CODEWITHSUNDEEP
pythonprobabilitybayesian-networkspomegranate
Shiny_and_Chrome
Shiny_and_Chrome

Reputation: 320

Representing disease and syptom in pomegranate bayes network

I'm trying to work out P(m|s) probability of meningitis given "stiff neck"

So I'm trying to represent this in the model: P(m|s) = (P(s|m) * P(m))/P(s)

P(s) = 0.1

P(m) = 0.0001

P(s|m) = 0.8

this is how I've represented it:

meningitis = DiscreteDistribution({"have meningitis": 0.0001, "no meningitis": 0.9999})

stiffNeckIfMeningitis = ConditionalProbabilityTable(
        [["have meningitis", "stiff", 0.8],
         ["have meningitis", "not stiff", 0.2]
        ],[meningitis])

I have to specify the values for ["no meningitis", "stiff", x] and ["no meningitis", "not stiff", x] which I did by manually working out the values, but should the bayes network not be doing this?

Should I have have a Discrete distribution like this too:

neck = DiscreteDistribution({ "stiff": 0.1, "not_sitff": 0.9})

I'm completely stumped on how to represent the relationships in the model = BayesNetwork() bit.

At the moment I have:

s1 = Node(meningitis, name="meningitis")
s2 = Node(stiffNeckIfMeningits, name="stiffNeckIfMeningits")
model.add_states(meningitis, stiffNeckIfMeningits)
model.add_edges(s1, s2)
model.bake()

Upvotes: 3

Views: 59

Answers (0)

Related Questions