Paper Title
Machine Learning based Cardiovascular Disease Diagnosis using Feature Selection

Abstract
It's important to determine the seriousness of cardiovascular disease in a patient. Machine learning techniques can solve this problem by considering risk and its connection with cardiovascular diseases. The majority of machine learning algorithms are influenced by dimensionality, whereby the time needed for a model with training data is reduced with an increasing number of features. However, the model tends to overfit, as data increases with the additional features. Feature selection, also known as the feature rankings, is used to avoid the problems with unnecessary features by reducing features in data sets with many functions. The group uses the naive approach of feature selection, evaluating the performance of each element in the powerset of the feature set. Performance is taken as the weighted average of each non-diagonal cell in a confusion matrix, so that more severe false negative cases have a higher weight. In the case of assessing heart disease, each feature is extracted using medical procedures with varying cost. This is taken into account when selecting the best feature set, so as to obtain the feature set that is most effective computationally, and in terms of cost. The group found that (feature set) is the most effective in terms of performance, saving up to (result) in medical costs per diagnosis. Keywords - Machine Learning, Feature Selection, Cardiovascular Disease, Cleveland Heart Dataset, SVM