Comparison of Linear Discriminant Analysis with Robust Linear Discriminant Analysis

Discriminant analysis is a multivariate technique related to separating distinct groups of objects and allocating new objects to predefined groups. Discriminant analysis produces a discriminant function, which is defined as a linear combination of independent variables used to classify objects into two or more groups or categories. Assumptions that must be met in linear discriminant analysis include the independent variables being multivariate normally distributed and the covariance matrices for each group being equal. It is also necessary to identify outliers in linear discriminant analysis, as the presence of outliers in the data set can cause the assumptions of linear discriminant analysis to be violated and result in inaccurate classification outcomes. Therefore, it is necessary to handle outliers using robust methods in discriminant analysis. One such robust method in discriminant analysis is the Minimum Covariance Determinant (MCD), which is highly effective in dealing with outliers and relatively easier to apply compared to other robust methods. The aim of this study is to compare the classification results of linear discriminant analysis with robust linear discriminant analysis on the dataset of diabetes patients at RSUD Padangsidimpuan in 2023. The results obtained from this dataset indicate that linear discriminant analysis achieved an accuracy of 85,71%, while robust linear discriminant analysis achieved an accuracy of 80,95%. These findings suggest that the use of linear discriminant analysis and robust linear discriminant analysis can yield different results depending on the characteristics of the data and the number of outliers in the dataset.