Quick Start

Install the package, fit the reference estimator on the Rust bus data, and read the structural parameters.

This is a mechanical first run. It shows the package shape on a known example; do the data and estimator checks before treating a result as evidence for a new problem.

Install

pip install econirl

Fit NFXP

NFXP is used here because the bundled bus example is the canonical small tabular dynamic discrete-choice problem.

from econirl.datasets import load_rust_bus
from econirl import NFXP

df = load_rust_bus()
model = NFXP(n_states=90, discount=0.9999, utility="linear_cost")
model.fit(df, state="mileage_bin", action="replaced", id="bus_id")

print(model.params_)

{'theta_c': 0.0010029257006533541, 'RC': 3.072263842893654}

Read the results

The fitted model exposes the reward parameters, their standard errors, the value function, and the implied policy.

print(model.summary())
print(model.se_)

Run a counterfactual

Change a structural primitive and read the new policy. This only makes sense for estimators that recover a reward object the model can re-solve under the changed primitive.

cf = model.counterfactual(RC=4.0)
print(cf.policy[50, 1])

0.055196291500871957

Next, check the data and the design before trusting a fit. See Loading Data and Pre-Estimation.