Add Poisson regression / GLM benchmark

This is a great project. Currently, it focuses on classification and squared error problems. It would be very nice to have one benchmark for Poisson GLMs, see e.g. https://github.com/scikit-learn/scikit-learn/pull/26707 with the French Motor Third-Party Liability Claims dataset https://www.openml.org/d/41214.

Start by reviewing the existing classification and squared-error benchmarks in this repository, then read scikit-learn PR 26707 and inspect the French Motor Third-Party Liability Claims dataset on OpenML (41214). Done means the project includes one Poisson GLM benchmark based on that dataset.

Written by the indexing model from the issue text.