Abstract

AbstractEarly diagnosis of cardiovascular diseases is an extreme necessity in today’s world since cardiovascular diseases are the most prevalent reason for an increase in mortality rate. Computer-based prediction methods such as Machine learning techniques, Artificial intelligence, and other latest technologies are used to analyze the clinical data for early diagnosis of disease. Heart disease prediction model is developed with clinical data using machine learning techniques. Clinical data is generated in the healthcare industry from various sources like electronic health record, sensor data, hospital data, IoT and social media data. Because of its variety of sources, the healthcare data may consist of irrelevant, noisy, redundant data and also consists of a large number of features. All these have a significant impact on prediction model accuracy. The selection of right features is the main step in developing prediction models using machine learning algorithms. Feature selection is the method of removing redundant, noisy, and inappropriate data. Feature selection process selects the important features from the clinical dataset for developing machine learning models. Feature selection approach improves the correctness and efficiency of machine learning prediction models. This research paper deliberates about the various feature selection methods for selecting significant attributes and for eliminating inappropriate attributes in the dataset. Wrapper, Filter, and Embedded methods are analyzed and implemented using the Kaggle heart disease dataset in Python to find the major risk factor of heart disease. The objective of this research article is to find the major risk factor for heart disease. From the implementation of feature selection techniques, we found that chest pain, maximum heart rate, ST depression induced by exercise relative to rest, number of major vessels colored by fluoroscopy, and exercise-induced angina as the important risk factor for heart disease.KeywordsHeart diseaseFeature selectionOptimization techniquesMachine learningElectronic health recordArtificial intelligenceHealthcare

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call