HEART DISEASE CLASSIFICATION USING ARTIFICIAL INTELLIGENCE, CNN
Date
2023-08-09
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Saudi Digital Library
Abstract
This thesis addresses the challenge of accurately classifying heart disease using a Heart Disease Classification model, which combines Convolutional Neural Networks (CNN) and BiLSTM. The study collects and preprocesses a comprehensive medical dataset related to heart disease. It then designs and trains CNN, RNN, and Combined models to capture complex patterns and temporal dependencies in the data. The results show that the Combined-2 (COMB-2) model outperforms others, achieving an accuracy of 0.964 and significant improvements in classification metrics. This research has implications for early diagnosis and improved healthcare in heart disease cases, showcasing the effectiveness of combining CNN and RNN models for classification.
Description
Heart disease remains a significant global health concern, necessitating accurate and efficient diagnostic approaches. This thesis addresses the challenges by proposing a Heart Disease Classification model, a combined model for prediction based on Convolutional Neural Networks (CNN), and Bi-LSTM. The goal is to create a solid and trustworthy system that may assist in the early identification and detection of heart- related disorders. The research problem revolves around accurately classifying heart disease based on medical data. Given the complexity and high-dimensional nature of heart-related datasets, traditional classification techniques face challenges. Therefore, this study explores the utilization of CNN, RNN, and Bi-LSTM models to capture intricate patterns and temporal dependencies in the data. The methodology involves several steps. Firstly, a comprehensive dataset containing medical records and associated features related to heart disease is collected and pre-processed. Next, CNN, RNN, and Combined models are designed and trained on the dataset, leveraging their respective architectures to learn discriminative features and capture temporal information. The models are optimized using suitable loss functions and evaluated using various performance metrics on MRI dataset. The results obtained from the evaluation demonstrate the effectiveness of the proposed Heart Disease Classification models. Among the evaluated models, the Combined-2 (COMB-2) model achieves the highest performance metrics, including accuracy, precision, recall, and F1 score, while maintaining a low loss value. COMB-2 achieves an accuracy of 0.964, precision of 0.962, recall of 0.965, F1 score of 0.963, and a loss of 0.0791. The significance of these findings lies in the potential impact on healthcare. The COMB-2 model showcases superior performance in accurately classifying heart disease cases, outperforming the individual CNN and RNN models. The contributions of this research extend to the field of medical diagnosis, providing insights into the efficacy of combined CNN and RNN models for heart disease classification. The information might help doctors make an early diagnosis, which would allow for prompt treatment and better prediction for the disease.
Keywords
HEART DISEASE CLASSIFICATION USING CNN, ARTIFICIAL INTELLIGENCE
Citation
1- Siontis, K. C., Noseworthy, P. A., Attia, Z. I., & Friedman, P. A. (2021). Artificial intelligenceenhanced electrocardiography in cardiovascular disease management. Nature Reviews
Cardiology, 18(7), 465–478
ISSN: 17595010, 17595002
Identifier: DOI: 10.1038/s41569-020-00503-2
page number: 18
Type: Journal
Language: English
2- Wang, H., Wei, J., Zheng, Q., Meng, L., Xin, Y., Yin, X., & Jiang, X. (2019). Radiationinduced heart disease: a review of classification, mechanism and prevention. International
journal of biological sciences, 15(10), 2128.
SSN: 14492288
Identifier: DOI: 10.7150/ijbs.35460
page number: 10
Type: Journal
Language: English