I make some algorithms to estimate the optimal values.
and i want to model it algorithms using probabilistic theorem.
but i dont know which algorithm or model is proper expression.
at first, i thinks it is a kind of bayesian network.
When a given graphical model is A -> B.
In typical bayesian networks,
P(A,B) = P(B|A)P(A)
And before estimates probability, we calcualte P(B|A) and P(A).
But in my case,
I have to calculate P(A) first
and when i estimate P(B|A), I use A' = arg max_A { P(A) } ex. P(B|A) can be N(B, A') ; A' is mean.
It feels something iterative, update or sequantial.
This property makes me interpret this algorithm hard.
I find below theorem or algorhtms.
bayesian network, markov chain, kalman filter, reculsive bayesian estimator. ...
In common sense, when estimate next probability, using a value that maximizes prior probability makes sense.
But how i can interprets it.
Please help me.
and i want to model it algorithms using probabilistic theorem.
but i dont know which algorithm or model is proper expression.
at first, i thinks it is a kind of bayesian network.
When a given graphical model is A -> B.
In typical bayesian networks,
P(A,B) = P(B|A)P(A)
And before estimates probability, we calcualte P(B|A) and P(A).
But in my case,
I have to calculate P(A) first
and when i estimate P(B|A), I use A' = arg max_A { P(A) } ex. P(B|A) can be N(B, A') ; A' is mean.
It feels something iterative, update or sequantial.
This property makes me interpret this algorithm hard.
I find below theorem or algorhtms.
bayesian network, markov chain, kalman filter, reculsive bayesian estimator. ...
In common sense, when estimate next probability, using a value that maximizes prior probability makes sense.
But how i can interprets it.
Please help me.