{
    "info": {
        "author": "Brian Williamson",
        "author_email": "brianw26@uw.edu",
        "bugtrack_url": null,
        "classifiers": [
            "License :: OSI Approved :: MIT License",
            "Operating System :: OS Independent",
            "Programming Language :: Python :: 2",
            "Programming Language :: Python :: 2.6",
            "Programming Language :: Python :: 2.7"
        ],
        "description": "# vimpy: nonparametric variable importance assessment in python\n\n[![PyPI version](https://badge.fury.io/py/vimpy.svg)](https://badge.fury.io/py/vimpy)\n[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)\n\n\n**Author:** Brian Williamson\n\n## Introduction\n\nIn predictive modeling applications, it is often of interest to determine the relative contribution of subsets of features in explaining an outcome; this is often called variable importance. It is useful to consider variable importance as a function of the unknown, underlying data-generating mechanism rather than the specific predictive algorithm used to fit the data. This package provides functions that, given fitted values from predictive algorithms, compute nonparametric estimates of deviance- and variance-based variable importance, along with asymptotically valid confidence intervals for the true importance.\n\n## Installation\n\nYou may install a stable release of `vimpy` using `pip` by running `python pip install vimpy` from a Terminal window. Alternatively, you may install within a `virtualenv` environment.\n\nYou may install the current dev release of `vimpy` by downloading this repository directly.\n\n## Issues\n\nIf you encounter any bugs or have any specific feature requests, please [file an issue](https://github.com/bdwilliamson/vimpy/issues).\n\n## Example\n\nThis example shows how to use `vimpy` in a simple setting with simulated data and using a single regression function. For more examples and detailed explanation, please see the `R` vignette (to come).\n\n```python\n## load required libraries\nimport numpy as np\nimport vimpy\nfrom sklearn.ensemble import GradientBoostingRegressor\nfrom sklearn.model_selection import GridSearchCV\n\n## -------------------------------------------------------------\n## problem setup\n## -------------------------------------------------------------\n## define a function for the conditional mean of Y given X\ndef cond_mean(x = None):\n    f1 = np.where(np.logical_and(-2 <= x[:, 0], x[:, 0] < 2), np.floor(x[:, 0]), 0) \n    f2 = np.where(x[:, 1] <= 0, 1, 0)\n    f3 = np.where(x[:, 2] > 0, 1, 0)\n\n    f6 = np.absolute(x[:, 5]/4) ** 3\n    f7 = np.absolute(x[:, 6]/4) ** 5\n\n    f11 = (7./3)*np.cos(x[:, 10]/2)\n\n    ret = f1 + f2 + f3 + f6 + f7 + f11\n\n    return ret\n\n## create data\nnp.random.seed(4747)\nn = 100\np = 15\ns = 1 # importance desired for X_1\nx = np.zeros((n, p))\nfor i in range(0, x.shape[1]) :\n    x[:,i] = np.random.normal(0, 2, n)\n\ny = cond_mean(x) + np.random.normal(0, 1, n)\n\n## -------------------------------------------------------------\n## preliminary step: get regression estimators\n## -------------------------------------------------------------\n## use grid search to get optimal number of trees and learning rate\nntrees = np.arange(100, 3500, 500)\nlr = np.arange(.01, .5, .05)\n\nparam_grid = [{'n_estimators':ntrees, 'learning_rate':lr}]\n\n## set up cv objects\ncv_full = GridSearchCV(GradientBoostingRegressor(loss = 'ls', max_depth = 1), param_grid = param_grid, cv = 5)\ncv_small = GridSearchCV(GradientBoostingRegressor(loss = 'ls', max_depth = 1), param_grid = param_grid, cv = 5)\n\n## fit the full regression\ncv_full.fit(x, y)\nfull_fit = cv_full.best_estimator_.predict(x)\n\n## fit the reduced regression\nx_small = np.delete(x, s, 1) # delete the columns in s\ncv_small.fit(x_small, full_fit)\nsmall_fit = cv_small.best_estimator_.predict(x_small)\n\n## -------------------------------------------------------------\n## get variable importance estimates\n## -------------------------------------------------------------\n## set