Home / Current Issue / Paper 1708822
An Ensemble-Based Machine Learning Approach for Predicting Thoracic Disease Risk
Subject area: Science,Engineering and Technology · Area of research: Machine Learning
Abstract
Thoracic diseases, including both pulmonary and cardiac conditions, are some of the principal reasons for global mortality. Accurate and early risk assessment is essential to the proper diagnosis and management of the condition. This project presents an ensemble learning-based approach for thoracic disease risk prediction using two publicly available Kaggle datasets: one comprising chest X-ray images labeled with multiple lung conditions, and the other including organised clinical information for diagnosis of heart disease. The chest X-ray dataset includes critical conditions like Pneumonia, Pneumothorax, and Cardiomegaly, identified through radiographic features, while the heart disease datasets involves such as age, blood pressure, cholesterol, EKG. Data Preprocessing standard scaling and one-hot encoding. Deep neural network models among other ones and Random Forest classifiers were implemented. The system attained a general level of accuracy of 82.17%, validating the strength of ensemble techniques. This dual-modality approach enhances prediction accuracy and promises to fit into intelligent diagnostic systems.
Keywords
Ensemble Learning, Thoracic Disease, Heart Disease Prediction, Chest X-ray Classification, Deep Neural Network, Random Forest, Machine Learning, Clinical Data, Medical Imaging, Risk Assessment
How to cite this paper
@article{1708822,
author = {Kavitha D D, Ruqhaiya Asif , Sarisha N, Sneha K C, Soundarya Rajan},
title = {An Ensemble-Based Machine Learning Approach for Predicting Thoracic Disease Risk},
journal = {Iconic Research And Engineering Journals},
year = {2025},
volume = {8},
number = {11},
pages = {2065-2072},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1708822.pdf},
abstract = {Thoracic diseases, including both pulmonary and cardiac conditions, are some of the principal reasons for global mortality. Accurate and early risk assessment is essential to the proper diagnosis and management of the condition. This project presents an ensemble learning-based approach for thoracic disease risk prediction using two publicly available Kaggle datasets: one comprising chest X-ray images labeled with multiple lung conditions, and the other including organised clinical information for diagnosis of heart disease. The chest X-ray dataset includes critical conditions like Pneumonia, Pneumothorax, and Cardiomegaly, identified through radiographic features, while the heart disease datasets involves such as age, blood pressure, cholesterol, EKG. Data Preprocessing standard scaling and one-hot encoding. Deep neural network models among other ones and Random Forest classifiers were implemented. The system attained a general level of accuracy of 82.17%, validating the strength of ensemble techniques. This dual-modality approach enhances prediction accuracy and promises to fit into intelligent diagnostic systems.},
keywords = {Ensemble Learning, Thoracic Disease, Heart Disease Prediction, Chest X-ray Classification, Deep Neural Network, Random Forest, Machine Learning, Clinical Data, Medical Imaging, Risk Assessment},
month = {May},
}