Complimentary Coursera offer for all new members
Get offer Join without the offer
For predicting time series with machine learning algorithms it is usual that you "lag" your columns many times (For Example 12 -24 times for monthly predictions). For example if you have initially 150 columns, you get after lagging them 24 times 24 * 150 --> 3600 columns.
Here comes my problem: SPSS Modeler is terribly slow. Even if you change something like a filter node, you add a source, you make any interaction like unlink nodes, deactivate node... and so on, you have to wait up to 1 minute for any change. All of that just because of many Columns.
What can I do here? Is there any setting i have to adjust? I already did dimension reduction with feature generation, filtering out columns with poor correlations...