Memory Limitation When Using Liner Regression

[MAAbdullahAlMubarah]

Dear All ,

This Post related to the following Kaggle competition: https://www.kaggle.com/c/competitive-data-science-predict-future-sales

I have prepared the data Until I reach it for prediction phase (check the data-set on this link): https://drive.google.com/open?id=1F_sNYLXrEzS6tAjhWkYCm87UdQcNo1ks

When I try to use the Liner regression (item_cnt_month is the target variable) I got the following error message : 
 Error: cannot allocate vector of size 285.1 GbAlteryx

Of course, I know it is related to the memory limitation as a result of the high number of Values of the dummy variables,

My Question is what is your suggestions to overcome this memory limitation?

[danilang · Accepted solution]

Hi @MAAbdullahAlMubarah

You might want to check out this thread, especially the posts on clustering.  You may find some ideas that you can apply

Dan

[MAAbdullahAlMubarah]

Thank you for your replay @danilang , but this related to clustering and my case is related to linear regression,Is it with the same concept?