Dynamic Bayesian Network Action Set

This example shows how you can use the dynBnet action along with a previously saved model to detect faults in new observations. The following data table contains the new observations from time periods 11 to 20:

data casuser.tank_data_2;
input Time f1out_rd f2out_rd;
datalines;
11        9.565        2.629
12        9.523        2.685
13        9.482        2.739
14        9.440        2.792
15        9.399        2.844
16        9.358        2.895
17        9.318        2.944
18        9.277        2.993
19        9.237        3.040
20        9.198        3.086
;

You need to invoke the ASTORE procedure as follows to upload the saved state from a file on disk so that it can be accessed as input by the dynBnet action:

 proc astore;
     upload rstore=casuser.tank_state_in store='\Documents\tank_state';
 run;

Finally, you invoke the dynBnet action as follows:

proc cas;
   loadactionset "dynamicBayesianNetwork";
   action dynBnet result = res /
      table = {data = {name = "tank_data_2" caslib = "casuser"}}
      inputState = {data = {name = "tank_state_in" caslib = "casuser"}}
      output = {data = {name = "tank_out_2" caslib="casuser" replace = 1}}
   ;
run;