obliquetree 1.0.5__cp310-cp310-win_amd64.whl

obliquetree-1.0.5.dist-info/METADATA

@@ -0,0 +1,113 @@
+Metadata-Version: 2.4
+Name: obliquetree
+Version: 1.0.5
+Summary: Traditional and Oblique Decision Tree
+Author-email: Samet Copur <sametcopur@yahoo.com>
+License-Expression: MIT
+Project-URL: Documentation, https://obliquetree.readthedocs.io/en/latest/
+Project-URL: Repository, https://github.com/sametcopur/obliquetree
+Project-URL: Tracker, https://github.com/sametcopur/obliquetree/issues
+Keywords: python,data-science,machine-learning,machine-learning-library,explainable-ai,decision-tree,oblique-tree
+Classifier: Programming Language :: Python :: 3
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: numpy>=2.2.1
+Requires-Dist: scipy>=1.15.0
+Dynamic: license-file
+
+# obliquetree
+
+`obliquetree` is an advanced decision tree implementation designed to provide high-performance and interpretable models. It supports both classification and regression tasks, enabling a wide range of applications. By offering traditional and oblique splits, it ensures flexibility and improved generalization with shallow trees. This makes it a powerful alternative to regular decision trees.
+
+
+![Tree Visualization](docs/source/_static/tree_visual.png)
+
+
+----
+
+## Getting Started
+
+`obliquetree` combines advanced capabilities with efficient performance. It supports **oblique splits**, leveraging **L-BFGS optimization** to determine the best linear weights for splits, ensuring both speed and accuracy.
+
+In **traditional mode**, without oblique splits, `obliquetree` outperforms `scikit-learn` in terms of speed and adds support for **categorical variables**, providing a significant advantage over many traditional decision tree implementations.
+
+When the **oblique feature** is enabled, `obliquetree` dynamically selects the optimal split type between oblique and traditional splits. If no weights can be found to reduce impurity, it defaults to an **axis-aligned split**, ensuring robustness and adaptability in various scenarios.
+
+In very large trees (e.g., depth 10 or more), the performance of `obliquetree` may converge closely with **traditional trees**. The true strength of `obliquetree` lies in their ability to perform exceptionally well at **shallower depths**, offering improved generalization with fewer splits. Moreover, thanks to linear projections, `obliquetree` significantly outperform traditional trees when working with datasets that exhibit **linear relationships**.
+
+-----
+## Installation
+To install `obliquetree`, use the following pip command:
+```bash
+pip install obliquetree
+```
+
+Using the `obliquetree` library is simple and intuitive. Here's a more generic example that works for both classification and regression:
+
+
+```python
+from obliquetree import Classifier, Regressor
+
+# Initialize the model (Classifier or Regressor)
+model = Classifier(  # Replace "Classifier" with "Regressor" if performing regression
+    use_oblique=True,       # Enable oblique splits
+    max_depth=2,            # Set the maximum depth of the tree
+    n_pair=2,               # Number of feature pairs for optimization
+    random_state=42,        # Set a random state for reproducibility
+    categories=[0, 10, 32], # Specify which features are categorical
+)
+
+# Train the model on the training dataset
+model.fit(X_train, y_train)
+
+# Predict on the test dataset
+y_pred = model.predict(X_test)
+```
+-----
+
+## Documentation
+For example usage, API details, comparisons with axis-aligned trees, and in-depth insights into the algorithmic foundation, we **strongly recommend** referring to the full [documentation](https://obliquetree.readthedocs.io/en/latest/).
+
+---
+## Key Features
+
+- **Oblique Splits**  
+  Perform oblique splits using linear combinations of features to capture complex patterns in data. Supports both linear and soft decision tree objectives for flexible and accurate modeling.
+
+- **Axis-Aligned Splits**  
+  Offers conventional (axis-aligned) splits, enabling users to leverage standard decision tree behavior for simplicity and interpretability.
+
+- **Feature Constraints**  
+  Limit the number of features used in oblique splits with the `n_pair` parameter, promoting simpler, more interpretable tree structures while retaining predictive power.
+
+- **Seamless Categorical Feature Handling**  
+  Natively supports categorical columns with minimal preprocessing. Only label encoding is required, removing the need for extensive data transformation.
+
+- **Robust Handling of Missing Values**  
+  Automatically assigns `NaN` values to the optimal leaf for axis-aligned splits.
+
+- **Customizable Tree Structures**  
+  The flexible API empowers users to design their own tree architectures easily.
+
+- **Exact Equivalence with `scikit-learn`**  
+  Guarantees results identical to `scikit-learn`'s decision trees when oblique and categorical splitting are disabled.
+
+- **Optimized Performance**  
+  Outperforms `scikit-learn` in terms of speed and efficiency when oblique and categorical splitting are disabled:
+  - Up to **50% faster** for datasets with float columns.
+  - Up to **200% faster** for datasets with integer columns.
+
+  ![Performance Comparison (Float)](docs/source/_static/sklearn_perf/performance_comparison_float.png)
+
+  ![Performance Comparison (Integer)](docs/source/_static/sklearn_perf/performance_comparison_int.png)
+
+
+----
+### Contributing
+Contributions are welcome! If you'd like to improve `obliquetree` or suggest new features, feel free to fork the repository and submit a pull request.
+
+-----
+### License
+`obliquetree` is released under the MIT License. See the LICENSE file for more details.

