abuza Tracker
https://communities.sas.com/kntur85557/tracker
abuza TrackerMon, 20 May 2024 15:32:45 GMT2024-05-20T15:32:45ZRe: Running Proc Logistic on a sparse matrix
https://communities.sas.com/t5/SAS-IML-Software-and-Matrix/Running-Proc-Logistic-on-a-sparse-matrix/m-p/521426#M4496
Thank you Rick, I will give it a try.<BR /><BR />Kind Regards,<BR />AlexisFri, 14 Dec 2018 11:54:34 GMThttps://communities.sas.com/t5/SAS-IML-Software-and-Matrix/Running-Proc-Logistic-on-a-sparse-matrix/m-p/521426#M4496abuza2018-12-14T11:54:34ZRunning Proc Logistic on a sparse matrix
https://communities.sas.com/t5/SAS-IML-Software-and-Matrix/Running-Proc-Logistic-on-a-sparse-matrix/m-p/521419#M4494
<P>Hi,</P><P> </P><P>I've been working with a big and sparse matrix (over 5,000 columns -dummy variables- and 2M observations). For every observation, only 20-30 columns of the <SPAN>5,000 contain a value. All of the 5,000 columns are independent variables, except for one, which is the dependent I need to explain. Running proc </SPAN>Logistic on that matrix, besides that technically results in an 'out of memory', is not a solutions as I would have to run it on a regularly basis.</P><P> </P><P><SPAN>Using IML, with <EM>sparse</EM>, <EM>dense</EM> and <EM>full</EM> functions I can get rid of the zeroes and store only the significant values. Hence, my question arises as to</SPAN><SPAN> how can I run proc logistic / glimmix on that resulting matrix, as it is totally distorted from the original one. And further, how could I be able to extract insights from the results of these procedures relating to the original matrix an its variables.</SPAN></P><P> </P><P><SPAN>Thank you for your time.</SPAN></P><P><SPAN>Kind Regards,</SPAN></P><P><SPAN>Alexis.</SPAN></P><P> </P><P> </P>Fri, 14 Dec 2018 11:09:45 GMThttps://communities.sas.com/t5/SAS-IML-Software-and-Matrix/Running-Proc-Logistic-on-a-sparse-matrix/m-p/521419#M4494abuza2018-12-14T11:09:45Z