| Peer-Reviewed

Discriminant Analysis Procedures Under Non-optimal Conditions for Binary Variables

Received: 12 October 2015     Accepted: 26 October 2015     Published: 10 December 2015
Views:       Downloads:
Abstract

The performance of four discriminant analysis procedures for the classification of observations from unknown populations was examined by Monte Carlo methods. The procedures examined were the Fisher Linear discriminant function, the quadratic discriminant function, a polynomial discriminant function and A-B linear procedure designed for use in situations where covariance matrices are equal. Each procedure was observed under conditions of equal sample sizes, equal covariance matrices, and in conditions where the sample was drawn from populations that have a multivariate normal distribution. When the population covariance matrices were equal, or not greatly different, the quadratic discriminant function performed similarly or marginally the same like Linear procedures. In all cases the polynomial discriminate function demonstrated the poorest, linear discriminant function performed much better than the other procedures. All of the procedures were greatly affected by non-normality and tended to make many more errors in the classification of one group than the other, suggesting that data be standardized when non-normality is suspected.

Published in American Journal of Theoretical and Applied Statistics (Volume 4, Issue 6)
DOI 10.11648/j.ajtas.20150406.32
Page(s) 602-609
Creative Commons

This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium or format, provided the original work is properly cited.

Copyright

Copyright © The Author(s), 2015. Published by Science Publishing Group

Keywords

Apparent Error Rates, Fisher’s Linear Discriminant, Quadratic Discriminant Function, A-B Discriminant Function, Polynomial Discriminant Function

References
[1] Anderson, T.W. and Bahadur, R.R. (1962).Classification into two multivariate normal distributions with different covariance matrices. Annals of Mathematics Statistics, 33.420-431.
[2] Barnard, M.M. (1935). The Secular variations of skull characteristics in four series of Egyptian skulls. Annals Eugenics, v. 6, 352-371.
[3] Benerjee, K.S. & Marcus, L.F. (1965).In a Minimax Classification Procedure. Biometrics, 52, 654-654
[4] Gilbert, S.E. (1968). “On Discrimination using Qualitative Variables” Journal of the American Statistical Association 1399-1418.
[5] Gold Stein M. &Wolf (1977). On the problem of Bias multinomial classification. Biometrics 33, 325-331.
[6] Hills, M. (1967). “Discriminantion and allocation with discrete data”, Applied Statistics. 16 237-250.
[7] Lachenbruch, P.A. (1975) Discriminant Analysis. Hafner Press New York:
[8] Marks, S. & Dunn, O.J. (1974). Discriminant functions when Covariance matrices are unequal. Journal of the American Statistical Association, 69, 555-559.
[9] Martins, D.C., & Bradley, R.R. (1972). Probability Models, Estimation and Classification for Multivariate Dichotomous Populations, Biometrics,23, 203-221
[10] Onyeagu S.I. (2003). Derivation of an optimal classification rule for discrete variables Journal of Nigerian Statistical Association, 73, 724-745.
[11] Oluadare. S. (2011). Rubust Linear classifier for equal Cost Ratios of misclassification. CBN Journal of Applied Statistics.(2) (1)
[12] Rao, C.R. (1965).Linear Statistical Inference and Its Applications: John Willey New York
[13] Richard A.J. & Dean W.W. (1988).Applied Multivariate Statistical Analysis.4th edition Prentice Hall. Inc. New Jessey.
[14] Smith, H.F. (1936). A discriminant function for plant selection. Ann. Eugn. 7, 240 – 250.
[15] Tou, J.T. & Gonzalez, R. C. (1974).Pattern Recognition Principles. Reading, Mass.; Addison –Wesley.
[16] Slah, B.Y. & Abdelwaheb Rebai (2007). Comparison between Statistical Approaches and linear programming for resolving classification problem. International Mathematics Forum, 2,(63), 3125-3141.
[17] Egbo, I., Onyeagu, S.I. & Ekezie, D.D. (2014). A comparison of multinomial classification rules for binary variables. International Journal of Mathematical Science and Engineering Applications (IJMSEA),8, 141-157.
[18] Ekezie, D.D. (2012). Comparison of seven Asymptotic Error Rate Expansion for the sample linear Discriminant function. Unpublished Ph.D thesis submitted to Department of Statistics, Imo State University, Owerri, Nigeria.
[19] Egbo, I., Onyeagu, S.I. & Ekezie, D.D. (2014).A Comparison of Multivariate Discrimination of Binary Data. International Journal of Mathematics and Statistics Studies, 2(4), 40-61.
Cite This Article
  • APA Style

