File size: 1,262 Bytes
cda703b | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 |
import pandas as pd
import numpy as np
import streamlit as st
import gc
@st.cache_data(show_spinner=False)
def get_baseline_default_rate(df,target_col='bad_loan'):
baseline_rate = df[target_col].mean()
return baseline_rate
@st.cache_data(show_spinner=False)
def calculate_woe_iv(df, feature, target='bad_loan'):
"""
Transforms a categorical variable into Weight of Evidence (WoE) scores
and computes the Total Information Value (IV) using vectorized operations.
"""
subset = df[[feature, target]].copy()
stats = subset.groupby(feature, observed=True)[target].agg([("Total", "count"), ("Bad", "sum")])
stats['Good'] = stats['Total'] - stats['Bad']
stats['Dist_Bad'] = stats['Bad'] / stats['Bad'].sum()
stats['Dist_Good'] = stats['Good'] / stats['Good'].sum()
stats['WoE'] = np.log((stats['Dist_Good'] + 0.001) / (stats['Dist_Bad'] + 0.001))
stats['IV_Contribution'] = (stats['Dist_Good'] - stats['Dist_Bad']) * stats['WoE']
woe_df = stats.sort_values(by='WoE', ascending=False).reset_index()
total_iv = woe_df['IV_Contribution'].sum()
del subset, stats
gc.collect()
return woe_df, total_iv
|