Some Informal Notes on Functionality in GeNIe

Graphical Network Interface (GeNIe) is a software package that can be used to create decision theoretic models intuitively using the graphical click-and-drop interface (currently Windows version only).

Some Informal Notes on Functionality in GeNIe

Postby Kaumudi » Thu Mar 21, 2019 12:20 am

Alright listed below are a few random things I have learnt while playing around with GeNIe functionality. Forum members feel free to add to this list or correct these if necessary.

1. In Learning Algorithm, if you use Bayesian Search, pay attention to the Algorithm Parameters:
I have tended to keep Max Parent Count at 8 and Iterations at 20, but the default sample size is 50, I change it to 100 because the way I have understood it, "Sample Size" stands for the confidence you have in the input data - 100 is max.
2. Seed is explained as follows in the GeNIe manual: "With all functionality that needs randomness, GeNIe relies on the JKISS random number generator with a period of 2127 = 1.7 x 1038. This is sufficient for all practical purposes. Non-zero seeds are hashed before they are used to initialize the generator. Zero seed causes the generator to be initialized with the value based on the system clock and, therefore, makes the output of the algorithm using the pseudo-random generator differ across runs."
3. Link probability lets you decide the amount of confidence you have in the fact that the nodes in your structure should be connected.
From manual:
" Link Probability (default 0.1) is a parameter used when generating a random starting network at the outset of each of the iterations. It essentially influences the connectivity of the starting network.
· Prior Link Probability (default 0.001) influences the (BDeu) score, by offering a prior over all edges. It comes into the formula in the following way: log Posterior score = log marginal likelihood (i.e., the BDeu) + |parents|*log(pll) + (|nodes|-| parents|-1)*log(1-pll))."

Ok - will talk about K-fold cross-validation in my next post
Kaumudi
 
Posts: 76
Joined: Tue Feb 14, 2017 12:38 pm

Return to GeNIe

Who is online

Users browsing this forum: No registered users and 2 guests