training set with positive label only

miky

Hello,

I need to use only positive labels in training set, how to do it? (like in SPSS)
I mean I can sign wich records are GOOD, but don't have opposites.
We are using Logistic Regression Model... (after Principal Component Analysis)

We have customers and finding new potencional customers (we don't have non-customers).

Example (fictional)

ID	label	pc_1	pc_2	pc_3
1	1	0.23	-0.65	0.89	this is my GOOD label
2	0	-0.15	0.32	0.62	I want to ommit this rows (we are faking them)
3	0	-0.37	0.87	0.24	I want to ommit this rows (we are faking them)

Thank you very much

MariusHelf

So you are creating fake data as negative examples? Do you want to use that data as training data? Then you can just learn the Logistic Regression as you would on any other dataset.
Please let us know about any problems you have.

Best regards,
Marius