streamlit_regression.py

import pickle
import streamlit as st 
import tensorflow as tf
from sklearn.preprocessing import StandardScaler, LabelEncoder, OneHotEncoder
import pandas as pd
import numpy as np

## Load the trained Model
model = tf.keras.models.load_model('regression_model.h5')

with open('lable_encoder_gender.pkl','rb') as file:
    lable_encoder_gender = pickle.load(file)
    
with open('one_hot_encoder_geo.pkl','rb') as file:
    onehot_encoder_geo = pickle.load(file)

with open('scaler.pkl','rb') as file:
    scaler = pickle.load(file)
    
## Streamlit app
st.title("Customer Churn Salary Predicition")

## user input

geography = st.selectbox('Geography',onehot_encoder_geo.categories_[0])
gender = st.selectbox('Gender',lable_encoder_gender.classes_)
age = st.slider('Age',18,92)
balance = st.number_input('Balance')
credit_score = st.number_input('Credit Score')
exited = st.selectbox('Exited',[0,1])
tenure = st.slider('Tenure',0,10)
num_of_products = st.slider('Number of Products',1,4)
has_cr_card = st.selectbox('Has Credit Card',[0,1])
is_active_member = st.selectbox('Is Active Member',[0,1])

## prepare the input data
input_data = pd.DataFrame({
    'CreditScore' : [credit_score],
    'Gender' : [lable_encoder_gender.transform([gender])[0]],
    'Age' : [age],
    'Tenure' : [tenure],
    'Balance' : [balance],
    'NumOfProducts' : [num_of_products],
    'HasCrCard' : [has_cr_card],
    'IsActiveMember' : [is_active_member],
    'Exited' : [exited]
})

## One Hot Encoding GeoGraphy
geo_encoded = onehot_encoder_geo.transform([[geography]]).toarray()
geo_encoded_df = pd.DataFrame(geo_encoded,columns=onehot_encoder_geo.get_feature_names_out(['Geography']))

## combine the data
input_data = pd.concat([input_data.reset_index(drop=True),geo_encoded_df],axis=1)

## Scale the input data
scaled_data = scaler.transform(input_data)

## prediction churn
prediction = model.predict(scaled_data)
predicted_salary = prediction[0][0]

st.write(f"Predicted Salary : {predicted_salary}")