fast-causal-shap 0.1.0__py3-none-any.whl → 0.1.2__py3-none-any.whl

fast_causal_shap-0.1.2.dist-info/METADATA

@@ -0,0 +1,96 @@
+Metadata-Version: 2.4
+Name: fast-causal-shap
+Version: 0.1.2
+Summary: A Python package for efficient causal SHAP computations
+Author-email: woonyee28 <ngnwy289@gmail.com>
+License: MIT
+Project-URL: Homepage, https://github.com/woonyee28/CausalSHAP
+Project-URL: Issues, https://github.com/woonyee28/CausalSHAP/issues
+Requires-Python: >=3.7
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: pandas>=1.0.0
+Requires-Dist: networkx>=2.0
+Requires-Dist: numpy>=1.18.0
+Requires-Dist: scikit-learn>=0.24.0
+Dynamic: license-file
+
+# Fast Causal SHAP
+
+Fast Causal SHAP is a Python package designed for efficient and interpretable SHAP value computation in causal inference tasks. It integrates seamlessly with various causal inference frameworks and enables feature attribution with awareness of causal dependencies.
+
+## Features
+
+- Fast computation of SHAP values for causal models
+- Support for multiple causal inference frameworks
+
+## Installation
+
+Install the stable version via PyPI:
+
+```bash
+pip install fast-causal-shap
+```
+
+Or, for the latest development version:
+
+```bash
+pip install git+https://github.com/woonyee28/CausalSHAP.git
+```
+
+## Usage
+```
+from fast_causal_shap.core import FastCausalSHAP
+
+# Predict probabilities and assign to training data
+predicted_probabilities = model.predict_proba(X_train)[:,1]
+X_train['target'] = predicted_probabilities
+
+# Initialize FastCausalInference
+ci = FastCausalInference(data=X_train, model=model, target_variable='target')
+
+# Load causal strengths (precomputed using R packages)
+ci.load_causal_strengths(result_dir + 'Causal_Effect.json')
+
+# Compute modified SHAP values for a single instance
+x_instance = X_train.iloc[33]
+
+print(ci.compute_modified_shap_proba(x_instance, is_classifier=True))
+```
+
+Format of the Causal_Effect.json:
+```
+[
+  {
+    "Pair": "Bacteroidia->Clostridia",
+    "Mean_Causal_Effect": 0.71292
+  },
+  {
+    "Pair": "Clostridia->Alphaproteobacteria",
+    "Mean_Causal_Effect": 0.37652
+  }, ......
+]
+```
+
+Fast Causal SHAP supports integration with structural algorithms such as:        
+1. Peter-Clarke (PC) Algorithm     
+2. IDA Algorithm     
+3. Fast Causal Inference (FCI) Algorithm    
+You can find example R code for these integrations here: [FastCausalSHAP R code examples](https://github.com/woonyee28/CausalSHAP/tree/main/code/r)
+
+
+## Citation
+If you use Fast Causal SHAP in your research, please cite:
+```
+@inproceedings{ng2025causal,
+  title={Causal SHAP: Feature Attribution with Dependency Awareness through Causal Discovery},
+  author={Ng, Woon Yee and Wang, Li Rong and Liu, Siyuan and Fan, Xiuyi},
+  booktitle={Proceedings of the International Joint Conference on Neural Networks (IJCNN)},
+  year={2025},
+  organization={IEEE}
+}
+```
+
+## License
+
+This project is licensed under the MIT License.

fast_causal_shap-0.1.2.dist-info/RECORD

@@ -0,0 +1,7 @@
+fast_causal_shap/__init__.py,sha256=d1piIUSJI81MVDdqtqzdteGKWO1ShUP0KUoqD8CoZa4,226
+fast_causal_shap/core.py,sha256=9FSFdPLV5TCQIW8p2Bu8W4icZjvNb5MnQOdDkOQHCTM,12505
+fast_causal_shap-0.1.2.dist-info/licenses/LICENSE,sha256=pAZXnNE2dxxwXFIduGyn1gpvPefJtUYOYZOi3yeGG94,1068
+fast_causal_shap-0.1.2.dist-info/METADATA,sha256=QhCik5sM_Sng_YXDnabuDKAf6Akt46th_dqqo2wR44Q,2796
+fast_causal_shap-0.1.2.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+fast_causal_shap-0.1.2.dist-info/top_level.txt,sha256=nAIqoFfVB4g6cJal-o9z4LmDYIX1lj1x15oJrlsT_4E,17
+fast_causal_shap-0.1.2.dist-info/RECORD,,

fast_causal_shap-0.1.0.dist-info/METADATA

@@ -1,50 +0,0 @@
-Metadata-Version: 2.4
-Name: fast-causal-shap
-Version: 0.1.0
-Summary: A Python package for efficient causal SHAP computations
-Author-email: woonyee28 <ngnwy289@gmail.com>
-License: MIT
-Project-URL: Homepage, https://github.com/woonyee28/CausalSHAP
-Project-URL: Issues, https://github.com/woonyee28/CausalSHAP/issues
-Requires-Python: >=3.7
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: pandas>=1.0.0
-Requires-Dist: networkx>=2.0
-Requires-Dist: numpy>=1.18.0
-Requires-Dist: scikit-learn>=0.24.0
-Dynamic: license-file
-
-# Fast Causal SHAP
-
-This folder contains the core modules and components for the **Fast Causal SHAP** Python package. Fast Causal SHAP provides efficient and interpretable SHAP value computation for causal inference tasks.
-
-## Features
-
-- Fast computation of SHAP values for causal models
-- Support for multiple causal inference frameworks
-
-## Installation
-
-Install Fast Causal SHAP using pip:
-
-```bash
-pip install fast-causal-shap
-```
-
-Or, for the latest development version:
-
-```bash
-pip install git+https://github.com/woonyee28/CausalSHAP.git
-```
-
-## Usage
-// To be added
-
-## Citation
-If you use this package in your research, please cite:
-// To be added
-
-## License
-
-This project is licensed under the MIT License.

fast_causal_shap-0.1.0.dist-info/RECORD

@@ -1,7 +0,0 @@
-fast_causal_shap/__init__.py,sha256=d1piIUSJI81MVDdqtqzdteGKWO1ShUP0KUoqD8CoZa4,226
-fast_causal_shap/core.py,sha256=9FSFdPLV5TCQIW8p2Bu8W4icZjvNb5MnQOdDkOQHCTM,12505
-fast_causal_shap-0.1.0.dist-info/licenses/LICENSE,sha256=pAZXnNE2dxxwXFIduGyn1gpvPefJtUYOYZOi3yeGG94,1068
-fast_causal_shap-0.1.0.dist-info/METADATA,sha256=g97MKoflaX8Wy_3jIKxPmZ5ZSLvNcyqrse8XvR6y7Iw,1264
-fast_causal_shap-0.1.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-fast_causal_shap-0.1.0.dist-info/top_level.txt,sha256=nAIqoFfVB4g6cJal-o9z4LmDYIX1lj1x15oJrlsT_4E,17
-fast_causal_shap-0.1.0.dist-info/RECORD,,