Abstract

Heart disease is a health problem of wide concern around the world. This paper aims to realize an early prediction of heart disease based on k-nearest neighbors algorithm to warn the users of their potential risks of heart disease in an early stage so that measures can be taken to minimize the danger. The dataset introduced contains basic information of body that is possibly related to heart disease, with 14 feature dimensions and 303 samples in total. To realize the proposition of early prediction, three different strategies for feature dimension choosing are introduced to evaluate the predictions based on all the data, data that can be measured at home with household devices and data that do not require any extra measurements and devices. Several indexes are introduced to evaluate and compare the performance of the models that are trained and the changes caused by the decrease of feature dimensions are analyzed. Finally, possible future work for improving the models is discussed.

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