Diastolic Dysfunction Prediction with Symptoms Using Machine Learning Approach

Abstract

Cardiac disease is the major cause of deaths all over the world, with 17.9 million deaths annually, as per World Health Organization reports. The purpose of this study is to enable a cardiologist to early predict the patient’s condition before performing the echocardiography test. This study aims to find out whether diastolic function or diastolic dysfunction using symptoms through machine learning. We used the unexplored dataset of diastolic dysfunction disease in this study and checked the symptoms with cardiologist to be enough to predict the disease. For this study, the records of 1285 patients were used, out of which 524 patients had diastolic function and the other 761 patients had diastolic dysfunction. The input parameters considered in this detection include patient age, gender, BP systolic, BP diastolic, BSA, BMI, hypertension, obesity, and Shortness of Breath (SOB). Various machine learning algorithms were used for this detection including Random Forest, J.48, Logistic Regression, and Support Vector Machine algorithms. As a result, with an accuracy of 85.45%, Logistic Regression provided promising results and proved efficient for early prediction of cardiac disease. Other algorithms had an accuracy as follow, J.48 (85.21%), Random Forest (84.94%), and SVM (84.94%). Using a machine learning tool and a patient’s dataset of diastolic dysfunction, we can declare either a patient has cardiac disease or not.

Reliability Awareness Multiple Path Installation in Software Defined Networking using Machine Learning Algorithm

“…A sloping straight line is used to represent the relationship between the variables in the linear regression model [14]. In our work, we discuss the problem of a connection failure in software-defined networking and propose a novel machine learning-based solution to mitigate the link failure issue [15]. The traditional approaches for link failure are not efficient to handle link failure as it involves the identification of available alternative links, and it also needs to compute the reliability of the links.…”

Critical Review of Blockchain Consensus Algorithms: challenges and opportunities

“…• Records in Blockchain are irreversible as one record, when updated with some account, is linked to all previous records, hence forming a chain. Numerous approaches [16] and algorithms are implied to make sure that the coming record is stored in the database permanently in chronological order and is accessible to all users or nodes in the network. • Transactions in Blockchain are programmed and restricted to computation logic.…”