Recent Advancements in Information and Communication Technology
Volume 1 | Issue 1 | Pages 01-13
Novel Automated Dataset Classification of Left Ventricular Hypertrophy Using Radial Basis Function Network & Support Vector Machine
S. Rajalaxmi 1 and S. Nirmala 1*
1 Muthayammal Centre for Advanced Research, Muthayammal Engineering College, Rasipuram, Namakkal, India.
An efficient classification of Left Ventricular Hypertrophy is proposed using Radial Basis Function Network and Support Vector Machine. This novel work is proposed with the advent of automation and quantitative aid to clinical experts in exact disease diagnosis. Four attributes namely sex, Left Ventricle End Diastolic Diameter (LVEDD), Intraventricular Septum Thickness (IVSD) and Posterior Wall dimension (PWD) are used for classification. Based on the parameters, calculated Left Ventricular Mass (Penn convention) and Relative Wall Thickness are given as inputs, and Radial Basis Function Network is trained in classifying the disease as Concentric Left Ventricular Remodeling, Concentric Left Ventricular Hypertrophy, Eccentric Left Ventricular Hypertrophy and no Left Ventricular Hypertrophy. Comparison of Classification results is done using Radial Basis Function network and Support Vector Machine, and training of the networks is achieved with 1625 input datasets. Accuracy of classification is compared with the diagnosis results of clinical experts and is witnessed as 98.5%.
Keywords: Radial Basis Function Network, Support Vector Machine, Left Ventricular Hypertrophy, Concentric Left Ventricular-Remodeling, Concentric Left Ventricular-Hypertrophy, Eccentric Left Ventricular-Hypertrophy