{
    "info": {
        "author": "Ralph Haygood",
        "author_email": "ralph@ralphhaygood.org",
        "bugtrack_url": null,
        "classifiers": [
            "Development Status :: 5 - Production/Stable",
            "Intended Audience :: Science/Research",
            "License :: OSI Approved :: MIT License",
            "Natural Language :: English",
            "Programming Language :: Python :: 2",
            "Programming Language :: Python :: 2.7",
            "Programming Language :: Python :: 3",
            "Programming Language :: Python :: 3.5",
            "Topic :: Scientific/Engineering"
        ],
        "description": "sklearn-gbmi: scikit-learn gradient-boosting-model interactions\n===============================================================\n\nThis distribution provides a Python module for computing Friedman and Popescu's *H* statistics, in order to look for\ninteractions among variables in scikit-learn gradient-boosting models\n(http://scikit-learn.org/stable/modules/ensemble.html#gradient-tree-boosting).\n\nSee Jerome H. Friedman and Bogdan E. Popescu, 2008, \"Predictive learning via rule ensembles\", *Ann. Appl. Stat.*\n**2**:916-954, http://projecteuclid.org/download/pdfview_1/euclid.aoas/1223908046, s. 8.1.\n\n\nInstallation\n------------\n\n    pip install sklearn-gbmi\n\n\nUsage\n-----\n\nGiven a scikit-learn gradient-boosting model `gbm` that has been fitted to a NumPy array or pandas data frame\n`array_or_frame` and a list of indices of columns of the array or columns of the data frame `indices_or_columns`, the\n*H* statistic of the variables represented by the elements of `array_or_frame` and specified by `indices_or_columns` can\nbe computed via\n\n    from sklearn_gbmi import *\n\n    h(gbm, array_or_frame, indices_or_columns)\n\nAlternatively, the two-variable *H* statistic of each pair of variables represented by the elements of `array_or_frame`\nand specified by `indices_or_columns` can be computed via\n\n    from sklearn_gbmi import *\n\n    h_all_pairs(gbm, array_or_frame, indices_or_columns)\n\n(Compared to iteratively calling `h`, calling `h_all_pairs` avoids redundant computations.)\n\n`indices_or_columns` is optional, with default value `'all'`. If it is `'all'`, then all columns of `array_or_frame` are\nused.\n\n`NaN` is returned if a computation is spoiled by weak main effects and rounding errors.\n\n*H* varies from 0 to 1. The larger *H*, the stronger the evidence for an interaction among the variables.\n\n\nExample\n-------\n\nSee the Jupyter notebook example.ipynb (https://github.com/ralphhaygood/sklearn-gbmi/blob/master/example.ipynb) for a\ncomplete example of how to use the module.\n\n\nNotes\n-----\n\n1. Per Friedman and Popescu, only variables with strong main effects should be examined for interactions. Strengths of\nmain effects are available as `gbm.feature_importances_` once `gbm` has been fitted.\n\n2. Per Friedman and Popescu, collinearity among variables can lead to interactions in `gbm` that are not present in the\ntarget function. To forestall such spurious interactions, check for strong correlations among variables before fitting\n`gbm`.",
        "description_content_type": null,
        "docs_url": null,
        "download_url": "https://github.com/ralphhaygood/sklearn-gbmi/tarball/1.0.0",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/ralphhaygood/sklearn-gbmi",
        "keywords": "boosted,boosting,data science,Friedman,gradient boosted,gradient boosting,H statistic,interaction,machine learning,Popescu,scikit learn,sklearn",
        "license": "MIT",
        "maintainer": null,
        "maintainer_email": null,
        "name": "sklearn-gbmi",
        "package_url": "https://pypi.org/project/sklearn-gbmi/",
        "platform": "UNKNOWN",
        "project_url": "https://pypi.org/project/sklearn-gbmi/",
        "project_urls": {
            "Download": "https://github.com/ralphhaygood/sklearn-gbmi/tarball/1.0.0",
            "Homepage": "https://github.com/ralphhaygood/sklearn-gbmi"
        },
        "release_url": "https://pypi.org/project/sklearn-gbmi/1.0.0/",
        "requires_dist": null,
        "requires_python": null,
        "summary": "Compute Friedman and Popescu's H statistics, in order to look for interactions among variables in scikit-learn gradient-boosting models.",
        "version": "1.0.0"
    },
    "last_serial": 2574547,
    "releases": {
        "1.0.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "5aea27938b5841a60517381759c0befc",
                    "sha256": "7b49aaf248eaef969f6c517c727ea4f8506bbd4b845fb920963a6535a5df0f53"
                },
                "downloads": -1,
                "filename": "sklearn-gbmi-1.0.0.zip",
                "has_sig": false,
                "md5_digest": "5aea27938b5841a60517381759c0befc",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 11144,
                "upload_time": "2017-01-14T21:27:04",
                "url": "https://files.pythonhosted.org/packages/e2/41/c2793c52cd0046efcba1ab41453c64039d47090f2bb5eb75bd123e9ffb74/sklearn-gbmi-1.0.0.zip"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "5aea27938b5841a60517381759c0befc",
                "sha256": "7b49aaf248eaef969f6c517c727ea4f8506bbd4b845fb920963a6535a5df0f53"
            },
            "downloads": -1,
            "filename": "sklearn-gbmi-1.0.0.zip",
            "has_sig": false,
            "md5_digest": "5aea27938b5841a60517381759c0befc",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 11144,
            "upload_time": "2017-01-14T21:27:04",
            "url": "https://files.pythonhosted.org/packages/e2/41/c2793c52cd0046efcba1ab41453c64039d47090f2bb5eb75bd123e9ffb74/sklearn-gbmi-1.0.0.zip"
        }
    ]
}