Hello,
As a complete beginner I am wondering how to interpret the results of the strength of influence.
Almost all my arrows are purple meaning the influence is ambiguous.
I dropped a screenshot of the values as an attachment.
Interpreting Strength of influence
Interpreting Strength of influence
- Attachments
-
- datastrength.PNG (63.97 KiB) Viewed 6156 times
-
marek [BayesFusion]
- Site Admin
- Posts: 438
- Joined: Tue Dec 11, 2007 4:24 pm
Re: Interpreting Strength of influence
Hi Elise,
The color of the arc will be displayed once you specify for the nodes in your model whether they are increasing (Low to high) or decreasing (High to low) order of outcomes. I assume you have done that. The color green means that the higher the outcomes of the parent, the higher the outcomes of the child. Red means the opposite, i.e., the higher the outcomes of the parent, the lower the outcomes of the child. Purple means that the relationship is not that clear, i.e., sometimes the higher state of the parent increases the state of the child and sometimes it decreases it. The above is all regardless of the actual measure used for the strength of the arcs. So, the numbers in your table do not tell me much.
I think that the section "Strength of influence" in GeNIe manual outlines it quite comprehensibly. If you want to learn more about the measures themselves, please look at the original source, notably Joost Koiter's thesis (listed in the manual):
Koiter, Joost R. (2006). Visualizing Inference in Bayesian Networks. M.Sc. thesis, Faculty of Electrical Engineering, Mathematics, and Computer Science, Department of Man-Machine Interaction, Delft University of Technology.
You should be able to find it online.
I hope this helps,
Marek
The color of the arc will be displayed once you specify for the nodes in your model whether they are increasing (Low to high) or decreasing (High to low) order of outcomes. I assume you have done that. The color green means that the higher the outcomes of the parent, the higher the outcomes of the child. Red means the opposite, i.e., the higher the outcomes of the parent, the lower the outcomes of the child. Purple means that the relationship is not that clear, i.e., sometimes the higher state of the parent increases the state of the child and sometimes it decreases it. The above is all regardless of the actual measure used for the strength of the arcs. So, the numbers in your table do not tell me much.
I think that the section "Strength of influence" in GeNIe manual outlines it quite comprehensibly. If you want to learn more about the measures themselves, please look at the original source, notably Joost Koiter's thesis (listed in the manual):
Koiter, Joost R. (2006). Visualizing Inference in Bayesian Networks. M.Sc. thesis, Faculty of Electrical Engineering, Mathematics, and Computer Science, Department of Man-Machine Interaction, Delft University of Technology.
You should be able to find it online.
I hope this helps,
Marek