Enhanced Predictive Model to Diagnose Coronary Heart Disease Using Decision Tree
As per WHO reports, Heart diseases or Cardiovascular Diseases (CVDs) rank as the leading cause of death worldwide. Approximately 17.7 million individuals succumbed to CVDs in 2015, accounting for 31% of all global fatalities. Among these fatalities, an estimated 7.4 million were attributed to corona...
Saved in:
Published in | 2024 International Conference on Data Science and Network Security (ICDSNS) pp. 1 - 5 |
---|---|
Main Authors | , , |
Format | Conference Proceeding |
Language | English |
Published |
IEEE
26.07.2024
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | As per WHO reports, Heart diseases or Cardiovascular Diseases (CVDs) rank as the leading cause of death worldwide. Approximately 17.7 million individuals succumbed to CVDs in 2015, accounting for 31% of all global fatalities. Among these fatalities, an estimated 7.4 million were attributed to coronary heart disease, one of the prevailing forms of heart disease.. According to the Global Burden of Disease study, nearly quarters (24.8%) of all deaths in India are due to CVD. Hence there is a need for careful and periodic methods of examining heart related diseases. The majority of cardiovascular diseases found in India are coronary heart diseases (CHD) The main contribution of this work is to develop efficient predictive models for Coronary heart diseases using decision trees. In this work, the predictive models are developed for diagnosis of heart disease using the cardiovascular disease dataset from UCI repositories with the objective of improving the accuracy. We have also obtained echocardiograph data obtained from M. S. Ramaiah hospital and used in this work. It is found that the data used in this work is heterogeneous in nature and the decision trees work well on heterogeneous data and also they are expressive and interpretative, we have opted to use decision trees. The predictive model namely Enhanced_Pred_Model CVD-DATA-DT is developed using both the datasets. It is found that the model shows marginal accuracy increase from 77% in baseline model to 80% by the proposed model. The proposed method has better performance in terms of accuracy as well as error rate. The models are also tested on the echocardiograph data for prediction of LVH event and has shown improvement in the performance. |
---|---|
DOI: | 10.1109/ICDSNS62112.2024.10690975 |