Python package for concise, transparent, and accurate predictive modeling. All sklearncompatible and easy to use.
π docs β’ π demo notebooks
Modern machinelearning models are increasingly complex, often making them difficult to interpret. This package provides a simple interface for fitting and using stateoftheart interpretable models, all compatible with scikitlearn. These models can often replace blackbox 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 scikitlearn models.
from sklearn.model_selection import train_test_split
from imodels import get_clean_dataset, HSTreeClassifierCV # import any model here
# 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 = HSTreeRegressorCV(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)
print(model) # print the model

Decision Tree with Hierarchical Shrinkage
Prediction is made by looking at the value in the appropriate leaf of the tree

 FocalNeuroFindings2 <= 0.50
  HighriskDiving <= 0.50
   Torticollis2 <= 0.50
    value: [0.10]
   Torticollis2 > 0.50
    value: [0.30]
  HighriskDiving > 0.50
   value: [0.68]
 FocalNeuroFindings2 > 0.50
  value: [0.42]
Installation
Install with pip install imodels
(see here for help).
Supported models
Model  Reference  Description 

Rulefit rule set  ποΈ, π, π  Fits a sparse linear model on rules extracted from decision trees 
Skope rule set  ποΈ, π  Extracts rules from gradientboosted trees, deduplicates them, then linearly combines them based on their OOB precision 
Boosted rule set  ποΈ, π, π  Sequentially fits a set of rules with Adaboost 
Slipper rule set  ποΈ, γ €γ €π  Sequentially learns a set of rules with SLIPPER 
Bayesian rule set  ποΈ, π, π  Finds concise rule set with Bayesian sampling (slow) 
Optimal rule list  ποΈ, π, π  Fits rule list using global optimization for sparsity (CORELS) 
Bayesian rule list  ποΈ, π, π  Fits compact rule list distribution with Bayesian sampling (slow) 
Greedy rule list  ποΈ, π  Uses CART to fit a list (only a single path), rather than a tree 
OneR rule list  ποΈ, γ €γ €π  Fits rule list restricted to only one feature 
Optimal rule tree  ποΈ, π, π  Fits succinct tree using global optimization for sparsity (GOSDT) 
Greedy rule tree  ποΈ, π, π  Greedily fits tree using CART 
C4.5 rule tree  ποΈ, π, π  Greedily fits tree using C4.5 
TAO rule tree  ποΈ, γ €γ €π  Fits tree using alternating optimization 
Iterative random forest 
ποΈ, π, π  Repeatedly fit random forest, giving features with high importance a higher chance of being selected 
Sparse integer linear model 
ποΈ, γ €γ €π  Sparse linear model with integer coefficients 
Greedy tree sums  ποΈ, γ €γ €π  Sum of small trees with very few total rules (FIGS) 
Hierarchical shrinkage wrapper 
ποΈ, γ €γ €π  Improve any treebased model with ultrafast, posthoc regularization 
Distillation wrapper 
ποΈ  Train a blackbox model, then distill it into an interpretable model 
More models  β  (Coming soon!) Lightweight Rule Induction, MLRules, … 
Docs ποΈ, Reference code implementation π, Research paper π
Demo notebooks
Demos are contained in the notebooks folder.
Quickstart demo
Shows how to fit, predict, and visualize with different interpretable modelsAutogluon demo
Fit/select an interpretable model automatically using Autogluon AutoMLQuickstart colab demo
Shows how to fit, predict, and visualize with different interpretable modelsClinical decision rule notebook
Shows an example of usingimodels
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 modelWhat's the difference between the models?
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, such as how they generate, select, and postprocess rules:
Rule candidate generation  Rule selection  Rule postprocessing 

Ex. RuleFit vs. 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 vs. 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 vs. SkopeRules
FPSkope and SkopeRules differ only in the way they generate candidate rules: FPSkope uses FPgrowth whereas SkopeRules extracts rules from decision trees.Support for different tasks
Different models support different machinelearning 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  Notes 

Rulefit rule set  RuleFitClassifier  RuleFitRegressor  
Skope rule set  SkopeRulesClassifier  
Boosted rule set  BoostedRulesClassifier  
SLIPPER rule set  SlipperClassifier  
Bayesian rule set  BayesianRuleSetClassifier  Fails for large problems  
Optimal rule list (CORELS)  OptimalRuleListClassifier  Requires corels, fails for large problems  
Bayesian rule list  BayesianRuleListClassifier  
Greedy rule list  GreedyRuleListClassifier  
OneR rule list  OneRClassifier  
Optimal rule tree (GOSDT)  OptimalTreeClassifier  Requires gosdt, fails for large problems  
Greedy rule tree (CART)  GreedyTreeClassifier  GreedyTreeRegressor  
C4.5 rule tree  C45TreeClassifier  
TAO rule tree  TaoTreeClassifier  TaoTreeRegressor  
Iterative random forest  IRFClassifier  Requires irf  
Sparse integer linear model  SLIMClassifier  SLIMRegressor  Requires extra dependencies for speed 
Greedy tree sums (FIGS)  FIGSClassifier  FIGSRegressor  
Hierarchical shrinkage  HSTreeClassifierCV  HSTreeRegressorCV  Wraps any sklearn treebased model 
Distillation  DistilledRegressor  Wraps any sklearncompatible models 
Extras
Datawrangling functions for working with popular tabular datasets (e.g. compas).
These functions, in conjunction with imodelsdata and imodelsexperiments, make it simple to download data and run experiments on new models.Explain classification errors with a simple posthoc function.
Fit an interpretable model to explain a previous model's errors (ex. in this notebookπ).Fast and effective discretizers for data preprocessing.
Discretizer  Reference  Description 

