Related Experiment Videos
Neural networks for the prediction of spirometric reference values
1University of Athens, Faculty of Nursing, Laboratory of Health Informatics, Greece. tbotsis@nurs.uoa.gr
Medical Informatics and the Internet in Medicine
|December 12, 2003
Summary
This study introduces artificial neural networks for predicting lung function, specifically forced vital capacity (FVC) and forced expiratory volume in one second (FEV1). The novel method shows improved accuracy compared to traditional prediction equations in the elderly Greek population.
Area of Science:
- Pulmonary Medicine
- Biomedical Engineering
- Artificial Intelligence
Background:
- Traditional prediction equations for lung function parameters like FVC and FEV1 have limitations.
- Accurate prediction of lung function is crucial for diagnosing and managing respiratory diseases.
Purpose of the Study:
- To develop and evaluate a novel method using artificial neural networks (ANNs) for predicting FVC and FEV1.
- To achieve better correlation between predicted and actual lung function values.
Main Methods:
- Training multiple ANNs on a Greek elderly population sample, divided into training and testing sets.
- Studying males and females separately for four distinct prediction cases (FVC/FEV1 for males/females).
- Utilizing different network architectures, transfer functions, and neuron/layer counts for ANN training.
Main Results:
- Trained ANNs demonstrated strong correlation between predicted and measured FVC and FEV1 values in the testing dataset.
- The ANN models achieved a higher degree of accuracy compared to existing prediction equations from Baltopoulos et al. using the same sample.
- The developed ANNs show potential for predicting other spirometric parameters beyond FVC and FEV1.
Conclusions:
- Artificial neural networks offer a promising alternative for predicting lung function parameters.
- ANN-based prediction models provide enhanced accuracy for FVC and FEV1 compared to conventional methods.
- This approach has the potential to be extended to predict various other spirometric measures.