Skip to content

Latest commit

 

History

History
5 lines (3 loc) · 642 Bytes

README.md

File metadata and controls

5 lines (3 loc) · 642 Bytes

Learn_CMDL

Learn_CMDL is a Java implementation of a new proposed scoring function for learning Bayesian networks called Complete Minimum Description Length (CMDL). CMDL can be efficiently computed through a graph quotient for a class of Bayesian networks that we denote by Covering Graphs. A score-based learning algorithm is implemented, considering the Greedy Hill Climber (GHC) as the search method. The program receives as input a data set with multivariate categorical observations and outputs the optimal structure.

For further details, please refer to the project website: https://margaridanarsousa.github.io/Learn_CMDL/.