imodels 1.0.1

Implementations of various interpretable models

Interpretable machine-learning models (imodels) 🔍

Python package for concise, transparent, and accurate predictive modeling. All sklearn-compatible and easily customizable.

docs • imodels overview • demo notebooks

imodels overview

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 BayesianRuleListClassifier, GreedyRuleListClassifier, SkopeRulesClassifier # see more models below
from imodels import SLIMRegressor, RuleFitRegressor

model = BayesianRuleListClassifier()  # initialize a model
model.fit(X_train, y_train)   # 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)
print(model) # print the rule-based model

-----------------------------
# the model consists of the following 3 rules
# if X1 > 5: then 80.5% risk
# else if X2 > 5: then 40% risk
# else: 10% risk

Installation

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

Supported models

Model	Reference	Description
Rulefit rule set	🗂️, 🔗, 📄	Extracts rules from a decision tree then builds a sparse linear model with them
Skope rule set	🗂️, 🔗	Extracts rules from gradient-boosted trees, deduplicates them, then forms a linear combination of them based on their OOB precision
Boosted rule set	🗂️, 🔗, 📄	Uses Adaboost to sequentially learn a set of rules
Slipper rule set	🗂️, 📄	Uses SLIPPER to sequentially learn a set of rules
Bayesian rule set	🗂️, 🔗, 📄	Uses a Bayesian or-of-and algorithm to find a concise rule set
Corels rule list	🗂️, 🔗, 📄	Learns an optimal compact rule list (rather than using a greedy heuristic)
Bayesian rule list	🗂️, 🔗, 📄	Learns a compact rule list distribution by sampling rule lists (slow)
Greedy rule list	🗂️, 🔗	Uses CART to learn a list (only a single path), rather than a decision tree
OneR rule list	🗂️, 📄	Learns rule list restricted to only one feature
Optimal rule tree	🗂️, 🔗, 📄	(In progress) Learns succinct trees using global optimization rather than greedy heuristics
Iterative random forest	🗂️, 🔗, 📄	(In progress) Repeatedly fit random forest, giving features with high importance a higher chance of being selected
Sparse integer linear model	🗂️, 📄	Forces coefficients to be integers
More models	⌛	(Coming soon!) Popular rule sets including Lightweight Rule Induction, MLRules

Docs 🗂️, Reference code implementation 🔗, Research paper 📄

See also our fast and effective discretizers for data preprocessing.

Discretizer	Reference	Description
MDLP	🗂️, 🔗, 📄	Discretize using entropy minimization heuristic
Simple	🗂️, 🔗	Simple KBins discretization
Random Forest	🗂️	Discretize into bins based on random forest split popularity

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

Rule set	Rule list	Rule tree	Algebraic models

Different models and algorithms vary not only in their final form but also in different choices made during modeling. In particular, many models differ in the 3 steps given by the table below.

ex. RuleFit and SkopeRules

RuleFit and SkopeRules differ only in the way they prune rules: RuleFit uses a linear model whereas SkopeRules heuristically deduplicates rules sharing overlap.

ex. Bayesian rule lists and greedy rule lists

Bayesian rule lists and greedy rule lists differ in how they select rules; bayesian rule lists perform a global optimization over possible rule lists while Greedy rule lists pick splits sequentially to maximize a given criterion.

ex. FPSkope and SkopeRules

FPSkope and SkopeRules differ only in the way they generate candidate rules: FPSkope uses FPgrowth whereas SkopeRules extracts rules from decision trees.

See the docs for individual models for futher descriptions.

Rule candidate generation	Rule selection	Rule pruning / combination

The code here contains many useful and customizable functions for rule-based learning in the util folder. This includes functions / classes for rule deduplication, rule screening, and converting between trees, rulesets, and neural networks.

Demo notebooks

Demos are contained in the notebooks folder.

imodels demo

Shows how to fit, predict, and visualize with different interpretable models

imodels colab demo

Shows how to fit, predict, and visualize with different interpretable models

clinical decision rule notebook

Shows an example of using imodels for deriving a clinical decision rule

posthoc analysis

We also include some demos of posthoc analysis, which occurs after fitting models: posthoc.ipynb shows different simple analyses to interpret a trained model and uncertainty.ipynb contains basic code to get uncertainty estimates for a model

Support for different tasks

Different models support different machine-learning tasks. Current support for different models is given below (each of these models can be imported directly from imodels (e.g. from imodels import RuleFitClassifier):

Model	Binary classification	Regression
Rulefit rule set	RuleFitClassifier	RuleFitRegressor
Skope rule set	SkopeRulesClassifier
Boosted rule set	BoostedRulesClassifier
SLIPPER rule set	SlipperClassifier
BOA rule set	BayesianRuleSetClassifier
Corels rule list	CorelsRuleListClassifier
Bayesian rule list	BayesianRuleListClassifier
Greedy rule list	GreedyRuleListClassifier
OneR rule list	OneRClassifier
Optimal rule tree
Iterative random forest
Sparse integer linear model	SLIMClassifier	SLIMRegressor

References

Readings

• Interpretable ML good quick overview: murdoch et al. 2019, pdf
• Interpretable ML book: molnar 2019, pdf
• Case for interpretable models rather than post-hoc explanation: rudin 2019, pdf
• Review on evaluating interpretability: doshi-velez & kim 2017, pdf

Reference implementations (also linked above)

The code here heavily derives from the wonderful work of previous projects. We seek to to extract out, unify, and maintain key parts of these projects.

• pycorels - by @fingoldin and the original CORELS team
• sklearn-expertsys - by @tmadl and @kenben based on original code by Ben Letham
• rulefit - by @christophM
• skope-rules - by the skope-rules team (including @ngoix, @floriangardin, @datajms, Bibi Ndiaye, Ronan Gautier)
• boa - by @wangtongada

Related packages

• gplearn: symbolic regression/classification
• pygam: generative additive models
• interpretml: boosting-based gam

Updates

• For updates, star the repo, see this related repo, or follow @csinva_
• Please make sure to give authors of original methods / base implementations appropriate credit!
• Contributing: pull requests very welcome!

If it's useful for you, please star/cite the package, and make sure to give authors of original methods / base implementations credit:

@software{
    imodels2021,
    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},
}