mass.
20. Hendry, D.F. (1980),”Econometrics-Alchemy or Science”, Economica, c.47.
21. Hjorth, J.S. (1994), Computer Intensive Statistical Methods Validation Model
Selection and Boostrap, New York: Champman & Hall.
22. Hurvich, C.M. ve Tsai, C. (1989), “ Regression and Time Series Model Selection
in Small Samples”, Biometrika, 76, pp.297-307.
23. Kullback, S. ve Leibler, R. A.(1951), “On Information and Sufficiency,” Annals of
Mathematical Statistics, 22, pp. 79-86.
24. Leamer, E. E.(1979), “Information Criteria for Choice of Regression Models: A
Comment”, Econometrica, 47, pp.507-510.
25. Lindley, D.(1956), “ On A Measures of the Information Provided By An
Experiment.” Annals of Mathematics and Statistics, 27, pp.986-1003.
26. Linhart, H. ve Zucchini, W (1986), Model Selection, John Wıley & Sons, New
York.
27. McQuarrie, A.D. ve Tsai, C. L., Regression and Time Series Model Selection,
World Sciencetific, 1998.
28. Nishi, R. (19849, “Asymptotic Properties of Criteria for Selection of Variable in
Multible Regression,” Annals of Statistics, 12, pp.758-765.
29. Onish, H. (1983),”A Variable Selection Procedure for Econometrics Models,”
Computational Statistics & Data Analysis, pp. 85-95.
30. Poirier, D. (1991),” Posterior Odds Analysis When All Competing Models are
False,”Econometric Society Meeting, Amsterdam, The Netherlands.
Raftery, A. E. (1995), Bayesian Model Selection in Social Research(with
Discussion by Andrew Gelman, Donald B. Rubin and Robert M. Hauser). In P.V.
Marsdn (Ed.), Socialogical Methodology, pp. 111-196.
31. Saho, J. (1993), “Linear Model Selection by Cross-Validation”, Journal of the
American Assosiational, Vol. 88.
32. ---------- (1996),”Bootstrap Model Selection”, Journal of the American Statistical
Association, 91, 434, pp.655-665.
33. Shao, J.ve. Tu, D (1996), The Jacknife and Bootstrap, Newyork, Springer.
Schwarz, G. (1978), ”Estimating the Dimensions of a Model,” The Annals of
Statistical, 6, pp.461-464.
34. Seber, G.A.F.(1977), Linear Regression Analysis, John Wıley & Sons, New York.

Türkiye’nin ilk İşletme Fakültesi olan İstanbul Üniversitesi İşletme Fakültesi bir ilke daha imza atmaya hazırlanıyor. Arastirmax.com "1. Liselerarası İşletme ve Ekonomi Proje Yarışması"nın sponsorlarından biri olmaktan gurur duymakta.