ImranK/DeepLearningCreditApp

1 2import pandas as pd3import gradio as gr4import tensorflow5import tensorflow as tf6 7# define the function to make predictions8def predict_loan_approval(loan_amount, mortdue, prop_value, reason, job, yoj, derog, delinq, clage, ninq, clno, debtinc):9    # load the input data into a pandas dataframe10    input_df = pd.DataFrame({11        "LOAN": [loan_amount],12        "MORTDUE": [mortdue],13        "VALUE": [prop_value],14        "REASON_DebtCon": [int("DebtCon" in reason)],15        "REASON_HomeImp": [int("HomeImp" in reason)],16        "REASON_Other" :[int("Other" in reason)],17        "JOB_Mgr": [int("Mgr" in job)],18        "JOB_Office": [int("Office" in job)],19        "JOB_ProfExe": [int("ProfExe" in job)],20        "JOB_Sales": [int("Sales" in job)],21        "JOB_Self": [int("Self" in job)],22        "JOB_Other": [int("Other" in job)],23        "YOJ": [yoj],24        "DEROG": [derog],25        "DELINQ": [delinq],26        "CLAGE": [clage],27        "NINQ": [ninq],28        "CLNO": [clno],29        "DEBTINC": [debtinc]30    })31    # add dummy variables for reason and job32    if "DebtCon" in reason:33        input_df["REASON_DebtCon"] = 134    else:35        input_df["REASON_DebtCon"] = 036    if "HomeImp" in reason:37        input_df["REASON_HomeImp"] = 138    else:39        input_df["REASON_HomeImp"] = 040        41    if "Mgr" in job:42        input_df["JOB_Mgr"] = 143    else:44        input_df["JOB_Mgr"] = 045    if "Office" in job:46        input_df["JOB_Office"] = 147    else:48        input_df["JOB_Office"] = 049    if "Other" in job:50        input_df["JOB_Other"] = 151    else:52        input_df["JOB_Other"] = 053    if "ProfExe" in job:54        input_df["JOB_ProfExe"] = 155    else:56        input_df["JOB_ProfExe"] = 057    if "Sales" in job:58        input_df["JOB_Sales"] = 159    else:60        input_df["JOB_Sales"] = 061    if "Self" in job:62        input_df["JOB_Self"] = 163    else:64        input_df["JOB_Self"] = 065    66    # load the pretrained model67    model = tf.keras.models.load_model("credit_prediction_model.h5")68 69    # make a prediction70    prediction = model.predict(input_df)71    pred = prediction[0][0]72    # return the prediction as "Approved" or "Not Approved"73    if pred  > 0.3:74        return "Risk score: "+str(round(pred))+" :Decision: "+"Not Approved"75    else:76        return str(round(pred,4))+":"+"Approved" 77 78 79# define the Gradio interface80inputs = [81    gr.inputs.Number(label="Loan Amount (000)"),82    gr.inputs.Number(label="Amount due on Existing Mortgage"),83    gr.inputs.Number(label="Value of Current Property"),84    gr.inputs.CheckboxGroup(85        label="Reason",86        choices=["DebtCon", "HomeImp"]87    ),88    gr.inputs.CheckboxGroup(89        label="Job",90        choices=['Office', 'Sales', 'Mgr', 'ProfExe', 'Self','Other']91    ),92    gr.inputs.Number(label="Years at Present Job"),93    gr.inputs.Number(label="Number of Major Derogatory Reports"),94    gr.inputs.Number(label="Number of Delinquent Credit Lines"),95    gr.inputs.Number(label="Age of Oldest Credit Line in Months"),96    gr.inputs.Number(label="Number of Recent Credit Inquiries"),97    gr.inputs.Number(label="Number of Credit Lines"),98    gr.inputs.Number(label="Debt-to-Income Ratio")99]100 101outputs = gr.outputs.Textbox(label="Loan Approval")102 103iface = gr.Interface(104    fn=predict_loan_approval,105    inputs=inputs,106    outputs=outputs,107    title="Loan Approval Prediction",108    description="Enter the details of the loan application to check for eligibility",109    theme="dark" # default theme110)111 112#iface.launch(share=True)113#iface.theme_toggle()