Comparison of linear discriminant analysis and logistic regression for data classification

Abstract

Linear discriminant analysis (LDA) and logistic regression (LR) are often used for the purpose of classifying populations or groups using a set of predictor variables. Assumptions of multivariate normality and equal variancecovariance matrices across groups are required before proceeding with LDA, but such assumptions are not required for LR and hence LR is considered to be much more robust than LDA. In this paper, several real datasets which are different in terms of normality, number of independent variables and sample size are used to study the performance of both methods. The methods are compared based on the percentage of correct classification and B index. The results show that overall, LR performs better regardless of the distribution of the data is normal or nonnormal. However, LR needs longer computing time than LDA with the increase in sample size. The performance of LDA was also tested by using various prior probabilities. The results show that the average percentage of correct classification and the B index are higher when the prior probability is set based on the group size rather than using equal probabilities for all groups.

Corporate political strategy and liability of foreignness: Similarities and differences between local and foreign firms in the South African Health Sector

“…Freixanet, 2012). Logistic Regression (LR) is a widely used data analysis method to determine membership of a group based on known key characteristics or a finite set of criteria/alternatives (Antipov & Pokryshevskaya, 2010;Liong & Foo, 2013). We chose an LR approach as it solves the problem of data heterogeneity in our classification of foreign and local firms and the potential problem of an inaccuracies arising from aggregate predictive methods (Antipov & Pokryshevskaya, 2010).…”

Determination of macro, micro and trace elements in citrus fruits by inductively coupled plasma–optical emission spectrometry (ICP‐OES), ICP–mass spectrometry and direct mercury analyzer

“…Significant differences ( P < 0.05) within means were analyzed by one‐way analysis of variance (ANOVA) using SPSS Statistics software, version 20 (IBM, New York, USA). Linear discriminant analysis (LDA), a multivariate statistical method for data analysis with categorical outcome variables, can construct a linear classification model and a boundary between two groups . In this study, LDA was performed in XLSTAT 2015 (Addinsoft, Paris, France).…”

“…Linear discriminant analysis (LDA), a multivariate statistical method for data analysis with categorical outcome variables, can construct a linear classification model and a boundary between two groups. 23 In this study, LDA was performed in XLSTAT 2015 (Addinsoft, Paris, France).…”

Quantification of complex modular architecture in plants

“…We finally employed a linear discriminant analysis (LDA) to determine whether the categorical and quantitative characters significantly vary among Riccardia species and between Riccardia , Afroriccardia and Aneura . LDA is designed for the analysis of variables that are normally distributed within each category (here the molecular species), is sensitive to multicollinearity, and is designed to work with matrices including more observations in the category with the lowest sampling size than variables (Press & Wilson, ; Pohar et al ., ; Liong & Foo, ). To reduce the number of variables, solve the problem of multicollinearity, and generate a set of continuous variables to avoid working on a large number of categorical characters that do not meet the criterion of normality in the case of the morphological data, we employed the PCA axes, which summarize the information included in the raw matrix, and are orthogonal, as variables in the LDA.…”