up the vimp object\nvimp = vimpy.vimp_regression(y, x, full_fit, small_fit, s)\n## get the naive estimator\nvimp.plugin()\n## get the corrected estimator\nvimp.update()\nvimp.onestep_based_estimator()\n## get a standard error\nvimp.onestep_based_se()\n## get a confidence interval\nvimp.get_ci()\n\n## -------------------------------------------------------------\n## get variable importance estimates using cross-validation\n## -------------------------------------------------------------\nfull_fits = [None]*V\nsmall_fits = [None]*V\nfor v in range(V):\n    cv_full.fit(x[folds == v, :], y[folds == v])\n    full_fits[v] = cv_full.best_estimator_.predict(x[folds == v, :])\n    x_small = np.delete(x[folds == v, :], s, 1) # delete the columns in s\n    cv_small.fit(x_small, full_fits[v])\n    small_fits[v] = cv_small.best_estimator_.predict(x_small)\n\n## set up the outcome and vimp object\nys = [y[folds == v] for v in range(V)]\nvimp_cv = vimpy.cv_vim(ys, x, full_fits, small_fits, V, folds, \"regression\", s)\n## get the naive estimator\nvimp_cv.plugin()\n## get the corrected estimator\nvimp_cv.update()\nvimp_cv.onestep_based_estimator()\n## get a standard error\nvimp_cv.onestep_based_se()\n## get a confidence interval\nvimp_cv.get_ci()\n```\n\n",
        "description_content_type": "text/markdown",
        "docs_url": null,
        "download_url": "",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/bdwilliamson/vimpy",
        "keywords": "",
        "license": "",
        "maintainer": "",
        "maintainer_email": "",
        "name": "vimpy",
        "package_url": "https://pypi.org/project/vimpy/",
        "platform": "",
        "project_url": "https://pypi.org/project/vimpy/",
        "project_urls": {
            "Homepage": "https://github.com/bdwilliamson/vimpy"
        },
        "release_url": "https://pypi.org/project/vimpy/1.0.0/",
        "requires_dist": null,
        "requires_python": "",
        "summary": "vimpy: nonparametric variable importance assessment in python",
        "version": "1.0.0"
    },
    "last_serial": 4424849,
    "releases": {
        "0.0.10": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "1068b3395e7df5035510f0213323c720",
                    "sha256": "a4f8191b722f25452aa1096209be0b9f950f6823e99073f55fe645cee659d94d"
                },
                "downloads": -1,
                "filename": "vimpy-0.0.10-py2-none-any.whl",
                "has_sig": false,
                "md5_digest": "1068b3395e7df5035510f0213323c720",
                "packagetype": "bdist_wheel",
                "python_version": "py2",
                "requires_python": null,
                "size": 5327,
                "upload_time": "2018-06-20T21:22:45",
                "url": "https://files.pythonhosted.org/packages/5c/56/31a983a5f11bdc1eeef090e17f20a206ffa4efd6a4bf8e9ea1b130ef2571/vimpy-0.0.10-py2-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "f390203a8654a098917af25f06d69a32",
                    "sha256": "8335bd5fdea1e93210ac12f4a86183be73e59c5c912b0489c13d85949864e705"
                },
                "downloads": -1,
                "filename": "vimpy-0.0.10.tar.gz",
                "has_sig": false,
                "md5_digest": "f390203a8654a098917af25f06d69a32",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 4325,
                "upload_time": "2018-06-20T21:22:46",
                "url": "https://files.pythonhosted.org/packages/37/39/4ef83238c7228d46c57adde85ada9db13396159a1b74960bf2655ad8e431/vimpy-0.0.10.tar.gz"
            }
        ],
        "0.0.3": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "fca093e553a585a052d423ef4993b15b",
                    "sha256": "5600873157a1ae96136cdfe50e0880f97395ded8887e48ef3bc2ed5379f034e4"
                },
                "downloads": -1,
                "filename": "vimpy-0.0.3-py2-none-any.whl",
                "has_sig": false,
                "md5_digest": "fca093e553a585a052d423ef4993b15b",
                "packagetype": "bdist_wheel",
                "python_version": "py2",
