{
    "info": {
        "author": "Andrew Ross",
        "author_email": "andrewslavinross@gmail.com",
        "bugtrack_url": null,
        "classifiers": [
            "Development Status :: 3 - Alpha",
            "Intended Audience :: Developers",
            "Intended Audience :: Science/Research",
            "License :: OSI Approved :: Apache Software License",
            "Natural Language :: English",
            "Operating System :: OS Independent",
            "Programming Language :: Python"
        ],
        "description": "Object-Oriented Decision Trees\n==============================\n\nThis repository will contain several variants of decision tree /\nensemble classification algorithms, written in an object-oriented style.\nMy immediate goal is to try to reproduce some of the results from `this\npaper <http://arxiv.org/abs/1507.05444>`__ on canonical correlation\nforests, which I am testing against `the same\ndatasets <https://bitbucket.org/twgr/ccf/src/master/Datasets>`__.\n\nWhere possible, external parameters names will match ``scikit-learn``'s\nimplementations of `decision\ntrees <http://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html>`__\nand `random\nforests <http://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html>`__.\n\nUsage\n-----\n\nOne major difference from ``scikit-learn`` is that datasets and their\nattributes are treated as first-class objects. Additionally, all\nclassifiers must be initialized with their training dataset (as opposed\nto calling ``fit``).\n\n.. code:: python\n\n    from oo_trees.dataset import Dataset\n    from oo_trees.decision_tree import DecisionTree\n    from oo_trees.random_forest import RandomForest\n\n    X = examples # numpy 2D numeric array\n    y = outcomes # numpy 1D array\n\n    dataset = Dataset(X, y)\n\n    training_dataset, test_dataset = dataset.random_split(0.75)\n\n    d_tree = DecisionTree(training_dataset)\n    forest = RandomForest(training_dataset)\n\n    print(d_tree.classify(test_dataset.X[0]))\n    print(forest.classify(test_dataset.X[0]))\n\n    d_tree_confusion_matrix = d_tree.performance_on(test_dataset)\n    forest_confusion_matrix = forest.performance_on(test_dataset)\n\n    print(d_tree_confusion_matrix.accuracy)\n    print(forest_confusion_matrix.accuracy)\n\nWhen initializing datasets, we assume all attributes of the training\nexamples are categorical. If that is not the case, you can pass in an\nadditional ``attribute_types`` variable on initialize:\n\n.. code:: python\n\n    from oo_trees.dataset import Dataset\n    from oo_trees.attribute import NumericAttribute, CategoricalAttribute\n\n    X = examples\n    y = outcomes\n\n    attributes = [\n      NumericAttribute(index=0, name='age'),\n      CategoricalAttribute(index=1, name='sex'),\n      NumericAttribute(index=2, name='income')\n    ]\n\n    dataset = Dataset(X, y, attributes)\n\nThe logic for finding the best split is differs for each attribute type,\nand in the future there may be additional type-specific parameters (such\nas importance or number-to-name mappings) useful for classification or\ndisplay.",
        "description_content_type": null,
        "docs_url": null,
        "download_url": "UNKNOWN",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "http://github.com/asross/oo_trees",
        "keywords": null,
        "license": "Apache Software License",
        "maintainer": null,
        "maintainer_email": null,
        "name": "oo_trees",
        "package_url": "https://pypi.org/project/oo_trees/",
        "platform": "any",
        "project_url": "https://pypi.org/project/oo_trees/",
        "project_urls": {
            "Download": "UNKNOWN",
            "Homepage": "http://github.com/asross/oo_trees"
        },
        "release_url": "https://pypi.org/project/oo_trees/0.0.1/",
        "requires_dist": null,
        "requires_python": null,
        "summary": "Object-oriented implementations of decision tree variants",
        "version": "0.0.1"
    },
    "last_serial": 1900284,
    "releases": {
        "0.0.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "a77b1b32863e724b09484c9c85570e12",
                    "sha256": "f67442d6b872e51a338791e0612210b7c9ba111b0e32efa738b5a90f6ddb7557"
                },
                "downloads": -1,
                "filename": "oo_trees-0.0.1.tar.gz",
                "has_sig": false,
                "md5_digest": "a77b1b32863e724b09484c9c85570e12",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 3938,
                "upload_time": "2016-01-11T18:33:57",
                "url": "https://files.pythonhosted.org/packages/ef/60/e5f388298fd023b70d8b9d1aa053089ca8ced85ee6bab5447ef941eb16bf/oo_trees-0.0.1.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "a77b1b32863e724b09484c9c85570e12",
                "sha256": "f67442d6b872e51a338791e0612210b7c9ba111b0e32efa738b5a90f6ddb7557"
            },
            "downloads": -1,
            "filename": "oo_trees-0.0.1.tar.gz",
            "has_sig": false,
            "md5_digest": "a77b1b32863e724b09484c9c85570e12",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 3938,
            "upload_time": "2016-01-11T18:33:57",
            "url": "https://files.pythonhosted.org/packages/ef/60/e5f388298fd023b70d8b9d1aa053089ca8ced85ee6bab5447ef941eb16bf/oo_trees-0.0.1.tar.gz"
        }
    ]
}