Abstract—Cardiovascular disease is the main cause of fatality for women and men in developing and developed countries. Hypertension also known as high blood pressure is one of the major type of cardiovascular disease which causes more deaths than other cardiovascular diseases such as heart failure, coronary artery disease, arrhythmia, congenital heart disease etc. It is the time to take necessary steps to control death rate due to hypertension. In this paper, an approach to predict the hypertension based on handwritten manuscript using K-NN, Minimum distance and SVM classifiers has been presented. The proposed approach extracts geometric features such as number of right and left slants, number of horizontal and vertical lines, total length of horizontal and vertical lines, total length of left and right diagonal lines, and regional features such as Euler, major axis length and extend from a manuscript. Based on this information it predicts the existence of high blood pressure. The dataset contains total 140 manuscripts. The dataset is prepared using manuscripts of high blood pressure and control group people. The proposed system provides accuracy of 76.47% using K-NN classifier, 70.59% using minimum distance classifier, and 85.29% using SVM classifier.

 

Index Terms—high blood pressure, writing features, handwriting analysis, manuscript

Cite: Seema V. Kedar, D. S. Bormane, and S. C. Patil, "Performance Comparison of K-NN, Minimum Distance and SVM Classifiers Used to Predict Hypertension Based on Manuscript," International Journal of Signal Processing Systems, Vol. 5, No. 1, pp. 39-43, March 2017. doi: 10.18178/ijsps.5.1.39-43