Your Client FinMan is a financial services company that provides various financial services like loan, investment funds, insurance etc. to its customers. FinMan wishes to cross-sell health insurance to the existing customers who may or may not hold insurance policies with the company. The company recommend health insurance to it's customers based on their profile once these customers land on the website. Customers might browse the recommended health insurance policy and consequently fill up a form to apply. When these customers fill-up the form, their Response towards the policy is considered positive and they are classified as a lead.
Once these leads are acquired, the sales advisors approach them to convert and thus the company can sell proposed health insurance to these leads in a more efficient manner.
Now the company needs your help in building a model to predict whether the person will be interested in their proposed Health plan/policy given the information about:
- Demographics (city, age, region etc.)
- Information regarding holding policies of the customer
- Recommended Policy Information
Submissions are evaluated on area under the ROC curve between the predicted probability and the observed target.
Test data is further randomly divided into Public (40%) and Private (60%) data. Your initial responses will be checked and scored on the Public data. The final rankings would be based on your private score which will be published once the competition is over.
- Public LB : 19th Rank
- Private LB : 12th Rank
- Feature Generation using pandas agg
- Feature selection using feature Importance and SHAP values
- Removing correlation
- Stratified KFolds prediction for LGBM and CatBoost
- Ensemble logic