Applicability of information criteria for neural network architecture selection

Christian Haefke, Christian Helmenstein

Research output: Contribution to conferencePaper

Abstract

In most of the empirical research on capital markets, stock market indexes are used as proxies for the aggregate market development. In previous work we found that a particular market segment of the Vienna stock exchange might be less efficient than the whole market and hence easier to forecast. Extending the focus of investigation in this paper, we use feedforward networks and linear models to predict the all share index WBI as well as various subindexes covering the highly liquid, semi-liquid, and initial public offering (IPO) market segment. In order to shed some light on network construction principles, we compare different models as selected by hold-out crossvalidation (HCV), Akaike's information criterion (AIC), and Schwartz' information criterion (SIC). The forecasts are subsequently evaluated on the basis of hypothetical trading in the out-of-sample period.

Original languageEnglish (US)
Pages293-301
Number of pages9
StatePublished - Dec 1 1996
EventProceedings of the IEEE/IAFE 1996 Conference on Computational Intelligence for Financial Engineering, CIFEr - New York, NY, USA
Duration: Mar 24 1996Mar 26 1996

Other

OtherProceedings of the IEEE/IAFE 1996 Conference on Computational Intelligence for Financial Engineering, CIFEr
CityNew York, NY, USA
Period3/24/963/26/96

Fingerprint

Network architecture
Neural networks
Liquids
Financial markets
Market segments
Information criterion
Stock exchange
Stock market index
Initial public offerings
Capital markets
Cross-validation
Market development
Feedforward networks
Empirical research
Akaike information criterion
Network model

ASJC Scopus subject areas

  • Computer Science(all)
  • Economics, Econometrics and Finance(all)
  • Engineering(all)

Cite this

Haefke, C., & Helmenstein, C. (1996). Applicability of information criteria for neural network architecture selection. 293-301. Paper presented at Proceedings of the IEEE/IAFE 1996 Conference on Computational Intelligence for Financial Engineering, CIFEr, New York, NY, USA, .

Applicability of information criteria for neural network architecture selection. / Haefke, Christian; Helmenstein, Christian.

1996. 293-301 Paper presented at Proceedings of the IEEE/IAFE 1996 Conference on Computational Intelligence for Financial Engineering, CIFEr, New York, NY, USA, .

Research output: Contribution to conferencePaper

Haefke, C & Helmenstein, C 1996, 'Applicability of information criteria for neural network architecture selection' Paper presented at Proceedings of the IEEE/IAFE 1996 Conference on Computational Intelligence for Financial Engineering, CIFEr, New York, NY, USA, 3/24/96 - 3/26/96, pp. 293-301.
Haefke C, Helmenstein C. Applicability of information criteria for neural network architecture selection. 1996. Paper presented at Proceedings of the IEEE/IAFE 1996 Conference on Computational Intelligence for Financial Engineering, CIFEr, New York, NY, USA, .
Haefke, Christian ; Helmenstein, Christian. / Applicability of information criteria for neural network architecture selection. Paper presented at Proceedings of the IEEE/IAFE 1996 Conference on Computational Intelligence for Financial Engineering, CIFEr, New York, NY, USA, .9 p.
@conference{06ee5e2d290241998748afb432c084ec,
title = "Applicability of information criteria for neural network architecture selection",
abstract = "In most of the empirical research on capital markets, stock market indexes are used as proxies for the aggregate market development. In previous work we found that a particular market segment of the Vienna stock exchange might be less efficient than the whole market and hence easier to forecast. Extending the focus of investigation in this paper, we use feedforward networks and linear models to predict the all share index WBI as well as various subindexes covering the highly liquid, semi-liquid, and initial public offering (IPO) market segment. In order to shed some light on network construction principles, we compare different models as selected by hold-out crossvalidation (HCV), Akaike's information criterion (AIC), and Schwartz' information criterion (SIC). The forecasts are subsequently evaluated on the basis of hypothetical trading in the out-of-sample period.",
author = "Christian Haefke and Christian Helmenstein",
year = "1996",
month = "12",
day = "1",
language = "English (US)",
pages = "293--301",
note = "Proceedings of the IEEE/IAFE 1996 Conference on Computational Intelligence for Financial Engineering, CIFEr ; Conference date: 24-03-1996 Through 26-03-1996",

}

TY - CONF

T1 - Applicability of information criteria for neural network architecture selection

AU - Haefke, Christian

AU - Helmenstein, Christian

PY - 1996/12/1

Y1 - 1996/12/1

N2 - In most of the empirical research on capital markets, stock market indexes are used as proxies for the aggregate market development. In previous work we found that a particular market segment of the Vienna stock exchange might be less efficient than the whole market and hence easier to forecast. Extending the focus of investigation in this paper, we use feedforward networks and linear models to predict the all share index WBI as well as various subindexes covering the highly liquid, semi-liquid, and initial public offering (IPO) market segment. In order to shed some light on network construction principles, we compare different models as selected by hold-out crossvalidation (HCV), Akaike's information criterion (AIC), and Schwartz' information criterion (SIC). The forecasts are subsequently evaluated on the basis of hypothetical trading in the out-of-sample period.

AB - In most of the empirical research on capital markets, stock market indexes are used as proxies for the aggregate market development. In previous work we found that a particular market segment of the Vienna stock exchange might be less efficient than the whole market and hence easier to forecast. Extending the focus of investigation in this paper, we use feedforward networks and linear models to predict the all share index WBI as well as various subindexes covering the highly liquid, semi-liquid, and initial public offering (IPO) market segment. In order to shed some light on network construction principles, we compare different models as selected by hold-out crossvalidation (HCV), Akaike's information criterion (AIC), and Schwartz' information criterion (SIC). The forecasts are subsequently evaluated on the basis of hypothetical trading in the out-of-sample period.

UR - http://www.scopus.com/inward/record.url?scp=0030413768&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0030413768&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:0030413768

SP - 293

EP - 301

ER -