Imodels

imodels is an open source Python package for concise, transparent and accurate predictive modeling. It gives data scientists a scikit learn compatible interface to interpretable models, so a classifier or regressor can be imported, fitted and used with the familiar fit and predict methods. The aim is to replace black box models such as random forests with simpler structures, such as rule lists, while keeping predictive accuracy and improving computational efficiency. The package supports a long list of model families, each linked to documentation, a research paper and a reference implementation. They include RuleFit rule sets, Skope rules, boosted and Bayesian rule sets, Bayesian and greedy rule lists, fast and frugal trees, OneR rule lists and greedy rule trees fitted with CART. Models can be drawn as diagrams, printed as text or explored interactively. The home page walks through a worked example on a clinical dataset for cervical spine injury prediction, and demo notebooks are provided. Installation is a single pip command. The project site does not publish any pricing, and the software is distributed as a package rather than a hosted service. The same maintainers point to two companion packages: imodelsX, which covers interpretability in text, and agentic imodels, which provides interpretability tools for tabular data with agents. The listing sits in the coding tools category, but the package is a machine learning library for tabular data rather than a coding assistant.