Improving Pneumonia Detection with AI
Author Information
Author(s): Chihaoui Mejda, Dhibi Naziha, Ferchichi Ahlem
Primary Institution: University of Ha'il, Hail, Saudi Arabia
Hypothesis
Can genetic algorithms enhance the performance of convolutional neural networks for pneumonia detection?
Conclusion
The study found that using genetic algorithms significantly improved the accuracy of pneumonia detection models.
Supporting Evidence
- The optimized CNN model achieved a training accuracy of 97% and a test accuracy of 94%.
- The error rate for the optimized models was low at 0.072.
- Genetic algorithms were effective in tuning hyperparameters and improving model performance.
Takeaway
This study shows that using smart computer programs can help doctors find pneumonia in X-ray pictures better and faster.
Methodology
The study used genetic algorithms to optimize the architecture and hyperparameters of convolutional neural networks and VGG16 for pneumonia detection.
Potential Biases
Potential overfitting due to optimization and limited dataset diversity.
Limitations
The dataset may not represent the full variability of pneumonia cases in real-world settings, which could affect model generalizability.
Statistical Information
P-Value
0.072
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website