How to Perform Discriminant Analysis in Medical Research? Explained with Illustrations

Abstract

Discriminant function analysis is the statistical analysis used to analyze data when the dependent variable or outcome is categorical and independent variable or predictor variable is parametric. It is a parametric technique to determine which weightings of quantitative variables or predictors best discriminates between two or more than two categories of dependent variables and does so better than chance. Discriminant analysis is used to find out the accuracy of a given classification system in predicting the sample into a particular group. Discriminant analysis includes the development of discriminant functions for each sample and deriving a cutoff score that is used for classifying the samples into different groups. Discriminant function analysis is a statistical analysis used to find out the accuracy of a given classification system or predictor variables. This article explains the basic assumptions, uses, and necessary requirements of discriminant analysis with a real-life clinical example. Whenever a new classification system is introduced, discriminant function analysis can be used to find out the accuracy with which the classification is able to differentiate a particular sample into different groups. Thus, it is a very useful tool in medical research where classification is required.

Keywords

discriminant analysis - discriminant equation - canonical discriminant functions - centroid

Publication History

Article published online:
01 June 2022

• References

• 1 Cramer D. Advanced Quantitative Data Analysis. New York, NY: McGraw-Hill Education; 2003
• 2 Chapter 25 Discriminant Analysis. Accessed March 3, 2022 from http://www.econ.upf.edu/~satorra/AnalisiMultivariant/Chapter25DiscriminantAnalysis.pdf
• 3 Johnson DH. How to measure habitat: a statistical perspective. US Forest Service General Technical Report RM. 1981 ;87: 53-57
• 4 Wahl PW, Kronmal RA. Discriminant functions when covariances are unequal and sample sizes are moderate. Biometrics 1977; 33: 479-484
• 5 Williams BK, Titus K. Assessment of sampling stability in ecological applications of discriminant analysis. Ecology 1988; 69 (04) 1275-1285
• 6 Dhamnetiya D, Goel MK, Dhiman B, Pathania OP. Gallstone disease and quantitative analysis of independent biochemical parameters: study in a tertiary care hospital of India. J Lab Physicians 2018; 10 (04) 448-452
• 7 Efron B. The efficiency of logistic regression compared to normal discriminant analysis. J Am Stat Assoc 1975; 70: 892-898
• 8 Harrell FE, Lee KL. . A comparison of the discrimination of discriminant analysis and logistic regression under multivariate normality, Biostatistics: Statistics in Biomedical, Public Health and Environmental Sciences. North-Holland, New York, United States;1985:333-343.
• 9 Beleites C, Geiger K, Kirsch M, Sobottka SB, Schackert G, Salzer R. Raman spectroscopic grading of astrocytoma tissues: using soft reference information. Anal Bioanal Chem 2011; 400 (09) 2801-2816
• 10 Hastie T, Tibshirani R, Friedman J. The Elements of Statistical Learning; Data mining, Inference and Prediction. New York: Springer Verlag; 2009
• 11 Lynn RD. “A comparison of tree-based and traditional classification methods: a thesis presented in partial fulfilment of the requirements for the degree of PhD in Statistics at Massey University.” PhD diss. Massey University; 1994
• 12 James G, Witten D, Hastie T, Tibshirani R. An introduction to statistical learning. New York: Springer; 2013
• 13 Friedman JH. Regularized discriminant analysis. J Am Statistic Assoc 1989; 84 (405) 165-75