Implementation:Interpretml Interpret MorrisSensitivity

Metadata

Field	Value
Sources	Repo: InterpretML, Doc: SALib
Domains	Sensitivity_Analysis, Feature_Analysis
Updated	2026-02-07
Type	Wrapper Doc (wraps SALib.sample.morris and SALib.analyze.morris)

Overview

Wrapper tool for Morris sensitivity analysis for blackbox models, integrating the SALib library into the InterpretML API.

Description

The MorrisSensitivity class performs all computation during __init__: it generates Morris sampling trajectories via SALib, evaluates the model on all samples, and computes μ* and σ statistics. The explain_global() method formats these results.

Usage

Use this for efficient global sensitivity screening of any model through the InterpretML visualization pipeline.

Code Reference

Field	Value
Source	interpretml/interpret
File	python/interpret-core/interpret/blackbox/_sensitivity.py
Lines	44-169
Import	`from interpret.blackbox import MorrisSensitivity`
External	SALib.sample.morris, SALib.analyze.morris

Signature:

class MorrisSensitivity(ExplainerMixin):
    available_explanations = ["global"]
    explainer_type = "blackbox"

    def __init__(self, model, data, feature_names=None, feature_types=None, sampler=None, **kwargs):
    def explain_global(self, name=None):

I/O Contract

Init inputs:

Parameter	Type	Required	Notes
model	predict function	Yes
data	ndarray	Yes
feature_names	list	No
feature_types	list	No
sampler	SamplerMixin	No
**kwargs	dict	No	Passed to SALib.analyze.morris.analyze

explain_global inputs:

Parameter	Type	Required	Notes
name	str	No

explain_global output: MorrisExplanation with μ* and σ per feature.

Usage Examples

from interpret.blackbox import MorrisSensitivity
from interpret import show

morris = MorrisSensitivity(rf.predict, X_train, feature_names=feature_names)
morris_global = morris.explain_global(name="Morris Sensitivity")
show(morris_global)

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment