Developing a User-Friendly Machine Learning Model for Heart Disease Prediction

The World Health Organization estimates that millions of people die from heart disease every year, and it constitutes one of the major threats to health in the world. Due to its high morbidity and fatality rate, early prediction and prevention are very significant. Most existing heart disease prediction models are complicated in structure and operations, which cannot be used easily by non-professionals in daily life. This also increases the cost of learning and operation and makes the applications of models less popular. Overcoming this challenge, the researchers in this paper followed machine learning techniques in an attempt to establish a non-expert user-friendly, reliable, effective, low-cost heart disease risk prediction model. The proposed model enhances its ease of use and usability. This paper is based on many publicly available clinical datasets, using machine learning algorithms in pre-screening to identify important factors that make much contribution to estimating the risk of heart disease. These key risk factors will be used in training in the prediction model so that this model will not just have high predictive performance but also a simple structure that is easily understandable and operable. Compare several algorithms and optimize to find a best-suited model for inexperienced users, trying to achieve the best balance between accuracy and user experience. We sincerely hope that with the help of this model, more and more users will be able to manage their personal health, record their real-time heart conditions, proactively protect their health, and find out the risk of heart diseases early so as to reduce deaths caused by cardiovascular diseases.