Hi,
I am looking for ways to build a look-alike model in SAS (either Base or Enterprise Miner).
The goal is to rank order or select a group of existing customers that have a high propensity to respond to upselling campaigns, base on the results of previous campaigns (or organic responders).
The challenge is that the previous campaigns only have reached out to a very selective segment of the whole customer base making it difficult to use the campaign results to train models and apply the results to the whole population. Reject inference also does not seem to be a solution. Not sure if MBR or Nearest neighborhood methods would help.
Any input would highly be appreciated.
Thanks,
Sam
