Enhancing Stroke Diagnosis in the Brain Using a Weighted Deep Learning Approach
This work addresses stroke diagnosis for medical applications, but it is incremental as it combines existing classifiers without introducing a fundamentally new approach.
The study tackled the problem of costly and time-consuming stroke diagnosis by proposing a weighted voting ensemble model, which achieved 94.91% accuracy on a private dataset for predicting strokes.
A brain stroke occurs when blood flow to a part of the brain is disrupted, leading to cell death. Traditional stroke diagnosis methods, such as CT scans and MRIs, are costly and time-consuming. This study proposes a weighted voting ensemble (WVE) machine learning model that combines predictions from classifiers like random forest, Deep Learning, and histogram-based gradient boosting to predict strokes more effectively. The model achieved 94.91% accuracy on a private dataset, enabling early risk assessment and prevention. Future research could explore optimization techniques to further enhance accuracy.