{
    "info": {
        "author": "Leandro Ferrado",
        "author_email": "ljferrado@gmail.com",
        "bugtrack_url": null,
        "classifiers": [],
        "description": "# pyoselm\n\n*A Python implementation of Online Sequential Extreme Machine Learning (OS-ELM) for online machine learning*\n\n\n[![Build Status](https://travis-ci.org/leferrad/pyoselm.svg?branch=master)](https://travis-ci.org/leferrad/pyoselm)\n\n### Dependencies\n\n- Numpy\n- Scipy\n- Scikit-learn\n\nOriginal publication: \n\n> Huang, G. B., Liang, N. Y., Rong, H. J., Saratchandran, P., & Sundararajan, N. (2005). \n  On-Line Sequential Extreme Learning Machine. Computational Intelligence, 2005, 232-237.\n\nLink: https://pdfs.semanticscholar.org/2ebd/fa3852e4ad68a0cfde9f0f69b95953d69178.pdf\n\nImplementation strongly based on the following repos:\n\n- https://github.com/ExtremeLearningMachines/ELM-MATLAB-and-Online.Sequential.ELM\n- https://github.com/dclambert/Python-ELM\n\n### Usage\n\n```python\nfrom pyoselm import OSELMRegressor, OSELMClassifier\nfrom sklearn.datasets import load_digits, make_regression\nimport random\n\n# --- Regression problem ---\n# Model\noselmr = OSELMRegressor(n_hidden=20, activation_func='tanh')\n# Data\nx, y = make_regression(n_samples=400, n_targets=1, n_features=10)\nn_batch = 20\n\n# Fit model with chunks of data\nfor i in range(20):\n    x_batch = x[i*n_batch:(i+1)*n_batch]\n    y_batch = y[i*n_batch:(i+1)*n_batch]\n\n    oselmr.fit(x_batch, y_batch)\n    print(\"Train score for batch %i: %s\" % (i+1, str(oselmr.score(x_batch, y_batch))))\n\n# Results\nprint(\"Train score of total: %s\" % str(oselmr.score(x, y)))\n\n# --- Classification problem ---\n# Model \noselmc = OSELMClassifier(n_hidden=20, activation_func='sigmoid')\n# Data\nx, y = load_digits(n_class=10, return_X_y=True)\n\n# Shuffle data (to have batches with more than one class)\nzip_x_y = zip(x, y)\nrandom.shuffle(zip_x_y)\nx, y = [x_y[0] for x_y in zip_x_y], [x_y[1] for x_y in zip_x_y]\n\n# Sequential learning\n# The first batch of data should have the same size as neurons in the model to achieve the 1st phase (boosting)\nbatches_x = [x[:oselmc.n_hidden]] + [[x_i] for x_i in x[oselmc.n_hidden:]]\nbatches_y = [y[:oselmc.n_hidden]] + [[y_i] for y_i in y[oselmc.n_hidden:]]\n\nfor b_x, b_y in zip(batches_x, batches_y):\n    oselmc.fit(b_x, b_y)\n\nprint(\"Train score of total: %s\" % str(oselmc.score(x, y)))\n\n```\n\nNOTE: Chuck-by-chunk is faster than one-by-one",
        "description_content_type": "",
        "docs_url": null,
        "download_url": "",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/leferrad/pyoselm",
        "keywords": "",
        "license": "LICENSE",
        "maintainer": "",
        "maintainer_email": "",
        "name": "pyoselm",
        "package_url": "https://pypi.org/project/pyoselm/",
        "platform": "",
        "project_url": "https://pypi.org/project/pyoselm/",
        "project_urls": {
            "Homepage": "https://github.com/leferrad/pyoselm"
        },
        "release_url": "https://pypi.org/project/pyoselm/0.1.0/",
        "requires_dist": null,
        "requires_python": "",
        "summary": "A Python implementation of Online Sequential Extreme Machine Learning (OS-ELM) for online machine learning",
        "version": "0.1.0"
    },
    "last_serial": 3702167,
    "releases": {
        "0.1.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "d1cee8cb067e31dfd7b77d2eef8677a4",
                    "sha256": "ad08801597bb33707eb4b8fce5a6efe4649a8dcb816306c61521368c77f518cd"
                },
                "downloads": -1,
                "filename": "pyoselm-0.1.0.tar.gz",
                "has_sig": false,
                "md5_digest": "d1cee8cb067e31dfd7b77d2eef8677a4",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 11967,
                "upload_time": "2018-03-24T19:37:56",
                "url": "https://files.pythonhosted.org/packages/c1/49/9a2b677e0796a12994b8f6a8702e69b34147ca683519417df54a8ff2d839/pyoselm-0.1.0.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "d1cee8cb067e31dfd7b77d2eef8677a4",
                "sha256": "ad08801597bb33707eb4b8fce5a6efe4649a8dcb816306c61521368c77f518cd"
            },
            "downloads": -1,
            "filename": "pyoselm-0.1.0.tar.gz",
            "has_sig": false,
            "md5_digest": "d1cee8cb067e31dfd7b77d2eef8677a4",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 11967,
            "upload_time": "2018-03-24T19:37:56",
            "url": "https://files.pythonhosted.org/packages/c1/49/9a2b677e0796a12994b8f6a8702e69b34147ca683519417df54a8ff2d839/pyoselm-0.1.0.tar.gz"
        }
    ]
}