Nautilus Systems, Inc. logo and menu bar Site Index Home
News Books
Button Bar Menu- Choices also at bottom of page About Nautilus Services Partners Case Studies Contact Us
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [Subscribe]

DM: RE:


From: Sarabjot S. Anand
Date: Thu, 9 Apr 1998 09:19:33 -0400 (EDT)
I would tend to support Hypothesis A. It also does depend on your 
validation technique. 

However, I am a bit concerned anout the non-churn customers that you 
used to build the model. The concern is similar to that discussed in 
our paper "A methodology for Cross Sales using Data Mining" available 
from http://inchinn.infj.ulst.ac.uk/htdocs/reports.html. How did you 
define non-chrun? If they were current customers - then tou have a 
problem as some of these are potentially churn customers that haven't 
executed their move as yet. How do you get around this? The model you 
have built differentiates between those who have left and those who 
have not and is subtly different from what you are trying to do - 
your current customers are your target set not the non-churn set.
Sarab

-----Original Message-----
From:   Hallberg Rassy [SMTP:hr38@hotmail.com]
Sent:   09 April 1998 11:23
To:     datamine-l@nautilus-sys.com
Subject:        

Dear Friend

I have now a problem and would like to share with you and to receive, 
if 
possible your opinion.
Hereinafter the picture of situation:
1.I'm presently involved in a customer profiling project for a large 
mobile operator.
2.The goal is to set up a system able to anticipate the likelyhood of 
churn of  customers
3.As a pilot step I extracted call records for 10000 active customers 
plus 4000  churned
4.Using SPSS neural connection I made up a neural network based on a 
set 
of 4000  active+4000 churned
5.The data was: calling patterns of july, agoust and september the 
target was:  churn/no churn situation in december
6.The results was promising: 90% of real churn anticipated, with a 
cut-off  probability of 80%
7.The same network was used on october, nov, dec. data to anticipate 
march churn  the results dropped to a terrific 11% with the same 
cut-off 
of 80%: totally  useless


I have formulated some hypotheses
A.The low time span (three month) is affected by seasonality
B.The data used are not sufficient to build a reliable network
C.The tool (SPSS Neural Connection) is not reliable

Could you give your opinion?
Many thanks in advance



______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com




[ Home | About Nautilus | Case Studies | Partners | Contact Nautilus ]
[ Subscribe to Lists | Recommended Books ]

logo Copyright © 1998 Nautilus Systems, Inc. All Rights Reserved.
Email: nautilus-info@nautilus-systems.com
Mail converted by MHonArc 2.2.0