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

SASEminer-Data-Volume-Issue-while-Modelling

Accepted Solution Solved
Reply
Highlighted
New Contributor
Posts: 2
Accepted Solution

SASEminer-Data-Volume-Issue-while-Modelling

I am using SAS Eminer-  HP BN Classifier Node .  I have 11 crore records . I am using  all data as training  dataset ,Since It is a huge data it takes 10-12 hours to run the Model . After first 5-6 hours of execution I am getting below error . Kindly let me know  if it is a Limitation of  SAS Eminer  ? Please suggest any alternative handle such enormous data  .

 

ERROR : The connection to the server has been lost . As a result , this diagram will be closed  .

 

Thanks

Sandeep


Accepted Solutions
Solution
Saturday
SAS Employee
Posts: 14

Re: SASEminer-Data-Volume-Issue-while-Modelling

Posted in reply to Sandeepgup11

There are a few different approaches to use in a case like this.  

 

* The literal cause is that your job run time exceeds the time limit that is imposed by your IT department.  Contact your IT department about giving you a longer time when you expect to be running longer jobs.

 

* If you have a client-server installation, then start your diagram, close it, and you can also close Enterprise Miner.  Processing continues even though the user interface is not open.  Since it is not open, there is no connection to lose.  Periodically (after lunch, later in the day, the next day if running over night) try to open the diagram.  It should open if the job is completed, but should be locked if it is still running.

 

 * Check your MEMSIZE setting.  If you are using a setting that is too small, processing time might take longer.  Likewise, increasing the MEMSIZE setting can sometimes enable a job to complete more quickly.  For details, go to http://support.sas.com/notes.  In the search area, try these terms:

 

     miner memsize

 

Select this result.
61376 - Overcoming "insufficient memory ." and "parameter larger than documented limit" error messag...

http://support.sas.com/kb/61/376.html

 

* In that same note, see the "Check whether you are trying to use a small-data-algorithm on a large data-set" section.

 

 

It is likely that your solution involves a combination of the above issues.

 

View solution in original post


All Replies
Solution
Saturday
SAS Employee
Posts: 14

Re: SASEminer-Data-Volume-Issue-while-Modelling

Posted in reply to Sandeepgup11

There are a few different approaches to use in a case like this.  

 

* The literal cause is that your job run time exceeds the time limit that is imposed by your IT department.  Contact your IT department about giving you a longer time when you expect to be running longer jobs.

 

* If you have a client-server installation, then start your diagram, close it, and you can also close Enterprise Miner.  Processing continues even though the user interface is not open.  Since it is not open, there is no connection to lose.  Periodically (after lunch, later in the day, the next day if running over night) try to open the diagram.  It should open if the job is completed, but should be locked if it is still running.

 

 * Check your MEMSIZE setting.  If you are using a setting that is too small, processing time might take longer.  Likewise, increasing the MEMSIZE setting can sometimes enable a job to complete more quickly.  For details, go to http://support.sas.com/notes.  In the search area, try these terms:

 

     miner memsize

 

Select this result.
61376 - Overcoming "insufficient memory ." and "parameter larger than documented limit" error messag...

http://support.sas.com/kb/61/376.html

 

* In that same note, see the "Check whether you are trying to use a small-data-algorithm on a large data-set" section.

 

 

It is likely that your solution involves a combination of the above issues.

 

New Contributor
Posts: 2

Re: SASEminer-Data-Volume-Issue-while-Modelling

Posted in reply to Sandeepgup11
Thanks Mike.
☑ This topic is solved.

Need further help from the community? Please ask a new question.

Discussion stats
  • 2 replies
  • 160 views
  • 1 like
  • 2 in conversation