G-Blend Network Development and Deployment

Dataset Collection

Gather clinical and demographic data for stroke prediction

Graph Construction

Build patient similarity graph using KNN and Euclidean distance

GAT Layer Implementation

Refine embeddings using attention scores between nodes

Model Evaluation

Evaluate G-Blend Network on hold-out test set using classification metrics

Data Preprocessing

Impute missing values, scale numerical features, encode categorical variables, and apply SMOTE

GCN Layer Implementation

Apply GCN to update node embeddings based on graph structure

OSGB Integration

Use GAT embeddings as input to DSGB for stroke risk prediction

API Deployment

Deploy G-Blend model as Flask RESTful API for real-time predictions

Flutter App Development

Develop cross-platform Flutter app for user interaction and risk display