Python package for concise, transparent, and accurate predictive modeling.
All sklearn-compatible and easy to use.
Check out our new packages! Interpretability in text: imodelsX, interpretability tools for tabular data with agents: agentic-imodels

📚 docs • 📖 demo notebooks

Modern machine-learning models are increasingly complex, often making them difficult to interpret. This package provides a simple interface for fitting and using state-of-the-art interpretable models, all compatible with scikit-learn. These models can often replace black-box models (e.g. random forests) with simpler models (e.g. rule lists) while improving interpretability and computational efficiency, all without sacrificing predictive accuracy! Simply import a classifier or regressor and use the fit and predict methods, same as standard scikit-learn models.

from imodels import get_clean_dataset, viz, HSTreeClassifierCV # import any imodels model here
from sklearn.model_selection import train_test_split

# prepare data (a sample clinical dataset)
X, y, feature_names = get_clean_dataset('csi_pecarn_pred')
X_train, X_test, y_train, y_test = train_test_split(
    X, y, random_state=42)

# fit the model
model = HSTreeClassifierCV(max_leaf_nodes=4)  # initialize a tree model and specify only 4 leaf nodes
model.fit(X_train, y_train, feature_names=feature_names)   # fit model
preds = model.predict(X_test) # discrete predictions: shape is (n_test, 1)
preds_proba = model.predict_proba(X_test) # predicted probabilities: shape is (n_test, n_classes)
viz.draw(model, X_train, y_train, feature_names=feature_names, class_names=["no CSI", "CSI"],
         title="Cervical spine injury", orientation="LR").save("model.svg")  # draw the model (or print(model) for text, viz.interactive(model) for interaction)

Installation

Install with pip install imodels (see here for help).

Supported models

🗂️ Docs   📄 Research paper   🔗 Reference code implementation

Demo notebooks

Demos are contained in the notebooks folder

Model categories

The final form of the above models takes one of the following forms, which aim to be simultaneously simple to understand and highly predictive:

model = FIGSClassifier().fit(X, y, categorical_features=['pet', 'city'])
model.predict(X)

import dtreeviz
from imodels import FIGSClassifier, shadow_tree

model = FIGSClassifier(max_rules=6).fit(X, y)
viz = dtreeviz.trees.DTreeVizAPI(shadow_tree(model, X, y))
viz.view()

from imodels import FIGSClassifier

model = FIGSClassifier(max_rules=4).fit(X_train, y_train, feature_names=feature_names)
model.get_rules()

                                               rule  prediction  tree
0                        FocalNeuroFindings2 <= 0.5       0.117     0
1                         FocalNeuroFindings2 > 0.5       0.427     0
2                             HighriskDiving <= 0.5      -0.008     1
3                              HighriskDiving > 0.5       0.550     1
4  PainNeck2 <= 0.5 and AlteredMentalStatus2 <= 0.5      -0.083     2
5   PainNeck2 <= 0.5 and AlteredMentalStatus2 > 0.5       0.048     2
6                                   PainNeck2 > 0.5       0.058     2

Extras

Our favorite methods

After developing and playing with imodels, we developed a few new models to overcome limitations of existing interpretable models.

FIGS: Fast interpretable greedy-tree sums

📄 Paper, 🔗 Post, 📌 Citation

Fast Interpretable Greedy-Tree Sums (FIGS) is an algorithm for fitting concise rule-based models. Specifically, FIGS generalizes CART to simultaneously grow a flexible number of trees in a summation. The total number of splits across all the trees can be restricted by a pre-specified threshold, keeping the model interpretable. Experiments across a wide array of real-world datasets show that FIGS achieves state-of-the-art prediction performance when restricted to just a few splits (e.g. less than 20).

FastRiskScore: sparse integer risk scores

🔗 Post, 🗂️ API

FastRiskScore fits a risk score: at most k conditions, each worth a few integer points, with a calibrated risk for every total. It came out of an autoresearch loop that started from FasterRisk, and adds a local search over the integer points to FasterRisk's beam search. Numeric columns are split at quantiles (n_thresholds, default 9), and the search needs numba (compiled once per machine, in about two minutes).

On 27 held-out TabArena datasets it fits about 225× faster than FasterRisk with a lower training loss, and on small problems it finds the best possible score in 49 of 50 cases (FasterRisk: 28).

GPGam: additive Gaussian processes over binned features

🔗 Post, 🗂️ API

GPGam fits a generalized additive model with pairwise interactions, in which every shape function is a Gaussian process over the quantile bins of its feature. Binning reduces the exact likelihood to a few sums over the data, so the likelihood alone picks the smoothness, drops irrelevant features and chooses the interactions: there is nothing to tune, fits are deterministic, and every curve comes with a posterior band.

It is the strongest interpretable model on our development suite (65 datasets) and on held-out TabArena and OpenML-CTR23 datasets.

FastSmallTree: provably optimal small decision trees

🔗 Post, 🗂️ API

FastSmallTree fits the decision tree that minimizes error plus a penalty per leaf over all trees on the binarized features, and certifies that no other tree scores better (the objective of GOSDT and STreeD). It came out of an autoresearch loop and compiles its branch-and-bound search with numba (about 20 seconds once per machine). regularization is the only hyperparameter.

On held-out TabArena benchmarks it matches the trees of existing optimal-tree packages while running faster, sometimes by more than 20×.

Hierarchical shrinkage: post-hoc regularization for tree-based methods

📄 Paper (ICML 2022), 🔗 Post, 📌 Citation

Hierarchical shrinkage is an extremely fast post-hoc regularization method which works on any decision tree (or tree-based ensemble, such as Random Forest). It does not modify the tree structure, and instead regularizes the tree by shrinking the prediction over each node towards the sample means of its ancestors (using a single regularization parameter). Experiments over a wide variety of datasets show that hierarchical shrinkage substantially increases the predictive performance of individual decision trees and decision-tree ensembles.

MDI+: Flexible Tree-Based Feature Importance

📄 Paper, 🔗 Post, 📌 Citation

MDI+ is a novel feature importance framework, which generalizes the popular mean decrease in impurity (MDI) importance score for random forests. At its core, MDI+ expands upon a recently discovered connection between linear regression and decision trees. In doing so, MDI+ enables practitioners to (1) tailor the feature importance computation to the data/problem structure and (2) incorporate additional features or knowledge to mitigate known biases of decision trees. In both real data case studies and extensive real-data-inspired simulations, MDI+ outperforms commonly used feature importance measures (e.g., MDI, permutation-based scores, and TreeSHAP) by substantional margins.

References

Please cite the package if you use it in an academic work :)

@software{
	singh2021imodels,
	title        = {imodels: a python package for fitting interpretable models},
	journal      = {Journal of Open Source Software},
	publisher    = {The Open Journal},
	year         = {2021},
	author       = {Singh, Chandan and Nasseri, Keyan and Tan, Yan Shuo and Tang, Tiffany and Yu, Bin},
	volume       = {6},
	number       = {61},
	pages        = {3192},
	doi          = {10.21105/joss.03192},
	url          = {https://doi.org/10.21105/joss.03192},
}