Comparing the Modeling Performance of Regression and Neural Networks as Data Quality Varies: A Business Value Approach

Under circumstances where data quality may vary (due to inaccuracies or lack of timeliness, for example), knowledge about the potential performance of alternate predictive models can help a decision maker to design a business-value-maximizing information system. This paper examines a real-world example from the field of finance to illustrate a comparison of alternative modeling tools. Two modeling alternatives are used in this example: regression analysis and neural network analysis. There are two main results: (1) Linear regression outperformed neural nets in terms of forecasting accuracy, but the opposite was true when we considered the business value of the forecast. (2) Neural net-based forecasts tended to be more robust than linear regression forecasts as data accuracy degraded. Managerial implications for financial risk management of mortgage-backed security portfolios are drawn from the results.

[1]  Charles H. Kriebel,et al.  MIS and information economics: augmenting rich descriptions with analytical rigor in information systems design , 1989, ICIS '89.

[2]  Michael H. Kutner.  Applied Linear Statistical Models , 1974 .

[3]  Averill M. Law,et al.  Simulation Modeling and Analysis , 1982 .

[4]  Jerry R. Green,et al.  The Effects of Interest Rates on Mortgage Prepayments , 1983 .

[5]  H. White.  Consequences and Detection of Misspecified Nonlinear Regression Models , 1981 .

[6]  Stavros A. Zenios,et al.  Complete prepayment models for mortgage-backed securities , 1992 .

[7]  Soumitra Dutta,et al.  Bond rating: A non-conservative application of neural networks , 1988 .

[8]  Robert J. Kauffman,et al.  Risk Management and Data Quality Selection: An Information Economics Approach , 1991 .

[9]  Philip D. Wasserman,et al.  Neural computing - theory and practice , 1989 .

[10]  Wayne H. Wagner,et al.  The Complete guide to securities transactions : enhancing investment performance and controlling costs , 1989 .

[11]  Daniel Martin,et al.  Early warning of bank failure: A logit regression approach , 1977 .

[12]  E. Mine Cinar,et al.  Neural Networks: A New Tool for Predicting Thrift Failures , 1992 .

[13]  Niv Ahituv,et al.  Assessing the value of information , 1989, ICIS '89.

[14]  Donald P. Ballou,et al.  Modeling Data and Process Quality in Multi-Input, Multi-Output Information Systems , 1985 .

[15]  William W. Bartlett,et al.  Mortgage-Backed Securities: Products, Analysis, Trading , 1989 .

[16]  E. Altman,et al.  ZETATM analysis A new model to identify bankruptcy risk of corporations , 1977 .

[17]  Ronald W. Hilton.  The Determinants of Information Value: Synthesizing Some General Results , 1981 .

[18]  Richard Roll,et al.  Prepayments on fixed-rate mortgage-backed securities , 1989 .

[19]  Frank J. Navratil.  THE ESTIMATION OF MORTGAGE PREPAYMENT RATES , 1985 .

[20]  Frank J. Fabozzi,et al.  Advances in Bond Analysis & Portfolio Strategies , 1987 .

[21]  H. White,et al.  Economic prediction using neural networks: the case of IBM daily stock returns , 1988, IEEE 1988 International Conference on Neural Networks.

[22]  F. Fabozzi.  The Handbook of Mortgage-Backed Securities , 1985 .

[23]  Melody Y. Kiang,et al.  Managerial Applications of Neural Networks: The Case of Bank Failure Predictions , 1992 .

[24]  Joel S. Demski,et al.  Information Analysis , 1980 .

[25]  Richard P. Lippmann,et al.  An introduction to computing with neural nets , 1987 .

[26]  Jack M. Guttentag,et al.  The Yield on Insured Residential Mortgages , 1974 .

[27]  Niv Ahituv,et al.  Assessing the value of information , 1989, ICIS '89.

[28]  P. Meyer,et al.  PREDICTION OF BANK FAILURES , 1970 .

[29]  Katherine Schipper,et al.  Application of Classification Techniques in Business, Banking and Finance. , 1983 .

[30]  W. Beaver.  Financial Ratios As Predictors Of Failure , 1966 .