Bank-Scheme-Subscription-Prediction

Logistic regression model focusing on significant features extraction using Somers' D and VIF (Variance Inflation Factor).

MODEL DETAILS

Logistic Regression : statsmodels.formula.api.logit
CUT-OFF : 0.43 (BEST Accuracy)
TRAIN ACCURACY : 0.91
TEST ACCURACY : 0.91
TEST PRECISION : 0.63
TEST RECALL : 0.43
TEST F1-SCORE : 0.51

SIGNIFICANT VARIABLES

Based on the analysis and the Logistic Regression Model evaluation, while campaigning for new Savings Scheme the bank may focus on :

1. previous_savings : number of contacts performed before this campaign and for this client (numeric)
2. cons_price_idx : consumer price index. monthly indicator (numeric)
3. nr_employed : number of employees. quarterly indicator (numeric)
4. duration : last contact duration, in seconds (numeric)