How I can improve my predictive model when i am prediting a RARE EVENT and my sample size is not large?, with EM?

[Mina]

I am modeling the Target “1” which just happens 5% of the time (70 observations for Target "1"). Hence I did the oversampling and adjusting the priorities as follows.

-I add a sample node to the DataSource (the originalpopulation N=1374) in the new diagram (without partition node)

-I add a SCORE Node to the model selected by the bestmodel node

-I add a DECISION node following the modeling node(select model)

At the decision node I set the prior probabilities as:

a) Level “1”, Count (70), Prior (0.5), Adjusted Prior(0.05)

b) Level “1”, Count (70), Prior (0.5), Adjusted Prior(0.95)

c) I applied the decisions by setting “yes” and I runthis node

Then, I run again the score node at the diagram, as the results are below.

The event classification table at the Decision node shows the following results:

FN (70), TN (70), FP (0) TP (0)

How I can imporve my predictive model when i am prediting a RARE EVENT and my sample size is not large?

THANK YOU.

[Doc_Duke]

This thread may have an answer for you.

http://communities.sas.com/thread/14564

Otherwise, repost to the Data Mining Forum, so the people who use EM (I don't currenty) will see the question.

Doc Muhlbaier

Duke