Categorical data represents variables whose values are labels from finite categories rather than numerical measurements. In a causal model, choosing these abstractions carefully matters because analysis depends on how variables are defined and encoded. section 3.1.8
Categorical variables support a discrete modeling workflow:
- Represent each variable through a finite set of named outcomes. section 3.1.8
- Fit conditional probability tables as causal Markov kernels. section 3.1.8
- Read a probability for each outcome after fixing parent-variable values; probabilities for those outcomes sum to 1. section 3.1.8
Transportation choices
A transportation survey encodes variables such as occupation, residence, and transportation choice categorically. For example, given occupation and residence values, a conditional probability table returns probabilities for car, train, or other transportation. section 3.1.8
Listing 3.4 Learning parameters for the causal Markov kernels
from pgmpy.models import BayesianNetwork
model = BayesianNetwork(
[
('A', 'E'),
('S', 'E'),
('E', 'O'),
('E', 'R'),
('O', 'T'),
('R', 'T')
]
)
model.fit(data) #1
causal_markov_kernels = model.get_cpds() #2
print(causal_markov_kernels) #2
Inspect a Python workflow that fits categorical data in a Bayesian network and retrieves its learned conditional probability tables. section 3.1.8
