Undestanding Bayesian network with OpenMarkov

I downloaded OpenMarkov software for probabilistic graphical models and tried it on mtcars dataset.

The mtcars.csv data looks like this:

In OpenMarkov GUI, I went to Tools > Learning and loaded mtcars.csv dataset. I then adjusted preprocessing settings to have Discretize with Equal width intervals for all variables.

I then chose Hill Climbing algorithm (default) and Automatic learning options. On learning, the result was as follows:

My question is what exactly does this figure represent? Does it represent a Bayesian network or some other type of probabilistic graphical models? Also, do arrows mean that hp affects cyl and carb; and cyl in turn affects disp and carb and so on?

Topic bayesian-networks markov

Category Data Science

1
answered at

First off, I did not know OpenMarkov. Anyway, from its website it has a particular focus on learning Bayesian networks (Bayes nets). Thus, I assume your figure represents a Bayes net, yes. Syntactically, it also qualifies since it is a directed acyclic graph.

The arrows (edges) represent influences (conditional dependencies) observed in the data. For instance, the conditional probability distribution of carb, P(carb | hp,cyl,disp), is defined by the values for hp,cyl,disp. However, arrows do not necessarily represent causal relationships.

About

Geeks Mental is a community that publishes articles and tutorials about Web, Android, Data Science, new techniques and Linux security.