Making small parts of large models

Top  Previous  Next

One will not always be able to do everything perfectly within the small model.  One needs to judiciously develop the small data set so it has all the features of the large data set.  I have found that most of the work can be done in the simpler setting.

Occasionally something happens in the full data set that I cannot reproduce in the small data set.  There are almost always be peculiarities and interrelationships introduced when I go to the full data set.  When I need to find a large data set only problem, I try one of the following strategies.

 

Save and restart to isolate problem area

Strategic sub-setting

Data reduction

 

Each of these is explained below.

 

Save and restart to isolate problem areas

Strategic sub-setting

Data reduction