Volume 18, No. 4, 2021
Predicting Heart Disease Using Feature Selection Techniques Based On Data Driven Approach
S.USHA , Dr.S.KANCHANA
Abstract
Machine learning techniques, a type of artificial intelligence, are being used in the health field to assist researchers in recognizing pathology before it becomes a major problem. Because healthcare is such an important aspect of a country's economy, researchers are exploring the level of uncertainty that arises when using machine learning algorithms for ways to anticipate the disease. The most significant concept in health data analysis is the prediction of cardiac disease from clinical data. The prediction helps physicians to take exact decisions regarding patients' health. The proposed model used Data collection, Data pre-processing, and Data Transformation methods to train the model. This model exploited feature selection methods: filter and wrapper with classification techniques to enhance the prediction of cardiac disease classification. The classification techniques, namely: Decision Tree, Logistic Regression, Random Forest, and Ada Boost are pragmatic to evaluate performance metrics. The performance metrics include Accuracy, F1-score, Precision, Sensitivity; Specificity reveals an improvement in the outcomes of the prediction.
Pages: 97-108
Keywords: Decision Tree, Logistic Regression, Random Forest, and Ada Boost, Filter, Wrapper