Data input and export
-
- Posts: 8
- Joined: Fri Feb 28, 2025 2:58 am
Data input and export
Hi, I have created a Bayesian network with many nodes, and I need to double-click on each node to enter the probability, is there a function that can be solved with a single window. There is also a function that can solve the problem with one window. The calculated posterior probabilities need to be opened one by one to copy the data for each node, which is very troublesome for models with many nodes, is there a button (function) that can be clicked to view all the posterior probabilities, which are all displayed in one pane, and it is easy for the whole to copy the data to other software such as excel.
- Attachments
-
- tmpF176.png (54.76 KiB) Viewed 25 times
-
- Site Admin
- Posts: 1474
- Joined: Mon Nov 26, 2007 5:51 pm
Re: Data input and export
GeNIe does not have the command to show all posterior probabilities. There's a Bulk Update option which iterates over rows in the datafile and outputs probabilities; the output file can be easily loaded into Excel. You can also switch node view to barchart, which gives you instant visualization of the probabilities in the main graphical view (but no copy to Excel).
The functionality you described (a command to show/copy all probabilitites) would be quite easy to implement. The question is, what data format should be used - one row for each node with node id in the first column, and probabilities in the 2nd, 3rd, ... columns? Please note that in general case nodes can have different number of outcomes. If you can create an Excel file with the layout, it would help us design this feature.
The functionality you described (a command to show/copy all probabilitites) would be quite easy to implement. The question is, what data format should be used - one row for each node with node id in the first column, and probabilities in the 2nd, 3rd, ... columns? Please note that in general case nodes can have different number of outcomes. If you can create an Excel file with the layout, it would help us design this feature.