Merge pull request #104 from sharath4444/temp

online payment fraud detection using ml algo

online_payment_fraud_detection/app.py

@@ -0,0 +1,45 @@
+import streamlit as st
+import numpy as np
+import pickle
+
+# Load the trained model
+with open("fraud_detection_model.pkl", "rb") as file:
+    model = pickle.load(file)
+
+# Streamlit app title and description
+st.title("Online Payment Fraud Detection System")
+st.markdown("""
+This application predicts whether an online payment transaction is fraudulent based on transaction details.
+Enter the transaction information below and click **Predict** to check for fraud.
+""")
+
+# Input section for transaction details
+st.subheader("Enter Transaction Details")
+transaction_type = st.selectbox("Transaction Type", ["CASH_OUT", "PAYMENT", "CASH_IN", "TRANSFER", "DEBIT"], help="Select the type of transaction.")
+amount = st.number_input("Transaction Amount", min_value=0.0, format="%.2f", help="Enter the transaction amount.")
+oldbalanceOrg = st.number_input("Original Balance (Before Transaction)", min_value=0.0, format="%.2f", help="Enter the account balance before the transaction.")
+newbalanceOrig = st.number_input("New Balance (After Transaction)", min_value=0.0, format="%.2f", help="Enter the account balance after the transaction.")
+
+# Map transaction types to numeric values
+transaction_map = {"CASH_OUT": 1, "PAYMENT": 2, "CASH_IN": 3, "TRANSFER": 4, "DEBIT": 5}
+transaction_type_num = transaction_map[transaction_type]
+
+# Predict fraud when button is clicked
+if st.button("Predict"):
+    # Prepare the input features for prediction
+    input_features = np.array([[transaction_type_num, amount, oldbalanceOrg, newbalanceOrig]])
+
+    # Perform prediction
+    prediction = model.predict(input_features)
+
+    # Display the result
+    if prediction[0] == "Fraud":
+        st.error("⚠️ This transaction is predicted as **Fraudulent**!")
+    else:
+        st.success("✅ This transaction is predicted as **Not Fraudulent**.")
+
+# Footer
+st.markdown("""
+---
+**Note:** This prediction is based on the trained model and may not be 100% accurate. Use this information as a guide, not a decision-making tool.
+""")