Internshala-Data-Science-Challenge

HACKATHON powered by ANALYTICS VIDHYA

Problem Statement

Your client is an Insurance company and they need your help in building a model to predict whether the policyholder (customer) will pay next premium on time or not.

Variable Definition

1. id : Unique ID of the policy

2. perc_premium_paid_by_cash_credit : Percentage of premium amount paid by cash or credit card

3. age_in_days : Age in days of policy holder

4. Income : Monthly Income of policy holder

5. Count_3-6_months_late : No of premiums late by 3 to 6 months

6. Count_6-12_months_late : No of premiums late by 6 to 12 months

7. Count_more_than_12_months_late : No of premiums late by more than 12 months

8. application_underwriting_score : Underwriting Score of the applicant at the time of application (No applications under the score of 90 are insured)

9. no_of_premiums_paid : Total premiums paid on time till now

10. sourcing_channel : Sourcing channel for application

11. residence_area_type : Area type of Residence (Urban/Rural)

target : 1 - premium paid on time , 0 - otherwise

Model Report

Accuracy : 95.5%

AUC Score (Train): 0.882333

SECURED COMPETITION RANK - 9