                "requires_python": null,
                "size": 1636,
                "upload_time": "2018-06-15T20:10:37",
                "url": "https://files.pythonhosted.org/packages/73/2e/839065471ab96309238fe8168da12427771faf0b17f1da10a171f43e9553/vimpy-0.0.3-py2-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "aa48b742fd06509f203b8f195abe50b9",
                    "sha256": "e8a883204bfd8613947701af5bd2adda80221c7ada70615cd62db0fe255c85ce"
                },
                "downloads": -1,
                "filename": "vimpy-0.0.3.tar.gz",
                "has_sig": false,
                "md5_digest": "aa48b742fd06509f203b8f195abe50b9",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 1491,
                "upload_time": "2018-06-15T20:10:38",
                "url": "https://files.pythonhosted.org/packages/b7/73/3da61bd95bb489f7f163df595f58ef2f12968d1c98f7286ceabf99f7ef92/vimpy-0.0.3.tar.gz"
            }
        ],
        "0.0.4": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "226fa9655736c627da185600d62d97f2",
                    "sha256": "c83844fa6b3ab86bf83b53c4b62e4db211b79b43872a154f892f8c20b4d5ca22"
                },
                "downloads": -1,
                "filename": "vimpy-0.0.4-py2-none-any.whl",
                "has_sig": false,
                "md5_digest": "226fa9655736c627da185600d62d97f2",
                "packagetype": "bdist_wheel",
                "python_version": "py2",
                "requires_python": null,
                "size": 2665,
                "upload_time": "2018-06-15T20:36:43",
                "url": "https://files.pythonhosted.org/packages/5c/89/8fc9384b3c7047fa58e5e891bf5141104388a9391b7ab74b064091f81e30/vimpy-0.0.4-py2-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "fad96b249ea322d96fadbdeab18e37d8",
                    "sha256": "307b131d8091538769f8e17acdebd6912c5c21eac695c87aebd8e881ef22bb73"
                },
                "downloads": -1,
                "filename": "vimpy-0.0.4.tar.gz",
                "has_sig": false,
                "md5_digest": "fad96b249ea322d96fadbdeab18e37d8",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 2828,
                "upload_time": "2018-06-15T20:36:44",
                "url": "https://files.pythonhosted.org/packages/b8/16/88a7847f8c66b4949517b2667197c79768c0e50f0acae08e9cc07b84b0d4/vimpy-0.0.4.tar.gz"
            }
        ],
        "0.0.5": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "75db7c9e963f42cb63c812e2d26dca36",
                    "sha256": "835023c5c9cfd2423f4dc4b0fdb06733066a8f1e2f2f4e15bddaefd62aabbb07"
                },
                "downloads": -1,
                "filename": "vimpy-0.0.5-py2-none-any.whl",
                "has_sig": false,
                "md5_digest": "75db7c9e963f42cb63c812e2d26dca36",
                "packagetype": "bdist_wheel",
                "python_version": "py2",
                "requires_python": null,
                "size": 2665,
                "upload_time": "2018-06-19T17:58:12",
                "url": "https://files.pythonhosted.org/packages/88/41/7264fcf02ac301b7f31b4b5891ca6e065ca1a71172475b95feb406a4b58a/vimpy-0.0.5-py2-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "50bdbd1093f88fe1256f84c01995aa15",
                    "sha256": "8f942ec8ce302264a74dde876c17f52bbff06a14fcdb0b745fafe91b472c4aa1"
                },
                "downloads": -1,
                "filename": "vimpy-0.0.5.tar.gz",
                "has_sig": false,
                "md5_digest": "50bdbd1093f88fe1256f84c01995aa15",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 2830,
                "upload_time": "2018-06-19T17:58:13",
                "url": "https://files.pythonhosted.org/packages/5e/4b/186aba5e20e133055108d44618699356cddd0cd8d95773c537a92b84c653/vimpy-0.0.5.tar.gz"
            }
        ],
        "0.0.9": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "0078e489c0f93087b45939862795f24a",
                    "sha256": "028004c63eacbd78145f04b4e0825fd66bd269944461db65002933c408250442"
