Skip to content

Predicted the likelihood that each insurance policyholder will make a service payment call | Comparative analysis of multiple ML models

Notifications You must be signed in to change notification settings

wajahat1064/InsuranceCompany_ClassImbalance

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

16 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Insurance Company Class Imbalance Project

Business Problem:

Insurance Company provides several ways for its policyholders to make payments. While the Insurance Company's service counselors can take payments over the phone, it is more cost-efficient for customers to make payments through their self-service channels such as online or through the automated phone system. Thus, they would like to use a predictive model to select people to receive a pre-emptive e-mail message designed to encourage them to pay online.

Task at hand:

We have been tasked with identifying which customers are likely to make a service payment call in the next 5 days. The Datasets folder contains highly imbalanced data on customers who have had a bill due in the next 5 days and whether they made a service payment call.

Predictive Modeling:

We will construct a model that predicts the likelihood that each policyholder will make a service payment call (CALL_FLAG=1).

Data Dictionary:

Var Name Description
DATE_FOR Date of Record Processing
Content Cell Content Cell
RTD_ST_CD Rated State of Policy
CustomerSegment Segment the Customer falls into
Tenure Years of Tenure with Company
Age Age of Policyholder
MART_STATUS Marital Status of Policyholder
GENDER Gender Of Primary Insured
CHANNEL1_6M # payments made through Channel 1 in last 6 months
CHANNEL2_6M # payments made through Channel 2 in last 6 months
CHANNEL3_6M # payments made through Channel 3 in last 6 months
CHANNEL4_6M # payments made through Channel 4 in last 6 months
CHANNEL5_6M # payments made through Channel 5 in last 6 months
METHOD1_6M # of payment made with method 1 (irrespective of channel)
RECENT_PAYMENT Payment made in last 15 days (1/0)
PAYMENTS_6M # of total payments in last 6 months
CHANNEL1_3M # payments made through Channel 1 in last 3 months
CHANNEL2_3M # payments made through Channel 2 in last 3 months
CHANNEL3_3M # payments made through Channel 3 in last 3 months
CHANNEL4_3M # payments made through Channel 4 in last 3 months
CHANNEL5_3M # payments made through Channel 5 in last 3 months
METHOD1_3M # of payment made with method 1 (irrespective of channel)
PAYMENTS_3M # of total payments in last 3 months
NOT_DI_3M Had this customer been enrolled in automated payments in the last 3 months? 1/0
NOT_DI_6M Had this customer been enrolled in automated payments in the last 6 months? 1/0
EVENT1_30_FLAG Has this customer been sent a cancellation notice in the last 30 days? 1/0
EVENT2_90_SUM How many cancellation notices have been sent in the last 90 days?
LOGINS How many times has this policy logged into self-service online in the last 30 days?
POLICYPURCHASECHANNEL How was this policy purchased? 1/0
------------- -------------
Call_Flag Was there a service payment 1/0? TARGET VARIABLE

About

Predicted the likelihood that each insurance policyholder will make a service payment call | Comparative analysis of multiple ML models

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published