SVM Hyperplane Misclassification Control by Finding Optimum Cost of Misclassification with Boundary Value Analysis Technique

Abstract

In Health care domain roughly 80% of data in electronic medical records consists of physicians’ unstructured notes. To unlock this important data we need a different approach than what we used to analyse structured data. That’s one place where machine learning comes in and for this SVM (Support Vector Machine) is extensively used to identify the handwritten digits and words based on pixel given as features. SVM uses the concept of Hyper Planes which leads to a boundary which classifies the data set. Key area of the Research paper is to get Optimum Hyper plane.