# Predict probabilities on testing data
predicted_probabilities <- predict(model2, newdata = test_data, type = "response")

# Convert predicted probabilities to binary predictions (1 if P>0.5, else 0)
predicted_classes <- ifelse(predicted_probabilities > 0.5, 1, 0)

# Compute confusion matrix for model evaluation
confusion_matrix <- table(Predicted = predicted_classes, Actual = test_data$upgraded)
print(confusion_matrix)

# Compute evaluation metrics
accuracy <- sum(predicted_classes == test_data$upgraded) / length(predicted_classes)
precision <- confusion_matrix[2,2] / sum(confusion_matrix[2,])
recall <- confusion_matrix[2,2] / sum(confusion_matrix[,2])

# Print evaluation metrics
print(paste("Accuracy: ", accuracy))
print(paste("Precision: ", precision))
print(paste("Recall: ", recall))