Geek asks and answers

Ensemble classifiers (Random Forest, Bagging, Boosting, etc.) in SSAS

I am using SSAS (SQL Server 2008 R2) to develop a classification model for a dataset where 80% of the values ​​are missing. Tree-based ensemble classifiers are supposedly the best solution (e.g. random forest).

Is there a good way to add an ensemble classifier to SSAS? For example, AdaBoost or any other Bagging or Boosting classifier?

I know that SSAS provides plugins, but I have not met anyone who made any ensemble decisions ... Not to mention that you can just download and start using.

If not, is there an efficient way to connect various classifiers in SSAS? I hope that I have lost something obvious that is.

+3
source share
1 answer

I'm not too familiar with the topic you are asking for, but technologically in SSAS you can register assemblies and use them in MDX. So maybe you can encode this in .NET and use the logic in SSAS. Please see the following MSDN page for more information in case it sounds like something worth exploring.

http://technet.microsoft.com/en-us/library/ms175398.aspx

Also, take a look at Data Mining. SSAS provides ready-made materials, as some classification goals can be achieved using the included algorithms:

http://technet.microsoft.com/en-us/library/ms175595.aspx

0
source

More articles:


All Articles