A Novel Framework for Classification of Cardiovascular Disease Using Local Linear Embedding as Feature Selection Method

Abstract

A huge amount of biological data has been collected which is related to cardiovascular disease in healthcare industry and it’s increasing day by day. This enormous amount of data is in irregular form. So it is difficult to extract useful information in limited time and within affordable cost range. That’s why we need some dimensionality reduction feature methods to process this data with combination of data mining techniques to extract useful information from this huge amount of data. Data mining techniques such as Decision tree, Naïve Bayes, Neural network, K-Nearest neighbor and Random Forest has been previously used by many researchers for classification of cardiovascular patients. In this work, LLE used as feature selection method on two datasets; Cleveland and Statlog before applying classification methods, Decision Tree (DTree), Random Forest (RF), Support vector Machine (SVM) and Neural Network (NN) to predict the cardiovascular disease in patients. The results show that Random Forest is the best classifier for the prediction of cardiovascular disease with highest prediction accuracy for both Cleveland and Statlog heart disease datasets. It shows AUC-ROC as 1 with 80% training datasets. This framework will also be time effective and cost effective.