Mining Echocardiography Data to Predict Heart Disease (Medphd)
No Thumbnail Available
Date
2012-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
Background: These days, a major challenge of health care is reaching to correct
diagnosis of specific disease condition. Poor clinical decision leads to catastrophic
consequences which are unacceptable. Decision making process at the health care
setting needs to be supported with more advanced technology including a computer
based information system.
Objective: This study aims at extracting hidden knowledge (patterns and
relationships) associated with echocardiography datasets and designing a predictive
model for heart disease detection using data mining techniques.
Methods: A Hybrid Data Mining methodology is followed, which is a six-step
knowledge discovery process. The data for this research obtained from International
Cardiovascular Hospital in Addis Abeba, Ethiopia.
This study investigates the use of different data mining techniques, Decision tree,
neural network, and support vector machine for classification tasks. On
Transthoracic Echocardiography report datasets, descriptive data summarization
and visualization were taken to gain understanding of the data. Moreover, missing
values, outliers data, data integration and transformation were managed at
preprocess stage of hybrid process model.
Results: The results show that all the models performed well, though J48 Decision
tree algorithms outperforms support vector machine, Multilayer Perceptron Neural
Network, registering 96.73%. The best attributes selected by J48 decision tree are
Left Atrium Systole Diameter, LV ejection fraction, and Tricuspid velocity. As per
discussion made with the cardiologist, one of the interesting rule, a patient with Left
atrium systole diameter less than or equal to 40 millimeter and LV ejection fraction
less than or equal to 51% blood pumped out of ventricles and Tricuspid velocity is
greater than 2.5 centimeter per second results Left Ventricle dysfunction and
Pulmonary hypertensive disorder.
Conclusion: The result thus obtained in this study is promising to apply data mining
for heart disease detection. To make usable the knowledge extracted in this study,
an attempt has made to design a knowledge-based system that shows the potential
to integration. It is a further research direction.Keywords: Echocardiography, Knowledge discovery process, Decision tree, neural
network, Support vector machine
Description
Keywords
Echocardiography, Knowledge discovery process, Decision tree, neural network, Support vector machine