Fig 1.
Overall Architecture of the Proposed Method.
Table 1.
Stroke dataset and their values/types.
Fig 2.
Formation of multimodal dataset for stroke risk prediction.
Fig 3.
Federated MLP-GRU architecture for stroke risk prediction.
Fig 4.
Comparison of the number of individuals with and without stroke.
Fig 5.
Reading, resizing, normalizing, and visualizing images the MRI images.
Fig 6.
Comparison of Stroke Between (A) Gender vs stroke (B) Smoking Status vs Stroke.
Fig 7.
Number of individuals with and without stroke.
Fig 8.
Comparative analysis between (A) Age, (B) Glucose, (C) BMI, and (D) Hypertension Distribution among Individuals.
Fig 9.
Relationship between(A) Age, (B) Glucose Level, and Stroke Occurrence.
Fig 10.
Pairplot of Heart Disease, Stroke, Hypertension, BMI, Average Glucose Level, and Age.
Fig 11.
Correlation matrix of health parameters and stroke risk factors.
Fig 12.
Training and Validation Accuracy of (A) Client 1, (B) Client 2, (C) Client 3 and (D) Global Model.
Fig 13.
Training and Validation Loss of(A) Client 1, (B) Client 2, (C) Client 3 and (D) Global Model.
Fig 14.
Confusion matrix for stroke prediction.
Fig 15.
Actual vs predicted of the brain CT images.
Fig 16.
ROC curve for ML models.
Fig 17.
Evaluation metrics of the proposed method.
Table 2.
Performance evaluation of the proposed method with other existing approaches.
Fig 18.
Comparative evaluation of the suggested and current approaches.