The Predictability of the Likeliness of Having a Heart Attack

Prisha Mathur

Abstract

Abstract: 

Heart Disease is the leading cause of deaths in the entire world. They are happening to more than 17.9 million people each year, and each year is a higher increase for heart diseases.  With using an accurate model to predict the likelihood of having a heart attack, you can get the help needed, and prevent further healthcare problems. This paper uses the logistic regression machine learning method to show the Predictability of the Likeliness of a Heart Attack. With AI models increasing in their accuracy with predictions, models are becoming safer. The methodology used for this dataset is to utilize the user’s information, and uses specific equations of linear models, neural networks, and decision tree to find the weight of the attributes, and evaluate the performance on the tested. For example, the neural network accuracy is 0.81 There are 14 attributes tested in this dataset. The model will use the user’s information to predict the likeliness of the heart attack. The methodology employing this is using logistic regression. First the model will predict the likeness of a heart attack based on the columns provided. It will use the neural network, decision tree, and linear model to predict, find out the weight of the attributes, and tell the model threshold predictability from 0 or 1.The significance in accurate prediction is so there isn’t a lack of trust with the patient and this dataset. If the data is not accurate, it can lead to misjudgement of the patient’s information.