Skip to content

Latest commit

 

History

History
27 lines (24 loc) · 1.25 KB

README.md

File metadata and controls

27 lines (24 loc) · 1.25 KB

AI_Algortihms

In this repository you will find implementation of two algoritms in AI field in Java. Please make sure that you run this repo on a good input and stand for the requirments.

Bayes Ball's algorithm

The algorithm checks whether two nodes are independent or not.

In case don't have evidence variables it's to easy to check if the nodes are independent or not.

Steps:

  • check if there is a path between two nodes.
  • path can be represented by common cause connection or causal chain
  • if there exits path between those two nodes then they are no possible independent.

in case we have evidence variables

Steps:

  1. Start with initial factors
  2. local CPTs instantiated by evidence.
  3. If an instantiated CPT becomes one-valued, discard the factor.
  • While there are still hidden variables (not Q or evidence):
    • Pick a hidden variable H
    • Join all factors mentioning H
    • Eliminate (sum out) H
    • If the factor becomes one-valued, discard the factor
  1. Join all remaining factors and normalize

E.g: