CSSM Procedure
Incorporating New Information
After you create an initial score store, such as mylib.inscore, the scoring process no longer needs the past data (
) that you used to create it. However, as time passes and a data packet, 
, arrives containing new information content, you need to create an updated score store, such as mylib.outscore, that incorporates the new information. Because they use the latest information, of course the forecasts for time points after 
that are based on mylib.outscore are better than the forecasts that are based on the old score store, mylib.inscore. In PROC CSSM, you can create the updated score store by using the data packet that contains new information (
) as the input data table and by using the SCORE statement with both the IN= and OUT= options as follows:
score in=mylib.inscore out=mylib.outscore;
After this, you can discard the old score store, mylib.inscore. The new score store, mylib.outscore, becomes the score store that is based on the full history up to time (
). If you created the initial score store by using OUT(k)= syntax, the updated score store also stores the data rows in 
. You can repeat this score-store updating process again and again as newer information arrives.