MDLP  ποΈ, π, π  Discretize using entropy minimization heuristic 
Simple  ποΈ, π  Simple KBins discretization 
Random Forest  ποΈ  Discretize into bins based on random forest split popularity 
Rulebased utils for customizing models
The code here contains many useful and customizable functions for rulebased learning in the util folder. This includes functions / classes for rule deduplication, rule screening, and converting between trees, rulesets, and neural networks.Our favorite models
After developing and playing with imodels
, we developed a few new models to overcome limitations of existing interpretable models.
FIGS: Fast interpretable greedytree sums
π Paper, π Post, π Citation
Fast Interpretable GreedyTree Sums (FIGS) is an algorithm for fitting concise rulebased 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 prespecified threshold, keeping the model interpretable. Experiments across a wide array of realworld datasets show that FIGS achieves stateoftheart prediction performance when restricted to just a few splits (e.g. less than 20).
Example FIGS model. FIGS learns a sum of trees with a flexible number of trees; to make its prediction, it sums the result from each tree.
Hierarchical shrinkage: posthoc regularization for treebased methods
π Paper (ICML 2022), π Post, π Citation
Hierarchical shrinkage is an extremely fast posthoc regularization method which works on any decision tree (or treebased 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 decisiontree ensembles.
HS Example. HS applies posthoc regularization to any decision tree by shrinking each node towards its parent.
References
Readings
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
 sklearnexpertsys  by @tmadl and @kenben based on original code by Ben Letham
 rulefit  by @christophM
 skoperules  by the skoperules team (including @ngoix, @floriangardin, @datajms, Bibi Ndiaye, Ronan Gautier)
 boa  by @wangtongada
Related packages
 gplearn: symbolic regression/classification
 pysr: fast symbolic regression
 pygam: generative additive models
 interpretml: boostingbased gam
 h20 ai: gams + glms (and more)
 optbinning: data discretization / scoring models
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},
}
Expand source code
"""
.. include:: ../readme.md
"""
# Python `imodels` package for interpretable models compatible with scikitlearn.
# Github repo available [here](https://github.com/csinva/imodels)
from .algebraic.slim import SLIMRegressor, SLIMClassifier
from .discretization.discretizer import RFDiscretizer, BasicDiscretizer
from .discretization.mdlp import MDLPDiscretizer, BRLDiscretizer
from .experimental.bartpy import BART
from .rule_list.bayesian_rule_list.bayesian_rule_list import BayesianRuleListClassifier
from .rule_list.corels_wrapper import OptimalRuleListClassifier
from .rule_list.greedy_rule_list import GreedyRuleListClassifier
from .rule_list.one_r import OneRClassifier
from .rule_set import boosted_rules
from .rule_set.boosted_rules import *
from .rule_set.boosted_rules import BoostedRulesClassifier
from .rule_set.brs import BayesianRuleSetClassifier
from .rule_set.fplasso import FPLassoRegressor, FPLassoClassifier
from .rule_set.fpskope import FPSkopeClassifier
from .rule_set.rule_fit import RuleFitRegressor, RuleFitClassifier
from .rule_set.skope_rules import SkopeRulesClassifier
from .rule_set.slipper import SlipperClassifier
from .tree.c45_tree.c45_tree import C45TreeClassifier
from .tree.cart_ccp import DecisionTreeCCPClassifier, DecisionTreeCCPRegressor, HSDecisionTreeCCPClassifierCV, \
HSDecisionTreeCCPRegressorCV
# from .tree.iterative_random_forest.iterative_random_forest import IRFClassifier
# from .tree.optimal_classification_tree import OptimalTreeModel
from .tree.cart_wrapper import GreedyTreeClassifier, GreedyTreeRegressor
from .tree.figs import FIGSRegressor, FIGSClassifier, FIGSRegressorCV, FIGSClassifierCV
from .tree.gosdt.pygosdt import OptimalTreeClassifier
from .tree.gosdt.pygosdt_shrinkage import HSOptimalTreeClassifier, HSOptimalTreeClassifierCV
from .tree.hierarchical_shrinkage import HSTreeRegressor, HSTreeClassifier, HSTreeRegressorCV, HSTreeClassifierCV
from .tree.tao import TaoTreeClassifier, TaoTreeRegressor
from .util.data_util import get_clean_dataset
from .util.distillation import DistilledRegressor
from .util.explain_errors import explain_classification_errors
from .importance.representation import TreeTransformer
CLASSIFIERS = [BayesianRuleListClassifier, GreedyRuleListClassifier, SkopeRulesClassifier,
BoostedRulesClassifier, SLIMClassifier, SlipperClassifier, BayesianRuleSetClassifier,
C45TreeClassifier, OptimalTreeClassifier, OptimalRuleListClassifier, OneRClassifier,
SlipperClassifier, RuleFitClassifier, TaoTreeClassifier,
FIGSClassifier, HSTreeClassifier, HSTreeClassifierCV] # , IRFClassifier
REGRESSORS = [RuleFitRegressor, SLIMRegressor, GreedyTreeClassifier, FIGSRegressor,
TaoTreeRegressor, HSTreeRegressor, HSTreeRegressorCV, BART]
ESTIMATORS = CLASSIFIERS + REGRESSORS
DISCRETIZERS = [RFDiscretizer, BasicDiscretizer, MDLPDiscretizer, BRLDiscretizer]
Submodules
imodels.algebraic

Generic class for models that take the form of algebraic equations (e.g. linear models).
imodels.discretization
imodels.experimental
imodels.importance

Feature importance methods for black box models
imodels.rule_list

Generic class for models that take the form of a list of rules.
imodels.rule_set

Generic class for models that take the form of a set of (potentially overlapping) rules.
imodels.tree

Generic class for models that take the form of a tree of rules.
imodels.util

Shared utilities for implementing different interpretable models.