Table 3 Pseudo code.
From: Responsible CVD screening with a blockchain assisted chatbot powered by explainable AI
START |
|---|
// Step 1: Data Loading |
Load dataset from the source |
// Step 2: Data Preprocessing |
Perform EDA |
Remove duplicates and check for null values |
Handle missing values if any |
Feature selection based on relevance |
Apply SMOTE for handling class imbalance (if needed) |
Normalize/Standardize numerical features |
// Step 3: Splitting Data |
Split data into Training set (70%) and Testing set (30%) |
// Step 4: Model Implementation |
Select an ML model (e.g., XGBoost) |
Train the model on the Training set |
Evaluate the model using the Testing set |
// Step 5: Model Evaluation |
Calculate performance metrics (e.g., Accuracy, Precision, Recall) |
Check if the model detects ‘CVD’ or ‘No CVD |
IF model performance is satisfactory THEN |
Proceed to XAI implementation |
ELSE |
Modify model or preprocessing steps and retrain |
// Step 6: XAI Implementation (e.g., SHAP & LIME) |
Generate explanations for model predictions using XAI techniques |
Display or store the explanations |
// Step 7: Make Predictions |
Input new data into the trained model |
Predict whether the patient has ‘CVD’ or ‘No CVD’ |
Display predictions and explanations |
// Step 8: Decision and Action |
IF ‘Health Disease’ is predicted THEN |
Recommend further medical consultation |
ELSE |
No further action required |
END |
