Add files via upload

Author: hassankhan2608

SVC.pkl

@@ -0,0 +1,92 @@
+import streamlit as st
+import pickle
+import numpy as np
+import pandas as pd
+from sklearn.preprocessing import StandardScaler
+
+# Load the LabelEncoder
+with open('encoder.pkl', 'rb') as f:
+    encoder = pickle.load(f)
+
+# Load the StandardScaler
+with open('standard_scaler.pkl', 'rb') as f:
+    scaler = pickle.load(f)
+
+# Load the SVC model
+with open('SVC.pkl', 'rb') as f:
+    model = pickle.load(f)
+
+# Define the UI
+st.set_page_config(page_title="Customer Churn Prediction", layout="wide")
+st.header("Customer Churn Prediction App")
+st.write("This app predicts whether a customer is likely to churn based on their demographic and banking information.")
+
+# Define the categorical feature mappings
+categorical_mappings = {
+    'Geography': ['France', 'Germany', 'Spain'],
+    'Gender': ['Female', 'Male'],
+    'Has Credit Card': ['No', 'Yes'],
+    'Is Active Member': ['No', 'Yes']
+}
+
+# Input widgets for categorical features
+col1, col2 = st.columns(2)
+selected_category = {feature: col1.selectbox(f'{feature}', options, index=None) for feature, options in list(categorical_mappings.items())}
+
+# Input widgets for numerical features
+col3, col4, col5, col6, col7, col8 = st.columns(6)
+creditscore = col3.slider('Credit Score', 300, 850, 500)
+age = col4.slider('Age', 18, 100, 30)
+tenure = col5.slider('Tenure', 0, 10, 5)
+balance = col6.number_input('Balance', min_value=0.0, max_value=250000.0, step=1000.0, value=0.0)
+num_of_products = col7.slider('Number of Products', 1, 4, 2)
+estimated_salary = col8.number_input('Estimated Salary', min_value=0.0, max_value=1000000.0, step=1000.0, value=0.0)
+
+# Map 'Yes'/'No' to 1/0 for 'Has Credit Card' and 'Is Active Member'
+has_cr_card = 1 if selected_category['Has Credit Card'] == 'Yes' else 0
+is_active_member = 1 if selected_category['Is Active Member'] == 'Yes' else 0
+
+# Button to predict
+if st.button('Predict'):
+    if None in selected_category.values():
+        missing_categories = [key for key, value in selected_category.items() if value is None]
+        missing_categories_str = ", ".join(missing_categories)
+        st.error(f"Please select values for the following categorical features: {missing_categories_str}")
+    else:
+        # Encode categorical features
+        encoded_features = []
+        for feature, value in selected_category.items():
+            # Skip 'Has Credit Card' and 'Is Active Member'
+            if feature in ['Has Credit Card', 'Is Active Member']:
+                continue
+            # Create a DataFrame with the selected category
+            df = pd.DataFrame({feature: [value]})
+            # One-hot encode the categorical feature
+            encoded_df = pd.get_dummies(df, columns=[feature], drop_first=True)
+            # Convert DataFrame to numpy array
+            encoded_value = encoded_df.values
+            # Append the encoded value
+            encoded_features.append(encoded_value)
+
+        # Scale numerical features
+        numerical_features = [creditscore, age, tenure, balance, num_of_products, estimated_salary]
+        scaled_features = scaler.transform([numerical_features])
+
+        # Combine encoded categorical features and scaled numerical features
+        input_data = np.concatenate(encoded_features + [scaled_features], axis=1)
+
+        # Append 'Has Credit Card' and 'Is Active Member'
+        input_data = np.append(input_data, [[has_cr_card, is_active_member]], axis=1)
+
+        # Make prediction
+        prediction = model.predict(input_data)[0]
+        prediction_text = 'Yes' if prediction == 1 else 'No'
+        if prediction_text == 'Yes':
+            styled_prediction = '<span style="font-size: 20px; color: green; font-weight: bold;">Yes</span>'
+        else:
+            styled_prediction = '<span style="font-size: 20px; color: red; font-weight: bold;">No</span>'
+        
+        # Display prediction with a subheader
+        st.subheader("Prediction Result")
+        st.markdown(f"The Customer is likely to churn: {styled_prediction}", unsafe_allow_html=True)
+