Open Access Open Access  Restricted Access Subscription or Fee Access

Robust Quadratic Discriminant Rule using Comedian

Sajesh T. A., M. R. Srinivasan

Abstract


Discriminant analysis aims to get rules that separate the groups of observations. Moreover, it can be used to classify new observations into one of the known groups. The classical discriminant techniques are based on the classical estimates of mean vector and variance covariance matrix. Unfortunately, these classical estimates are highly influenced by the extreme observations and hence any analysis based on these estimates will provide misleading results when the data contain outlying observations. This study proposes robust discriminant rules based on the highly robust Comedian estimates of location and scatter. Efficacy of the proposed is examined through simulation study.


Keywords: Classification, Discriminant analysis, Comedian, Robust statistics

Cite this Article T. A. Sajesh, M. R. Srinivasan. Robust Quadratic Discriminant Rule using Comedian. Research & Reviews: Journal of Statistics. 2019; 8(2): 41–47p.


Keywords


Classification, Discriminant analysis, Comedian, Robust statistics

Full Text:

PDF

References


Hubert M , Van Driesen K. Fast and Robust Discriminant Analysis. Computational Statistics and Data Analysis.2004; 45: 301 – 20p.

Rousseeuw PJ. Multivariate Estimation with High Breakdown Point. Fourth Pannonian Symposium on Mathematical Statistics and Probability. 1983.

Rousseeuw P.J. Least median of squares regression. Journal of the American Statistical Association. 1984; 79 (388): 871–80p.

Sajesh TA, Srinivasan MR. Outlier Detection for High Dimensional Data using Comedian Approach. Journal of statistical computation and simulation. 2012, 82(5):745–57p.


Refbacks

  • There are currently no refbacks.