Class LIMEColumnar

java.lang.Object

org.tribuo.classification.explanations.lime.LIMEBase

org.tribuo.classification.explanations.lime.LIMEColumnar

All Implemented Interfaces:

ColumnarExplainer<Regressor>, TabularExplainer<Regressor>

public class LIMEColumnar
extends LIMEBase
implements ColumnarExplainer<Regressor>

Uses the columnar data processing infrastructure to mix text and tabular data.

If the supplied RowProcessor doesn't reference any text or binarised fields then it delegates to LIMEBase.explain(org.tribuo.Example<org.tribuo.classification.Label>), though it's still more expensive at construction time.

See:

 Ribeiro MT, Singh S, Guestrin C.
 "Why should I trust you?: Explaining the predictions of any classifier"
 Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 2016.
 

Field Summary

Fields inherited from class org.tribuo.classification.explanations.lime.LIMEBase

DISTANCE_DELTA, evaluator, explanationTrainer, innerModel, kernelWidth, numSamples, numTrainingExamples, regressionFactory, rng, WIDTH_CONSTANT

Constructor Summary

Constructors

Constructor	Description
LIMEColumnar(SplittableRandom rng, Model<Label> innerModel, SparseTrainer<Regressor> explanationTrainer, int numSamples, RowProcessor<Label> exampleGenerator, Tokenizer tokenizer)	Constructs a LIME explainer for a model which uses the columnar data processing system.

Method Summary

Modifier and Type	Method	Description
LIMEExplanation	explain(Map<String,String> input)	Explains the supplied data.
protected com.oracle.labs.mlrg.olcut.util.Pair<LIMEExplanation, List<Example<Regressor>>>	explainWithSamples(Map<String,String> input)
protected String	nameFeature(String fieldName, String name, int idx)	Generate the feature name by combining the word and index.

Methods inherited from class org.tribuo.classification.explanations.lime.LIMEBase

explain, explainWithSamples, kernelDist, measureDistance, samplePoint, trainExplainer, transformOutput

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

LIMEColumnar

public LIMEColumnar(SplittableRandom rng,
 Model<Label> innerModel,
 SparseTrainer<Regressor> explanationTrainer,
 int numSamples,
 RowProcessor<Label> exampleGenerator,
 Tokenizer tokenizer)

Constructs a LIME explainer for a model which uses the columnar data processing system.

Parameters:

rng - The rng to use for sampling.

innerModel - The model to explain.

explanationTrainer - The trainer for the sparse model used to explain.

numSamples - The number of samples to generate in each explanation.

exampleGenerator - The RowProcessor which converts columnar data into an Example.

tokenizer - The tokenizer to use on any text fields.

Method Details

explain

public LIMEExplanation explain(Map<String,String> input)

Description copied from interface: ColumnarExplainer

Explains the supplied data. The Map is first converted into an Example using a RowProcessor, before being supplied to the internal Model.

Specified by:

explain in interface ColumnarExplainer<Regressor>

Parameters:

input - The data to explain.

Returns:

An Explanation for this data.

explainWithSamples

protected com.oracle.labs.mlrg.olcut.util.Pair<LIMEExplanation, List<Example<Regressor>>> explainWithSamples(Map<String,String> input)

nameFeature

protected String nameFeature(String fieldName,
 String name,
 int idx)

Generate the feature name by combining the word and index.

Parameters:

fieldName - The name of the column this text feature came from.

name - The word.

idx - The index.

Returns:

A string representing both of the inputs.