Open AccessSpeculating the Threat of Cardiovascular Disease Using Classifiers with User-Focused Security Evaluations
Author(s) -
Chandrasekaran S.R.,
Dr.Sabiyath Fatima N.
Publication year - 2022
Publication title -
webology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.259
H-Index - 18
ISSN - 1735-188X
DOI - 10.14704/web/v19i1/web19372
Subject(s) - random forest , disease , computer science , machine learning , classifier (uml) , artificial intelligence , precision and recall , stacking , feature (linguistics) , data mining , medicine , linguistics , philosophy , physics , nuclear magnetic resonance
In recent decades, cardiovascular disease (CVD) is the most common type of disease that is prevailing all over the world. It is a class of diseases that involve the heart and its vessels. Strokes and heart attacks are normally critical events that are largely provoked by congestion that restricts blood from streaming to the parts of the body. The principle aim of this research is to find the feature that accounts for cardiovascular disease risks. The collection of data from the hospitals and laboratories can determine the risk of patients having cardiovascular disease by analysing the trends and correlations between the dataset. First, the data undergoes a security process that involves user-level security. The data is further processed to show the comparison between the 12 features and find out the top few features that account for the risk of positive cardiovascular disease. This Machine learning techniques can be widely used in the medical field, to determine the risk of cardiovascular disease early. These collected data from the patients may be used to warn them. To identify the risk of having cardiovascular disease early there are some proposed machine learning algorithms like K-nearest neighbours, XG Boost, Gradient boost and Random Forest Classifier by measuring the metrics like precision, accuracy, f-1 score and recall. Out of all these algorithms, XG Boost yields the highest accuracy of 90%. To increase the overall accuracy stacking algorithm is used to combine all the base learner algorithms to produce a result at once. After stacking the overall accuracy boosted to 92% for the respective dataset.