I already build one data model in that around 30 to 40 lacs of records are available. In that I have number of predictors like channel, Product, Agent, Proofs submitted by customers, Age, etc. From this model I am getting number of customers are withdrawing the policy.
So I want the my model should tell me if the certain combination is get's repeated, then it should raise a alarm for me.
For e.g. From Total data size of 10000 and from that If Total 100 customers withdrawing the policies have the Address Proof PAN card, Purchase product CUM, Agent Number is 0001, Branch code is 113, Age bucket is 30 To 40 years.
So I want out-put like below in one sheet...
TOP 10 combinations (on the basis of %) that constitute withdrawn policies.
Address Proof
Product
Agent Number
Branch code
Age Bucket
%
PAN card
CUM
123
1234
30 To 40 Years
30%
Driving license
AUM
234
3456
50 To 60 Years
20%
If any new policy received with that combinations then it should tell me the count in one Table.
Address Proof
Product
Agent Number
Branch code
Age Bucket
Total Policies Received
PAN card
CUM
123
1234
30 To 40 Years
1000
Driving license
AUM
234
3456
50 To 60 Years
5000
If you provide any application for reference then it is vary useful.