Using LIME Explainer

This example illustrates how to use the LIME tabular explainer to explain a prediction.

This example shows how to use the tabular LIME implementation – fatf.transparency.predictions.surrogate_explainers.TabularBlimeyLime – to explain a prediction of a black-box probabilistic model.

Out:

{'setosa': {'*petal length (cm)* <= 1.60': 0.9307866199252784,
            '*petal width (cm)* <= 0.30': 0.04259129805843401,
            '*sepal length (cm)* <= 5.10': -0.0027663756624473936,
            '3.00 < *sepal width (cm)* <= 3.30': -0.011263216273426888},
 'versicolor': {'*petal length (cm)* <= 1.60': -0.5593822883064358,
                '*petal width (cm)* <= 0.30': 0.07065024115814264,
                '*sepal length (cm)* <= 5.10': 0.025538640983435418,
                '3.00 < *sepal width (cm)* <= 3.30': 0.0663293593939103},
 'virginica': {'*petal length (cm)* <= 1.60': -0.37140433161884245,
               '*petal width (cm)* <= 0.30': -0.11324153921657666,
               '*sepal length (cm)* <= 5.10': -0.022772265320988014,
               '3.00 < *sepal width (cm)* <= 3.30': -0.05506614312048342}}

# Author: Kacper Sokol <k.sokol@bristol.ac.uk>
# License: new BSD

from pprint import pprint

import fatf

import fatf.utils.data.datasets as fatf_datasets
import fatf.utils.models as fatf_models

import fatf.transparency.predictions.surrogate_explainers as fatf_surrogates

import fatf.vis.lime as fatf_vis_lime

print(__doc__)

# Fix random seed
fatf.setup_random_seed(42)

# Load data
iris_data_dict = fatf_datasets.load_iris()
iris_X = iris_data_dict['data']
iris_y = iris_data_dict['target']
iris_feature_names = iris_data_dict['feature_names'].tolist()
iris_class_names = iris_data_dict['target_names'].tolist()

# Train a model
clf = fatf_models.KNN()
clf.fit(iris_X, iris_y)

# Create a LIME explainer
lime = fatf_surrogates.TabularBlimeyLime(
    iris_X,
    clf,
    feature_names=iris_feature_names,
    class_names=iris_class_names)

# Choose an index of the instance to be explained
index_to_explain = 42

# Explain an instance
lime_explanation = lime.explain_instance(
    iris_X[index_to_explain, :], samples_number=500)

# Display the textual explanation
pprint(lime_explanation)

# Plot the explanation
fatf_vis_lime.plot_lime(lime_explanation)