Stroke Detection System

• Tech Stack: Python, BiLSTM, Genetic Algorithms, HTML/CSS/JS
• Github URL: GitHub Link

• Engineered an advanced diagnostic stroke model by integrating Genetic Algorithm-driven feature selection with BiLSTM neural networks to process complex clinical and neuroimaging datasets.
• Optimized high-performance classification as measured by 94.05% accuracy, a 0.9929 ROC AUC, and a 96.15% recall rate by tuning hyper-parameters to minimize critical false negatives.
• Developed an interactive web interface using HTML, CSS, and JavaScript to visualize real-time risk assessments, significantly enhancing clinical accessibility and user experience.