Optimized Approach for Heart Disease Prediction by using Enhancing Diagnostic Method through Meta-Features

Abstract

In today's world, pollution is worsening, and bad habits like not eating regularly, eating lots of junk food, and not exercising enough are becoming more common. This can cause health problems, and the global need for the detection of heart diseases is increasing. The American Heart Association, in 2023, has given an update that heart disease is the leading cause of death. Implementing machine learning models has given significant results, but due to the limitations of requirement balanced data, repeat training model complexity has led to unreliable results in some cases. Hence, the proposed model overcomes the limitations by comparing machine learning models for cardiovascular diseases. Performance of proposed work is evaluated with the base models KNN and DT with and without smooth. Thus, the comparison included increased accuracy in DT, but the proposed model GBM as a meta learner has led the performance metric with 92% accuracy, with recall of 0.89% and F1 score of 0.86%. Thus, the proposed approach has achieved the highest accuracy in cardiovascular diseases using meta-subjects. Future developments for the research will focus on applying the model to a larger dataset and analysing cases according to complex machine learning models.