Python based AI Work Causal Bayesian Network
Budget: ₹600 – ₹1,500 INR
You are given a dataset of patients’ survival cases Y in response to the voluntary intake of a treatment X (i.e. placebo or medicine). Unfortunately, the success of the treatment depends also on their mental health Z, which affects both their chances of survival and willingness to take the medicine.
Assuming X ∈ {nothing, placebo, medicine} is the possible treatment, Y ∈ {survive, ¬survive} is the outcome, and Z ∈ {healthy, ¬healthy} is the mental condition:
a) Represent the above situation with a Causal Bayesian Network (CBN) and its conditional probability tables (CPTs).
b) Implement and document a Python program that reads the given dataset and automatically computes the intervention’s probability P(Y | do(X)) for all the possible values of X and Y.
c) Modify the CBN to introduce another variable W ∈ {gene, ¬gene}, which indicates whether the patient is genetically predisposed to some mental problem. This influences the actual mental condition Z according to the following probabilities:
P(Z = healthy | W = gene) = 0.6
P(Z = healthy | W = ¬gene) = 0.8
P(W = gene) = 0.1
Implement and document a Python program for the new CBN that computes the average causal effect (ACE) of the genetic profile W on the patient’s survival Y.
Assuming X ∈ {nothing, placebo, medicine} is the possible treatment, Y ∈ {survive, ¬survive} is the outcome, and Z ∈ {healthy, ¬healthy} is the mental condition:
a) Represent the above situation with a Causal Bayesian Network (CBN) and its conditional probability tables (CPTs).
b) Implement and document a Python program that reads the given dataset and automatically computes the intervention’s probability P(Y | do(X)) for all the possible values of X and Y.
c) Modify the CBN to introduce another variable W ∈ {gene, ¬gene}, which indicates whether the patient is genetically predisposed to some mental problem. This influences the actual mental condition Z according to the following probabilities:
P(Z = healthy | W = gene) = 0.6
P(Z = healthy | W = ¬gene) = 0.8
P(W = gene) = 0.1
Implement and document a Python program for the new CBN that computes the average causal effect (ACE) of the genetic profile W on the patient’s survival Y.