{
    "info": {
        "author": "Nikolay Novik",
        "author_email": "nickolainovik@gmail.com",
        "bugtrack_url": null,
        "classifiers": [
            "Development Status :: 2 - Pre-Alpha",
            "Framework :: AsyncIO",
            "Intended Audience :: Developers",
            "License :: OSI Approved :: Apache Software License",
            "Operating System :: POSIX",
            "Programming Language :: Python :: 3",
            "Programming Language :: Python :: 3.6",
            "Programming Language :: Python :: 3.7"
        ],
        "description": "mldissect\n=========\n.. image:: https://travis-ci.com/ml-libs/mldissect.svg?branch=master\n    :target: https://travis-ci.com/ml-libs/mldissect\n.. image:: https://codecov.io/gh/ml-libs/mldissect/branch/master/graph/badge.svg\n    :target: https://codecov.io/gh/ml-libs/mldissect \n.. image:: https://api.codeclimate.com/v1/badges/bc29bc214f39b54ef30a/maintainability\n   :target: https://codeclimate.com/github/ml-libs/mldissect/maintainability\n   :alt: Maintainability\n\n\n**mldissect** is model agnostic predictions explainer, library can show\ncontribution of each feature of your prediction value.\n\nFeatures\n========\n* Supports predictions explanations for classification and regression\n* Easy to use API.\n* Works with ``pandas`` and ``numpy``\n\n\nInstallation\n------------\nInstallation process is simple, just::\n\n    $ pip install mldissect\n\nBasic Usage\n===========\n\n.. code:: python\n\n    # lets train a model\n    boston = load_boston()\n    columns = list(boston.feature_names)\n    X, y = boston['data'], boston['target']\n    X_train, X_test, y_train, y_test = train_test_split(\n        X, y, test_size=.2, random_state=seed\n    )\n\n    clf = LassoCV()\n    clf.fit(X_train, y_train)\n\n    # select first observation in test split\n    observation = X_test[0]\n    # RegressionExplainer uses training data or sample of training data\n    # for large dataset to figure out contributions of each feature\n    explainer = RegressionExplainer(clf, X_train, columns)\n    result = explainer.explain(observation)\n    # print/visualize explanation\n    explanation = Explanation(result)\n    explanation.print()\n\n\nresult::\n\n    +----------+---------+--------------------+\n    | Feature  | Value   | Contribution       |\n    +----------+---------+--------------------+\n    | baseline | -       | 22.611881188118804 |\n    | LSTAT    | 7.34    | 3.6872             |\n    | PTRATIO  | 16.9    | 1.3652             |\n    | CRIM     | 0.06724 | 0.2323             |\n    | B        | 375.21  | 0.1195             |\n    | RM       | 6.333   | 0.0411             |\n    | INDUS    | 3.24    | 0.0312             |\n    | CHAS     | 0.0     | 0.0                |\n    | NOX      | 0.46    | 0.0                |\n    | TAX      | 430.0   | -0.3794            |\n    | AGE      | 17.2    | -0.5127            |\n    | ZN       | 0.0     | -0.6143            |\n    | DIS      | 5.2146  | -1.0792            |\n    | RAD      | 4.0     | -1.0993            |\n    +----------+---------+--------------------+\n\n\nAlgorithm\n=========\nAlgorithm is based on ideas describe in paper *\"Explanations of model predictions\nwith live and breakDown packages\"* https://arxiv.org/abs/1804.01955\n\n\nDifference with pyBreakDown\n===========================\n``pyBreakDown`` is similar project, but there is key differences:\n\n* `mldissect` is maintained\n* Has tests and good code coverage.\n* Classification is working properly.\n* Multi class support.\n* Top down approach is not implemented.\n* Friendly license.\n\n\nRequirements\n------------\n\n* Python_ 3.6+\n* numpy_\n\n.. _Python: https://www.python.org\n.. _numpy: http://www.numpy.org/\n\nCHANGES\n=======",
        "description_content_type": "",
        "docs_url": null,
        "download_url": "https://pypi.python.org/pypi/mldissect",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/ml-libs/mldissect",
        "keywords": "mldissect",
        "license": "Apache 2",
        "maintainer": "",
        "maintainer_email": "",
        "name": "mldissect",
        "package_url": "https://pypi.org/project/mldissect/",
        "platform": "POSIX",
        "project_url": "https://pypi.org/project/mldissect/",
        "project_urls": {
            "Download": "https://pypi.python.org/pypi/mldissect",
            "Homepage": "https://github.com/ml-libs/mldissect"
        },
        "release_url": "https://pypi.org/project/mldissect/0.0.1a2/",
        "requires_dist": null,
        "requires_python": "",
        "summary": "mldissect - model agnostic explanations",
        "version": "0.0.1a2"
    },
    "last_serial": 4450183,
    "releases": {
        "0.0.1a0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "230a577ef07f89f5ec61e07863b92032",
                    "sha256": "35a49690289cc57a510dd5ba3663953edb367410f163d01375e46d9aba8fdfb8"
                },
                "downloads": -1,
                "filename": "mldissect-0.0.1a0.tar.gz",
                "has_sig": false,
                "md5_digest": "230a577ef07f89f5ec61e07863b92032",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 8632,
                "upload_time": "2018-09-26T21:23:28",
                "url": "https://files.pythonhosted.org/packages/2c/cc/2ff03c7496399f714442e83d3a6af40c065c12864dd3f5661e7131b5d656/mldissect-0.0.1a0.tar.gz"
            }
        ],
        "0.0.1a2": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "78192f0f66780678a99303b1de86a435",
                    "sha256": "76076b4cbe86a09c9e038785b62ec6ca18682fee9f2fccbc39711a09a1da7d86"
                },
                "downloads": -1,
                "filename": "mldissect-0.0.1a2.tar.gz",
                "has_sig": false,
                "md5_digest": "78192f0f66780678a99303b1de86a435",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 8674,
                "upload_time": "2018-11-04T17:23:47",
                "url": "https://files.pythonhosted.org/packages/f0/17/0a2c31d9a6d70c6347c3f254fa757fe5b60944dc19bb9af82b18d033ffd0/mldissect-0.0.1a2.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "78192f0f66780678a99303b1de86a435",
                "sha256": "76076b4cbe86a09c9e038785b62ec6ca18682fee9f2fccbc39711a09a1da7d86"
            },
            "downloads": -1,
            "filename": "mldissect-0.0.1a2.tar.gz",
            "has_sig": false,
            "md5_digest": "78192f0f66780678a99303b1de86a435",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 8674,
            "upload_time": "2018-11-04T17:23:47",
            "url": "https://files.pythonhosted.org/packages/f0/17/0a2c31d9a6d70c6347c3f254fa757fe5b60944dc19bb9af82b18d033ffd0/mldissect-0.0.1a2.tar.gz"
        }
    ]
}