Explainer API for Local Classifier per parent node #minor

Pipfile

@@ -19,5 +19,5 @@ sphinx-rtd-theme = "0.5.2"
 
 [extras]
 ray = "*"
-shap = "*"
+shap = "0.44.1"
 xarray = "*"

README.md

@@ -34,6 +34,7 @@ HiClass is an open-source Python library for hierarchical classification compati
 - **[Hierarchical metrics](https://hiclass.readthedocs.io/en/latest/api/utilities.html#hierarchical-metrics):** HiClass supports the computation of hierarchical precision, recall and f-score, which are more appropriate for hierarchical data than traditional metrics.
 - **[Compatible with pickle](https://hiclass.readthedocs.io/en/latest/auto_examples/plot_model_persistence.html):** Easily store trained models on disk for future use.
 - **[BERT sklearn](https://hiclass.readthedocs.io/en/latest/auto_examples/plot_bert.html):** Compatible with the library [BERT sklearn](https://github.com/charles9n/bert-sklearn).
+- **[Hierarchical Explanability]():**  HiClass allows explaining hierarchical models using the [SHAP](https://github.com/shap/shap) package.
 
 **Any feature missing on this list?** Search our [issue tracker](https://github.com/scikit-learn-contrib/hiclass/issues) to see if someone has already requested it and add a comment to it explaining your use-case. Otherwise, please open a new issue describing the requested feature and possible use-case scenario. We prioritize our roadmap based on user feedback, so we would love to hear from you.
 
@@ -113,7 +114,7 @@ pip install hiclass"[<extra_name>]"
 Replace <extra_name> with one of the following options:
 
 - ray: Installs the ray package, which is required for parallel processing support.
-- xai: Installs the shap and xarray packages, which are required for explaining Hiclass predictions.
+- xai: Installs the shap and xarray packages, which are required for explaining Hiclass' predictions.
 
 ### Option 2: Conda
 
@@ -199,6 +200,9 @@ pipeline.fit(X_train, Y_train)
 predictions = pipeline.predict(X_test)
 ```
 
+## Explaining Hierarchical Classifiers
+Hierarchical classifiers can provide additional insights when combined with explainability methods. HiClass allows explaining hierarchical models using SHAP values. Different hierarchical models yield different insights. More information on explaining [Local classifier per parent node](https://colab.research.google.com/drive/1rVlYuRU_uO1jw5sD6qo2HoCpCz6E6z5J?usp=sharing), [Local classifier per node](), and [Local classifier per level]() is available on [Read the Docs](https://hiclass.readthedocs.io/en/latest/algorithms/explainer.html).
+
 ## Step-by-step walk-through
 
 A step-by-step walk-through is available on our documentation hosted on [Read the Docs](https://hiclass.readthedocs.io/en/latest/index.html).

docs/examples/plot_lcppn_explainer.py

@@ -0,0 +1,61 @@
+# -*- coding: utf-8 -*-
+"""
+============================================
+Explaining Local Classifier Per Parent Node
+============================================
+
+A minimalist example showing how to use HiClass Explainer to obtain SHAP values of LCPPN model.
+A detailed summary of the Explainer class has been given at Algorithms Overview Section for :ref:`Hierarchical Explainability`.
+SHAP values are calculated based on a synthetic platypus diseases dataset that can be downloaded `here <https://gist.githubusercontent.com/ashishpatel16/9306f8ed3ed101e7ddcb519776bcbd80/raw/3f225c3f80dd8cbb1b6252f6c372a054ec968705/platypus_diseases.csv>`_.
+"""
+from sklearn.ensemble import RandomForestClassifier
+from hiclass import LocalClassifierPerParentNode, Explainer
+import requests
+import pandas as pd
+import shap
+
+# Download training data
+url = "https://gist.githubusercontent.com/ashishpatel16/9306f8ed3ed101e7ddcb519776bcbd80/raw/3f225c3f80dd8cbb1b6252f6c372a054ec968705/platypus_diseases.csv"
+path = "platypus_diseases.csv"
+response = requests.get(url)
+with open(path, "wb") as file:
+    file.write(response.content)
+
+# Load training data into pandas dataframe
+training_data = pd.read_csv(path).fillna(" ")
+
+# Define data
+X_train = training_data.drop(["label"], axis=1)
+X_test = X_train[:100]  # Use first 100 samples as test set
+Y_train = training_data["label"]
+Y_train = [eval(my) for my in Y_train]
+
+# Use random forest classifiers for every node
+rfc = RandomForestClassifier()
+classifier = LocalClassifierPerParentNode(
+    local_classifier=rfc, replace_classifiers=False
+)
+
+# Train local classifier per parent node
+classifier.fit(X_train, Y_train)
+
+# Define Explainer
+explainer = Explainer(classifier, data=X_train, mode="tree")
+explanations = explainer.explain(X_test.values)
+print(explanations)
+
+# Filter samples which only predicted "Respiratory" at first level
+respiratory_idx = classifier.predict(X_test)[:, 0] == "Respiratory"
+
+# Specify additional filters to obtain only level 0
+shap_filter = {"level": 0, "class": "Respiratory", "sample": respiratory_idx}
+
+# Use .sel() method to apply the filter and obtain filtered results
+shap_val_respiratory = explanations.sel(shap_filter)
+
+# Plot feature importance on test set
+shap.plots.violin(
+    shap_val_respiratory.shap_values,
+    feature_names=X_train.columns.values,
+    plot_size=(13, 8),
+)

docs/requirements.txt

@@ -9,3 +9,5 @@ pandas==1.4.2
 ray==1.13.0
 numpy
 git+https://github.com/charles9n/bert-sklearn.git@master
+shap==0.44.1
+xarray==2023.1.0