added confidences to predictions

app.py

@@ -19,7 +19,7 @@ def train_iris_model(algorithm):
     if algorithm == 'KNN':
         model = KNeighborsClassifier()
     elif algorithm == 'SVM':
-        model = SVC()
+        model = SVC(probability=True)
     elif algorithm == "logistic regression":
         model = LogisticRegression()
     elif algorithm == 'Random Forest':
@@ -28,6 +28,7 @@ def train_iris_model(algorithm):
         model = AdaBoostClassifier()
     elif algorithm == 'Decision tree':
         model = DecisionTreeClassifier()
+
     model.fit(X_train, y_train)
 
     return model
@@ -39,7 +40,12 @@ def predict_iris_species(model, input_data):
     # Make predictions using the trained model
     prediction = model.predict(input_data)
 
-    return prediction
+    # Check if the model has a predict_proba method
+    if hasattr(model, 'predict_proba'):
+        confidence = model.predict_proba(input_data).max()
+        return prediction, confidence
+    else:
+        return prediction, None
 
 def main():
     st.title("Iris Species Prediction App")
@@ -50,7 +56,6 @@ def main():
     # Train the model based on user's choice
     trained_model = train_iris_model(algorithm)
 
-
     st.sidebar.header("User Input")
     sepal_length = st.sidebar.slider("Sepal Length", 0.0, 10.0, 5.0)
     sepal_width = st.sidebar.slider("Sepal Width", 0.0, 10.0, 5.0)
@@ -58,7 +63,7 @@ def main():
     petal_width = st.sidebar.slider("Petal Width", 0.0, 10.0, 5.0)
 
     input_values = [sepal_length, sepal_width, petal_length, petal_width]
-    prediction_result = predict_iris_species(trained_model, input_values)
+    prediction_result, confidence = predict_iris_species(trained_model, input_values)
 
     species_mapping = {0: 'Iris-setosa', 1: 'Iris-virginica', 2: 'Iris-versicolor'}
     predicted_species = species_mapping.get(prediction_result[0], 'Unknown')
@@ -72,6 +77,9 @@ def main():
     st.subheader("Prediction:")
     st.success(f"Predicted Species: {predicted_species}")
 
+    if confidence is not None:
+        st.info(f"Confidence of prediction: {confidence * 100:.2f}%")
+
     # Display relevant images based on prediction
     if predicted_species == 'Iris-setosa':
         st.image('setosa_image.jpg', caption='Iris-setosa', use_column_width=True)