Discriminant analysis for heart disease attributes
Author(s): Rajarathinam A
Abstract:
This study employs multivariate analysis to investigate the relationship between heart disease parameters and outcomes. Rigorous assessment of normality, multicollinearity, and covariance matrix equality ensures analysis validity. Data normalization via the Box-Cox method enhances normality, facilitating robust statistical analyses. Multivariate analysis of Variance (MANOVA) uncovers significant heart disease parameter variations across Outcome variables. Linear discriminant analysis (LDA) assesses heart disease parameters' capacity to classify individuals, emphasizing gender as a discriminating factor. Results highlight the importance of heart disease parameters in understanding population characteristics and their implications for medical research and clinical practice. The confusion matrix reflects the classification accuracy of a heart disease prediction model, achieving 72.9% overall accuracy in distinguishing between individuals with and without heart disease.