Regularized Maximum Likelihood Estimation for the Random Coefficients Model in Python

We present PyRMLE (Python regularized maximum likelihood estimation), a Python module that implements regularized maximum likelihood estimation for the analysis of Random coefficient models. PyRMLE is simple to use and readily works with data formats that are typical to Random coefficient problems. The module makes use of Python’s scientific libraries NumPy and SciPy for computational efficiency. The main implementation of the algorithm is executed purely in Python code, which takes advantage of Python’s high-level features. The module has been applied successfully in numerical experiments and real data applications. We demonstrate an application of the package in consumer demand.