Ensemble learning
![[Deleted User]](https://us.v-cdn.net/6038102/uploads/defaultavatar/nLP0QHCLH24WL.jpg)
[Deleted User]
New Altair Community Member
Hi
When we use " Ensemble learning" and " Group model operator " why the algorithm which has high accuracy alone effect the same on group model? ( when we combine some algorithms with high accuracy the result doesnt improve more!!!!!) so in this situation " Ensemble learning " doesnt have any meaning.
mbs
When we use " Ensemble learning" and " Group model operator " why the algorithm which has high accuracy alone effect the same on group model? ( when we combine some algorithms with high accuracy the result doesnt improve more!!!!!) so in this situation " Ensemble learning " doesnt have any meaning.
mbs
Tagged:
0
Best Answers
-
Hi
Finally I find the answer from this link. The answer is:
"Keep in mind just by adding layers and more models to your stacking algorithm, does not mean you’ll get a better predictor".
https://medium.com/@rrfd/boosting-bagging-and-stacking-ensemble-methods-with-sklearn-and-mlens-a455c0c982de
mbs0 -
Well, the more important part is that Group Models is actually not building an ensemble at all. Group Model is simply collecting a set of models (predictive models and preprocessing models) and the resulting model is just applying them in the same order. If you want to build ensembles, you need to use operators like Vote etc.Hope this helps,
Ingo1 -
The Vote operator is a very simple ensemble---it takes multiple independent ML algorithms, builds a separate model for each, and then makes the final prediction based on voting from the predictions of each of the individual models.
There is a nice tutorial process you can see with that operator to understand how it is set up and what the output looks like.
1 -
Just put the Vote operator in your process and then click on the link for the tutorial process in the help.1
-
To clarify further, the Help menu is context driven, it will reflect the currently selected operator. The screenshot you have shared indicates you have no operator highlighted (shown in orange outline instead of black) and so it is just the default help for the entire process.1
Answers
-
Hi
Finally I find the answer from this link. The answer is:
"Keep in mind just by adding layers and more models to your stacking algorithm, does not mean you’ll get a better predictor".
https://medium.com/@rrfd/boosting-bagging-and-stacking-ensemble-methods-with-sklearn-and-mlens-a455c0c982de
mbs0 -
Well, the more important part is that Group Models is actually not building an ensemble at all. Group Model is simply collecting a set of models (predictive models and preprocessing models) and the resulting model is just applying them in the same order. If you want to build ensembles, you need to use operators like Vote etc.Hope this helps,
Ingo1 -
@IngoRM
hi
Thank you very much for useful information. Could you please tell me more about Vote operator and ensemble learning process with RM?
Thank you in advance for your help
mbs1 -
The Vote operator is a very simple ensemble---it takes multiple independent ML algorithms, builds a separate model for each, and then makes the final prediction based on voting from the predictions of each of the individual models.
There is a nice tutorial process you can see with that operator to understand how it is set up and what the output looks like.
1 -
1
-
Just put the Vote operator in your process and then click on the link for the tutorial process in the help.1
-
0
-
Please see below screenshot. The help window is at the bottom right corner (if not enable it from View --> SHow Panen --> Help). Click on the jump to tutorial and then click on the tutorial, it will open a new process.
0 -
0
-
You should click on the Vote operator once, then you will get it2
-
To clarify further, the Help menu is context driven, it will reflect the currently selected operator. The screenshot you have shared indicates you have no operator highlighted (shown in orange outline instead of black) and so it is just the default help for the entire process.1
