Logistic Model for Normality by Neural Networks

  • Published : 2003.02.28

Abstract

We propose a new logistic regression model of normality curves for normal(diseased) and abnormal(nondiseased) classifications by neural networks in data mining. The fitted logistic regression lines are estimated, interpreted and plotted by the neural network technique. A few goodness-of-fit test statistics for normality are discussed and the performances by the fitted logistic regression lines are conducted.

References

  1. Neural Network Haykin, Simon
  2. Technometrics v.19 Goodness-of-fit tests based on nonlinearity in probability plots LaBrecque, J.
  3. J of Information and Optimization Sciences v.20-3 The Goodness-of-fit tests of normality by ROC curves Lee, J.-Y.;Rhee, S.-W.
  4. J of Information and Optimization Sciences v.19-3 A transformed quantile-quantile plot for normal and bimodal distributions Lee, J.-Y.;Woo, J. S.;Rhee, S.-W.
  5. The American Statistician v.36 An objective graphical method for testing normal distributional assumptions using probability plots Mage, D. T.
  6. Biometrika v.52 An analysis-of-variance test for normality (complete sample) Shapiro, S. S.;Wilk, M. B.
  7. Biometrika v.55 Probability plotting methods for the analysis of data Wilk, M. B.;Gnanadesikan, R.