{
    "info": {
        "author": "Tang Han",
        "author_email": "aloofness54@gmail.com",
        "bugtrack_url": null,
        "classifiers": [
            "License :: OSI Approved :: MIT License",
            "Operating System :: OS Independent",
            "Programming Language :: Python :: 3"
        ],
        "description": "# auto-modelling\n\nAuto-modelling is a convenient library to train and tune machine models automatically.\n\nIts main features include the following:\n\n1. preprocessing columns in all datatypes. (numeric, categorical, text)\n2. train machine models and tune parameters automatically.\n3. return top n best models with optimized parameters.\n4. Apply **stacking** technique to combine the n best models returned by the repo or self-determined fitted models together to get an even better result.\n\nThe machine learning models include the following:\n- Classification:\n    - ExtraTreesClassifier\n    - RandomForestClassifier\n    - KNeighborsClassifier\n    - LogisticRegression\n    - XGBClassifier\n- Regression:\n    - ExtraTreesRegressor\n    - GradientBoostingRegressor\n    - AdaBoostRegressor\n    - DecisionTreeRegressor\n    - RandomForestRegressor\n    - XGBRegressor\n- Stack:\n    - for classify: LogisticRegression\n    - for regression: LinearRegression\n\nreference: https://github.com/EpistasisLab/tpot/blob\n\n# Installation\n\n`pip install auto-modelling`\n\n# Usage Example\n```\nfrom auto_modelling.classification import GoClassify\nfrom auto_modelling.regression import GoRegress\nfrom auto_modelling.preprocess import DataManager\nfrom auto_modelling.stack import Stack\n\n# preprocessing data\ndm = DataManager(directory = 'preprocess_tools')\ntrain, test = dm.drop_sparse_columns(x_train, x_test)\ntrain, test = dm.process_data(x_train, x_test)\n# the encoders are stored in the directory called data_process_tools.\n\n# use the same processing tools to process new data\npredict_data = dm.process_predict_data(predict_x)\n# predict_x should have the same format as x_train/x_test\n\n# classification\nclf = GoClassify(n_best=1)\nbest = clf.train(x_train, y_train)\ny_pred = best.predict(x_test)\n\n# regression\nreg = GoRegress(n_best=1)\nbest = reg.train(x_train, y_train)\ny_pred = best.predict(x_test)\n\n# get top 3 best models\nclf = GoClassify(n_best=3)\nbests = clf.train(x_train, y_train)\ny_preds = [m.predict(x_test) for m in bests]\n\n# Stack top 3 best models\nstack = Stack(n_models = 3)\nlevel_0_models, level_1_model = stack.train(x_train, y_train, x_test, y_test)\n```\n\nThere are examples `test.py` and `sample.py` in the root directory of this package. run\n`python test.py`/`python sample.py`.\n\n# Development Guide\n\n- Clone the repo\n\n- Create the virtual environment\n```\nmkvirtualenv auto\nworkon auto\npip install requirements.txt\n```\nif you have issues in installing `xgboost` \nrefrence: \nhttps://xgboost.readthedocs.io/en/latest/build.html#\nhttps://www.ibm.com/developerworks/community/blogs/jfp/entry/Installing_XGBoost_on_Mac_OSX?lang=en\n\n# Note\n\n- TO DO: Feature selection, evaluation metricss\n\n# Thoughts\n\n- Ideally, any dataframe being throw into this repo, it should be processed.\n\n1. pre-processing \n\n    - drop column that have too many null(Done)\n    - fill na for both numeric and non-numeric values(Done)\n    - encoded for non-numeric values(Done)\n    - scale values if needed\n    - balance the dataset if needed\n\n2. model-training\n\n    - mode = `classification`, `regression`, `auto`(Done)\n    - split data-set\n    - tuning parameters and model selection (Done)\n    - feature selection\n    - return a model with parameters, columns and a script to process x_test(Done)\n    - stacking with customized fitted models (Done)\n\n3. model-evualation\n# Other reference\n\n[Packaging your project](https://packaging.python.org/tutorials/packaging-projects/)\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/TangHan54/auto-modelling.git",
