Building models with SAS Enterprise Miner, SAS Factory Miner, SAS Visual Data Mining and Machine Learning or just with programming

Why ensemble is decreasing my overall lift?

Reply
Regular Contributor
Posts: 190

Why ensemble is decreasing my overall lift?

I am trying to create ensemble of 3 models: 

1. Regression, cumulative events percentage captured at 20%: 56.8

2. Neural Network, cumulative events percentage captured at 20%: 59.3

3. SVM, cumulative events captured at 20%: 54.01

3. Decision Tree, cumulative events captured at 20%: 48.05

 

Ensemble, cumulative results captured at 20%: 54.18

 

My target is binary. and In ensemble, I have set the posterior probability to maximum. So it will take the maximum probability out of all models, I believe.

Super Contributor
Posts: 337

Re: Why ensemble is decreasing my overall lift?

Posted in reply to munitech4u

Hi Munitech,

Right, maximum will take the maximum posterior probability of your models.

What is the misclassification for your 3 models and their ensemble?

 

Thanks,

M

Regular Contributor
Posts: 190

Re: Why ensemble is decreasing my overall lift?

Posted in reply to M_Maldonado
Regression: Train>0.012276, Test: 0.012101
SVM: Train>0.012515, Test: 0.012374
Decision Tree: Train>0.012245, Test: 0.012069
NN: Train: Train>0.01227, Test: 0.012095
Super Contributor
Posts: 337

Re: Why ensemble is decreasing my overall lift?

Posted in reply to munitech4u

My bad, I don't know how to count... I meant to ask, the Misc of your 4 models and their ensemble.

I am curious if the ensemble of all 4 models is getting worse--and if the Tree or the SVM are tripping it off...

 

Since you are at it, can you include the classification charts as well?

Maybe that will give us a better light of what's going on.

 

Thanks!

Ask a Question
Discussion stats
  • 3 replies
  • 417 views
  • 1 like
  • 2 in conversation