SQL Server Analysis Services & The Number of Input Attributes
-
Monday, February 11, 2013 9:18 PM
How many input attributes could be used for training?
Usually for training I use a table in which each column is an attribute.
I read that there is a limitation in SQL to have up to 1,024 columns in one table.
Is there a better way for training using Analysis Services (for example, to build the mining model based on some EAV table)?
I'm interested to use much more than 1,024 input attributes in my models...
Tnx.
- Edited by CoffeeCake Monday, February 11, 2013 9:19 PM
- Moved by furmanggMVP Monday, February 11, 2013 10:43 PM
All Replies
-
Tuesday, February 12, 2013 7:13 AMModerator
Hi CoffeeCake,
In Standard Edition, it is limit to 5000 attributes you can have in a mining model, please refer to the similar threads:
http://social.msdn.microsoft.com/Forums/lv/sqldatamining/thread/09444f74-65f4-48c8-8bfb-c11f4ec5bd2b
http://social.msdn.microsoft.com/Forums/en-US/sqldatamining/thread/3f822aac-b0c8-48f0-9ffa-8f76187edf9bThanks,
Eileen
If you have any feedback on our support, please click hereEileen Zhao
TechNet Community Support
- Marked As Answer by CoffeeCake Thursday, February 14, 2013 11:22 PM
-
Wednesday, February 13, 2013 7:35 PM
Thank you Eileen.
Analysis Services is a great tool, but I may need to work with databases that contain hundreds of thousands of attributes, such as databases that contain genetics-related information.
See, for example, the following paper: The application of naive Bayes model averaging to predict Alzheimer’s disease from genome-wide data.
Any ideas for appropriate tools?
How about R and its Machine Learning toolboxes?
Tnx.
- Edited by CoffeeCake Thursday, February 14, 2013 11:19 PM
- Marked As Answer by CoffeeCake Thursday, February 14, 2013 11:22 PM
-
Thursday, February 14, 2013 11:21 PM
By the way - is there a reason why the limitation is not presented at the Maximum Capacity Specifications?
- Marked As Answer by CoffeeCake Thursday, February 14, 2013 11:22 PM
.png)
