Home » Machine Learning/Artificial Intelligence

Applications of PGMs and relation and conversion of Bayes and Markov's model

In this article, we will know more about PGMs: its application, Relation between Bayes network and Markov's Random field and conversion of Bayes network into Markov's random field.
Submitted by Bharti Parmar, on March 16, 2019

Application of PGMs: (Use cases)

• Netflix, Amazon, facebook all use PGM recommend what is best for you.
• This algorithm, Use the latent factor model and non-negative matrix factorization.
• In politics American presidential polls using PGMs for prediction. FiveThirtyEight company that makes a prediction about American presidential polls using PGMs.
• PGM is also inferred whether one is a MODI and KEJRIWAL supporter.

How Bayes network and Markov Random Fields Model related to each other:

	For P(A, B) = P(A) * P(B/A)	For P(A,B) α Ф(A,B)
	For chain: P(A,B) = P(A) P(B/A) P(C/B)	For chain: P(A,B) α Ф(A,B) Ф(B,C)
	Shared parents: P(A,B,C) = P(A) P(B/A) P(C/B)	Shared parents: P(A,B,C) α Ф(A,B) Ф(A,C) Ф(A,B) ← P(A) P(B\A) Ф(A,C) ← P(C\A)
	Two parent shared a child: P(A,B,C) = P(A) P(B) P(C\A,B)	A and B are independent given C P(A,B,C) α Ф(A,C) Ф(B,C)

Converting Bayes network into Markov random fields:

Moralizing parents: P (A, B, C) α Ф(A,C) Ф(B,C) where A and B are independent given C

• Moralize all co-parents
• More challenging is that Lose marginal independence of parents.
  

Basically, it given an interference class of all directed graphical models and a class of all undirected graphical model are different but overlap each other.

Conclusion: In this article we have learnt about application of PGMs, how bayes network is related to markov's random field and conversion of bayes network into Markov's network model.