Anil,
This code seems similar to something I've done with CEM, so I think that the syntax
is right when it comes to specifying cutpoints.
Looking at your note, it seems like you want them all to be in the "same bins."
I just want to make sure that you mean that you want to have a similar way to break up
each of these revenue variables - from the names of the vars, it looks like you are doing
lags. I think this makes sense if the revenue vars have a similar distribution from
period to period - you are looking for them to have a similar binning structure, but not
be in the "same bins," right? Each a bin applies to just one variable, I think.
So, when we evaluate if datapoints fall into a bin, it is with respect to one variable.
So, if you are evaluating two points and whether or not they match with respect to two
revenue variables, it will look to see if they are in the same bin with respect to var1
AND with respect to the separate bin for var2. It has to satisfy both in order to be a
match.
Does that make sense?
EXL
--
ERIC LIN
Technology and Operations Management
Harvard Business School
Boston, MA 02163
elin at
hbs.edu
On Apr 13, 2010, at 1:07 PM, Doshi, Anil wrote:
I am trying to run a CEM with user-defined coarsening. Essentially there are a ton of
revenue covariates, and I would like them all to be in the same 'bins'. Can
someone confirm that this is the right way to attach a number of cutpoints into the
function? I have a cutpoint for each covariate conjoined by list(). Thanks in advanced.
Anil
salescut.point05 <- seq(0,13,.05) #set up bins
mat.point05 <- cem(treatment = "lrsales", data = winlow, drop =
c("X","lrsales2"),
+ cutpoints = list(lgsales2_lag5 = salescut.point05, lrsales_lag5
+salescut.point05, lgsales2_lag4 = salescut.point05, lrsales_lag4 =
+salescut.point05, lgsales2_lag3 = salescut.point05, lrsales_lag3 =
+salescut.point05, lgsales2_lag2 = salescut.point05, lrsales_lag2 =
+salescut.point05, lgsales2_lag1 = salescut.point05, lrsales_lag1 =
+salescut.point05))
Anil Doshi
Doctoral Student | Technology and Operations Management
Harvard Business School
302 Wyss Hall
Boston, MA 02163
tel 646-244-5396
email adoshi at hbs.edu<mailto:adoshi at hbs.edu>
<ATT00001.c>