bayes network representation Flashcards
(21 cards)
what are bayesian networks
probabilistic graphical models that use edges to display cause-effect relationships and bayes theoreon for probabilistic inference
what do edges represent in bayes network graph
displays cause and effect relationship
what is bayes theorem used for bayes network graph
probabilistic inference
what is bayes thereom equation
P(A|B) = ( P(B|A)*P(A) ) / P(B)
real world applications of bayes network
legal tech
chemistry
cyber security
advantages of bayes network
▶ Graphical representation: Provides a visual representation of joint
probability distributions of different random variables –
interpretability.
▶ Powerful: Can capture complex relationships between random
variables.
▶ Combine data and prior knowledge: Incorporates prior knowledge and
updates with statistically significant information from data.
▶ Generative approach: Able to generates new data similar to existing
data
disadvantages of bayes network
Requires prior knowledge of many probabilities.
Sometimes computationally intractable
what are the three main task needed for bayes network
1) inference: from observations ->
2) training: learn the model parameter ( usually calculating the probabilities )
3) structure determination: identify what is connected to what
how can we represent the joint probability distributions of random variables
bayes network is a directed acrylic graph (DAG)
what does DAG stand for
directed acrylic graph
what dies the bayes network graph consists of
set of nodes -> represent random variable
set of DIRECTED edges -> representing cause and effect relationship and connects those nodes
what does a directed edge represent
“directed dependency” or
“directed influence”
also called “direct cause”
basically a parent node ponitng to a child node
(PN) –> (CN)
what is the conditional distribution for each node given its parents
P( Xi | Parents (Xi) )
discrete random variables for bayes network
conditional distributions are represented as a conditional probability table (CPT) - the distribution over Xi for each combination of parent values
describe essence of a bayesian network
A compact representation of a joint probability distribution in terms
of conditional distributions
what is the equation of bayes network
P(X1, X2, . . . , Xn) =
n
i =1
P(Xi |Parents(Xi ))
what are the 4 probabilstic relationships of BN (the standard structures)
direct cause
indirect cause
common cause
common effect
draw a direct cause relationship
A -> B
draw a indirect cause relationship
A –> B –> C
draw a common cause relationship
B <– A –> C
draw a common effect relationship
A –> C <– B