        "keywords": "",
        "license": "",
        "maintainer": "",
        "maintainer_email": "",
        "name": "auto-modelling",
        "package_url": "https://pypi.org/project/auto-modelling/",
        "platform": "",
        "project_url": "https://pypi.org/project/auto-modelling/",
        "project_urls": {
            "Homepage": "https://github.com/TangHan54/auto-modelling.git"
        },
        "release_url": "https://pypi.org/project/auto-modelling/1.2.5/",
        "requires_dist": [
            "sklearn",
            "ipython",
            "numpy",
            "pandas",
            "setuptools",
            "xgboost",
            "joblib"
        ],
        "requires_python": "",
        "summary": "A light package for automatic model tuning and stacking",
        "version": "1.2.5"
    },
    "last_serial": 5713626,
    "releases": {
        "1.2.5": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "627b3f96650826304729de44585ea865",
                    "sha256": "16b9644ca661101ff41dd94417ec57c6d623d7945c6ca2827d06ea938f1e3e3a"
                },
                "downloads": -1,
                "filename": "auto_modelling-1.2.5-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "627b3f96650826304729de44585ea865",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": null,
                "size": 12297,
                "upload_time": "2019-08-22T07:06:39",
                "url": "https://files.pythonhosted.org/packages/db/b2/3369d9649e7f38f67069067d5139c5e15c787217355bffa08ebe6a619c24/auto_modelling-1.2.5-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "2ed2a9f07d8a3ba02f42481dc4c164a3",
                    "sha256": "bb228f0548eea4254f08ada4458765dcea2b92c630796f51ce4324fe089bada3"
                },
                "downloads": -1,
                "filename": "auto_modelling-1.2.5.tar.gz",
                "has_sig": false,
                "md5_digest": "2ed2a9f07d8a3ba02f42481dc4c164a3",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 6151,
                "upload_time": "2019-08-22T07:06:41",
                "url": "https://files.pythonhosted.org/packages/62/ab/fde8ec954cc70791744e29c310ba2a914c9c173b7b2bd904cbcbe963ec39/auto_modelling-1.2.5.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "627b3f96650826304729de44585ea865",
                "sha256": "16b9644ca661101ff41dd94417ec57c6d623d7945c6ca2827d06ea938f1e3e3a"
            },
            "downloads": -1,
            "filename": "auto_modelling-1.2.5-py3-none-any.whl",
            "has_sig": false,
            "md5_digest": "627b3f96650826304729de44585ea865",
            "packagetype": "bdist_wheel",
            "python_version": "py3",
            "requires_python": null,
            "size": 12297,
            "upload_time": "2019-08-22T07:06:39",
            "url": "https://files.pythonhosted.org/packages/db/b2/3369d9649e7f38f67069067d5139c5e15c787217355bffa08ebe6a619c24/auto_modelling-1.2.5-py3-none-any.whl"
        },
        {
            "comment_text": "",
            "digests": {
                "md5": "2ed2a9f07d8a3ba02f42481dc4c164a3",
                "sha256": "bb228f0548eea4254f08ada4458765dcea2b92c630796f51ce4324fe089bada3"
            },
            "downloads": -1,
            "filename": "auto_modelling-1.2.5.tar.gz",
            "has_sig": false,
            "md5_digest": "2ed2a9f07d8a3ba02f42481dc4c164a3",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 6151,
            "upload_time": "2019-08-22T07:06:41",
            "url": "https://files.pythonhosted.org/packages/62/ab/fde8ec954cc70791744e29c310ba2a914c9c173b7b2bd904cbcbe963ec39/auto_modelling-1.2.5.tar.gz"
        }
    ]
}