Step 1.3 Make the relevant nodes import/export nodes
We want some of the nodes of this template to be ‘joined’ to other templates. Specifically, the nodes “discovered defects” and “inserted defects” will be joined. We therefore need to make each of these nodes “input nodes”. To do this, open up the Node Properties window for each node in turn and
select “input node” as shown:
Once you press OK the node will change colour to denote that it is now
recognised as an input node.
Finally, make sure you save both the template file and name it with the same
name as shown:
At this point you can ‘test out’ the template to see how it runs, by clicking the lightning icon in the
toolbar:
There will be a delay as the template compiles (a lot of processing goes on to
compute all the values of a very large NPT). Then you should see the following
display in a new window:
Double-clicking on, say the “inserted defects” node will unveil a window showing the probability distribution for this node
(the diagram below only shows part of the distribution as it has too many
states to fit into one screen):
Open up the probability windows for each of the nodes. You are now ready to
enter evidence and propagate it.
Click the value “19” for inserted defects, “11” for discovered defects and then click the button:
This will propagate the values and you will see the computed value of residual
defects appear.
You can ‘retract’ evidence by simply clicking on an already selected value (the tick will be
cleared). You can enter new evidence and propagate again. You can tick as many
values as you like in any probability window. So, if for example, if you believe
the inserted defects is a number between 5 and 10 but don’t know which you can select all of those values to reflect your uncertainty.
When you exit the compiled template window you will be presented with the
message:
This enables you to save the evidence you entered as a case file if you wish.