obliquetree-1.0.5.dist-info/RECORD

@@ -0,0 +1,21 @@
+obliquetree/__init__.py,sha256=-0R6INeDNsCbWWWS8pcJMgfonWcykI_0JcvKvlfOxBo,106
+obliquetree/_pywrap.py,sha256=eUsDuXWkvmEWAVW_a7KRbNbBuKG_dCHjy2rqBRAnnrE,28022
+obliquetree/utils.py,sha256=ewYp_3HZRvgBMysMlnelZdDuMCv49wsPAuYD_6du4GY,31149
+obliquetree/src/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+obliquetree/src/base.cp310-win_amd64.pyd,sha256=9xTBtNkFBsbonmZdeQ-C9wrxFEcm1lWZ6wKZLItYPLA,168960
+obliquetree/src/base.cpp,sha256=Qm2uoVtY_XXlu0rFzlEgmjXbAwq_81QVUdgYCKnVMHs,1466782
+obliquetree/src/ccp.cp310-win_amd64.pyd,sha256=sY0zeJPAdJqoL-uucqCR8Z-9PAEOnw-BaPtQ4GOn2IQ,101376
+obliquetree/src/ccp.cpp,sha256=oxEFvDbNoQJBKquilFy6lWU1612Q9zLOsXrfFwST4IU,972943
+obliquetree/src/metric.cp310-win_amd64.pyd,sha256=73I0PBDmDOG8kpbEqzvBWU9t4lt7N8ZXYBMDKGCnJ_8,121344
+obliquetree/src/metric.cpp,sha256=1q8iVfciap4mdOjZJbsReD8TfEracDl_ndDY-_IEzQY,1216075
+obliquetree/src/oblique.cp310-win_amd64.pyd,sha256=jSmx-_NHJjBvFk14FFrKFbZtmqnsz5BpshK5H8inr_Q,196608
+obliquetree/src/oblique.cpp,sha256=ffKhhm5vQwFy7vWF1lDAvfZWpoMZeV5dOZOg0dGVuBU,1409335
+obliquetree/src/tree.cp310-win_amd64.pyd,sha256=9htc88qYn_21zxhPaGn0rSJliY_MSMrL9YxPzxAHmik,113152
+obliquetree/src/tree.cpp,sha256=Jofyp872pxXcwytHNhv3oi_JpvIxvi6SL9B586VbDoI,1167664
+obliquetree/src/utils.cp310-win_amd64.pyd,sha256=cMHym4tmAu8zQ-pjrbC8vHKBYWg81OEuGTqVXWzA-og,147968
+obliquetree/src/utils.cpp,sha256=9UoaFPO85gFKQ39xxy4glE_Yk5ciYH0N92EPGDtJ1S0,1263005
+obliquetree-1.0.5.dist-info/licenses/LICENSE,sha256=WfEPDyxevjVRSepuK7Lvl2VDl5pqtwjvMoNu2iWhcD8,1090
+obliquetree-1.0.5.dist-info/METADATA,sha256=Gw4fErzCt4_XAYmygxlUP9Q0Mh14MECNoxWN26G6bHw,5727
+obliquetree-1.0.5.dist-info/WHEEL,sha256=KUuBC6lxAbHCKilKua8R9W_TM71_-9Sg5uEP3uDWcoU,101
+obliquetree-1.0.5.dist-info/top_level.txt,sha256=m-5N4-iAS5MsFOdk8y1r2ya_i5rVQBgPayHBA-K26qg,12
+obliquetree-1.0.5.dist-info/RECORD,,

obliquetree-1.0.5.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (80.9.0)
+Root-Is-Purelib: false
+Tag: cp310-cp310-win_amd64
+

obliquetree-1.0.5.dist-info/licenses/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 Samet Çopur
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

obliquetree-1.0.5.dist-info/top_level.txt

@@ -0,0 +1 @@
+obliquetree