                },
                "downloads": -1,
                "filename": "vimpy-0.0.9-py2-none-any.whl",
                "has_sig": false,
                "md5_digest": "0078e489c0f93087b45939862795f24a",
                "packagetype": "bdist_wheel",
                "python_version": "py2",
                "requires_python": null,
                "size": 5323,
                "upload_time": "2018-06-20T20:29:57",
                "url": "https://files.pythonhosted.org/packages/ba/25/62bd194731a06bbeff8762823e93c20c120920ea2075d87f348a46a3b2c3/vimpy-0.0.9-py2-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "a92e7d7def571e17c6afafe2a0331f18",
                    "sha256": "6cfd437c7584e01c2cfdd2424aea9aa687b32b2d261cde2be4d6753003ec1e37"
                },
                "downloads": -1,
                "filename": "vimpy-0.0.9.tar.gz",
                "has_sig": false,
                "md5_digest": "a92e7d7def571e17c6afafe2a0331f18",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 4324,
                "upload_time": "2018-06-20T20:29:58",
                "url": "https://files.pythonhosted.org/packages/d3/52/9acc45faad3cd65aaf3a27ecc6c4d58eea67917f004b2237b090f3dcb451/vimpy-0.0.9.tar.gz"
            }
        ],
        "1.0.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "6c5983826249c920ee58520543c576b9",
                    "sha256": "040e5bf5c1782a15daf3c62ab895b892e2497aef101b8e5393a5b86d4922c535"
                },
                "downloads": -1,
                "filename": "vimpy-1.0.0-py2-none-any.whl",
                "has_sig": false,
                "md5_digest": "6c5983826249c920ee58520543c576b9",
                "packagetype": "bdist_wheel",
                "python_version": "py2",
                "requires_python": null,
                "size": 6880,
                "upload_time": "2018-10-28T17:37:26",
                "url": "https://files.pythonhosted.org/packages/38/46/3ddcfa54ce80f05f91d8c2cfb057a05df573e0a930578b566f365786b9e7/vimpy-1.0.0-py2-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "0486431ffdae309b8cbea51a3f343f9a",
                    "sha256": "340989839b1e648d6f7b3c2301ab8263b9465fe5267b3f3b9c981d143fd9b124"
                },
                "downloads": -1,
                "filename": "vimpy-1.0.0.tar.gz",
                "has_sig": false,
                "md5_digest": "0486431ffdae309b8cbea51a3f343f9a",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 4970,
                "upload_time": "2018-10-28T17:37:28",
                "url": "https://files.pythonhosted.org/packages/be/c1/73654fa26d68f3b3cd7e2aa6a4932e7b4a04b8208b0e0eb5d8b9de9d5a70/vimpy-1.0.0.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "6c5983826249c920ee58520543c576b9",
                "sha256": "040e5bf5c1782a15daf3c62ab895b892e2497aef101b8e5393a5b86d4922c535"
            },
            "downloads": -1,
            "filename": "vimpy-1.0.0-py2-none-any.whl",
            "has_sig": false,
            "md5_digest": "6c5983826249c920ee58520543c576b9",
            "packagetype": "bdist_wheel",
            "python_version": "py2",
            "requires_python": null,
            "size": 6880,
            "upload_time": "2018-10-28T17:37:26",
            "url": "https://files.pythonhosted.org/packages/38/46/3ddcfa54ce80f05f91d8c2cfb057a05df573e0a930578b566f365786b9e7/vimpy-1.0.0-py2-none-any.whl"
        },
        {
            "comment_text": "",
            "digests": {
                "md5": "0486431ffdae309b8cbea51a3f343f9a",
                "sha256": "340989839b1e648d6f7b3c2301ab8263b9465fe5267b3f3b9c981d143fd9b124"
            },
            "downloads": -1,
            "filename": "vimpy-1.0.0.tar.gz",
            "has_sig": false,
            "md5_digest": "0486431ffdae309b8cbea51a3f343f9a",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 4970,
            "upload_time": "2018-10-28T17:37:28",
            "url": "https://files.pythonhosted.org/packages/be/c1/73654fa26d68f3b3cd7e2aa6a4932e7b4a04b8208b0e0eb5d8b9de9d5a70/vimpy-1.0.0.tar.gz"
        }
    ]
}