An Integrated Machine Learning Approach for Congestive Heart Failure Prediction.
M Sheetal Singh1, Khelchandra Thongam1, Prakash Choudhary2
1Department of Computer Science and Engineering, National Institute of Technology Manipur, Langol, Imphal 795004, Manipur, India.
Diagnostics (Basel, Switzerland)
|April 13, 2024
Summary
Early detection of congestive heart failure (CHF) is crucial. This study uses machine learning, specifically a deep neural network, to accurately predict CHF, potentially reducing healthcare costs and improving patient outcomes.
Area of Science:
- Cardiology
- Artificial Intelligence
- Health Informatics
Background:
- Congestive heart failure (CHF) is a significant global health concern, affecting over 26 million people worldwide.
- The prevalence of CHF is increasing, highlighting the need for effective early detection and diagnosis methods.
- Current diagnostic costs can be reduced through advanced prediction techniques.
Purpose of the Study:
- To enhance the early diagnosis of congestive heart failure (CHF) using machine learning.
- To reduce the cost of CHF diagnosis by utilizing a minimum set of features for prediction.
- To compare the performance of a deep neural network (DNN) with other machine learning classifiers for CHF prediction.
Main Methods:
- Utilized the Cardiovascular Health Study (CHS) dataset for training and evaluation.
- Implemented a novel pre-processing technique integrating C4.5 for feature selection/outlier removal and K-nearest neighbor (KNN) for missing data imputation.
- Compared a deep neural network (DNN) classifier against six traditional machine learning algorithms (KNN, LR, NB, RF, SVM, DT) using seven statistical metrics.
Main Results:
- The proposed integrated approach, particularly the DNN, demonstrated superior performance in CHF prediction compared to other ML algorithms.
- Achieved high performance metrics: 97.03% F1-score, 95.30% accuracy, 96.49% sensitivity, and 97.58% precision.
- The study's methodology effectively reduced the number of medical tests required, indicating potential cost savings for patients.
Conclusions:
- The developed machine learning model, especially the DNN, offers a promising tool for accurate and cost-effective early prediction of congestive heart failure.
- The integrated pre-processing technique enhances data quality and model performance for cardiovascular health studies.
- Early prediction of CHF through advanced AI can significantly reduce mortality and morbidity associated with the condition.
Related Concept Videos
01:30Pulse rhythm
790
Pulse rhythm refers to the pattern of pulsations within specific intervals, offering valuable insights into the regularity or irregularity of the heart's beats as observed through the pattern of pulsation within specific intervals. A regular pulse exhibits a consistent heart rate with uniform waveforms and pulsation force, variations of which can be classified as normal, weak, or bounding.
Conversely, an irregular pulse pattern is termed dysrhythmia, stemming from disruptions in cardiac...
Conversely, an irregular pulse pattern is termed dysrhythmia, stemming from disruptions in cardiac...
790
01:17Pathophysiology of Heart Failure
1.6K
Heart failure (HF) is a progressive syndrome involving ventricles that leads to inadequate cardiac output. It can be classified based on location and output or ejection fraction. Ejection fraction (EF) is an essential measurement in the diagnosis and surveillance of HF. Reduced EF corresponds to systolic heart failure (HFrEF). However, HF with preserved ejection fraction (HFpEF) is becoming increasingly prevalent. Also known as diastolic HF, this form of HF is related to aging. The...
1.6K