    I. Egbo. (2015). Discriminant Analysis Procedures Under Non-optimal Conditions for Binary Variables. American Journal of Theoretical and Applied Statistics, 4(6), 602-609. https://doi.org/10.11648/j.ajtas.20150406.32

    Copy | Download

    ACS Style

    I. Egbo. Discriminant Analysis Procedures Under Non-optimal Conditions for Binary Variables. Am. J. Theor. Appl. Stat. 2015, 4(6), 602-609. doi: 10.11648/j.ajtas.20150406.32

    Copy | Download

    AMA Style

    I. Egbo. Discriminant Analysis Procedures Under Non-optimal Conditions for Binary Variables. Am J Theor Appl Stat. 2015;4(6):602-609. doi: 10.11648/j.ajtas.20150406.32

    Copy | Download

  • @article{10.11648/j.ajtas.20150406.32,
      author = {I. Egbo},
      title = {Discriminant Analysis Procedures Under Non-optimal Conditions for Binary Variables},
      journal = {American Journal of Theoretical and Applied Statistics},
      volume = {4},
      number = {6},
      pages = {602-609},
      doi = {10.11648/j.ajtas.20150406.32},
      url = {https://doi.org/10.11648/j.ajtas.20150406.32},
      eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajtas.20150406.32},
      abstract = {The performance of four discriminant analysis procedures for the classification of observations from unknown populations was examined by Monte Carlo methods. The procedures examined were the Fisher Linear discriminant function, the quadratic discriminant function, a polynomial discriminant function and A-B linear procedure designed for use in situations where covariance matrices are equal. Each procedure was observed under conditions of equal sample sizes, equal covariance matrices, and in conditions where the sample was drawn from populations that have a multivariate normal distribution. When the population covariance matrices were equal, or not greatly different, the quadratic discriminant function performed similarly or marginally the same like Linear procedures. In all cases the polynomial discriminate function demonstrated the poorest, linear discriminant function performed much better than the other procedures. All of the procedures were greatly affected by non-normality and tended to make many more errors in the classification of one group than the other, suggesting that data be standardized when non-normality is suspected.},
     year = {2015}
    }
    

    Copy | Download

  • TY  - JOUR
    T1  - Discriminant Analysis Procedures Under Non-optimal Conditions for Binary Variables
    AU  - I. Egbo
    Y1  - 2015/12/10
    PY  - 2015
    N1  - https://doi.org/10.11648/j.ajtas.20150406.32
    DO  - 10.11648/j.ajtas.20150406.32
    T2  - American Journal of Theoretical and Applied Statistics
    JF  - American Journal of Theoretical and Applied Statistics
    JO  - American Journal of Theoretical and Applied Statistics
    SP  - 602
    EP  - 609
    PB  - Science Publishing Group
    SN  - 2326-9006
    UR  - https://doi.org/10.11648/j.ajtas.20150406.32
    AB  - The performance of four discriminant analysis procedures for the classification of observations from unknown populations was examined by Monte Carlo methods. The procedures examined were the Fisher Linear discriminant function, the quadratic discriminant function, a polynomial discriminant function and A-B linear procedure designed for use in situations where covariance matrices are equal. Each procedure was observed under conditions of equal sample sizes, equal covariance matrices, and in conditions where the sample was drawn from populations that have a multivariate normal distribution. When the population covariance matrices were equal, or not greatly different, the quadratic discriminant function performed similarly or marginally the same like Linear procedures. In all cases the polynomial discriminate function demonstrated the poorest, linear discriminant function performed much better than the other procedures. All of the procedures were greatly affected by non-normality and tended to make many more errors in the classification of one group than the other, suggesting that data be standardized when non-normality is suspected.
    VL  - 4
    IS  - 6
    ER  - 

    Copy | Download

Author Information
  • Department of Mathematics, Alvan Ikoku University of Education, Owerri, Nigeria

  • Sections