{
    "info": {
        "author": "zhaofeng-shu33",
        "author_email": "616545598@qq.com",
        "bugtrack_url": null,
        "classifiers": [
            "Development Status :: 4 - Beta",
            "Operating System :: OS Independent",
            "Programming Language :: Python :: 2.7",
            "Programming Language :: Python :: 3.6",
            "Programming Language :: Python :: 3.7"
        ],
        "description": "# ace_nn\n[![Build Status](https://travis-ci.org/zhaofeng-shu33/ace_nn.svg?branch=master)](https://travis-ci.org/zhaofeng-shu33/ace_nn)\n\n## Introduction\n\nThis repo contains experimental implementation of ace algorithm via neural network. It is shown by **xiangxiang-xu** that calculating optimal features by *Alternating Conditional Expectation* is equivalent to maximize *H-score*. \n\n## How to run\n\nThree examples are provided ( one for continuous variable and the other twos are for discrete variable) and their results are the same as `ace`. The main function is `ace_nn` and its parameters are very similar to [`ace_cream`](https://github.com/zhaofeng-shu33/ace_cream). \n\n```python\nimport numpy as np\nfrom ace_nn import ace_nn\n# discrete case, binary symmetric channel with crossover probability 0.1\nN_SIZE = 1000\nx = np.random.choice([0,1], size=N_SIZE)\nn = np.random.choice([0,1], size=N_SIZE, p=[0.9, 0.1])\ny = np.mod(x + n, 2)\n# set both x(cat=0) and y(cat=-1) as categorical type\ntx, ty = ace_nn(x, y, cat=[-1,0], epochs=100)\n\n# continuous case\nx = np.random.uniform(0, np.pi, 200)\ny = np.exp(np.sin(x)+np.random.normal(size=200)/2)\ntx, ty = ace_nn(x, y)\n```\n\nFor more detail, run `help(ace_nn)` to see the parameters and returns of this function.\n\n## Further discussion\n\nCurrently, the neural networks used to approximate optimal $f(x)$ and $g(y)$ are two-layer MLP with `tanh` as activation function. More turns of epochs are needed for large alphabet  $|\\mathcal{X}|$ and  $|\\mathcal{Y}|$ and the running time is not short.\n\nAlso, `batch_size` and `hidden_units_num` can be hypertuned, and there is no guarantee that current configuration of neural network is optimal for solving ace.\n\n## Application\nwe can use `ace_nn(x, y, return_hscore = True)` to calculate a lower bound of $\\frac{\\norm{B}_F^2}{2}$",
        "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/zhaofeng-shu33/ace_nn",
        "keywords": "",
        "license": "Apache License Version 2.0",
        "maintainer": "zhaofeng-shu33",
        "maintainer_email": "616545598@qq.com",
        "name": "ace-nn",
        "package_url": "https://pypi.org/project/ace-nn/",
        "platform": "",
        "project_url": "https://pypi.org/project/ace-nn/",
        "project_urls": {
            "Homepage": "https://github.com/zhaofeng-shu33/ace_nn"
        },
        "release_url": "https://pypi.org/project/ace-nn/0.2/",
        "requires_dist": null,
        "requires_python": "",
        "summary": "Alternating Conditional Exceptation with Neural Network",
        "version": "0.2"
    },
    "last_serial": 4883796,
    "releases": {
        "0.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "5340b179c82ee26c5fe384289669c511",
                    "sha256": "1ac0af77e8709d5ff5d5af905f318c471aca4ddb2826a5024f587e8def270ba9"
                },
                "downloads": -1,
                "filename": "ace_nn-0.1.tar.gz",
                "has_sig": false,
                "md5_digest": "5340b179c82ee26c5fe384289669c511",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 3799,
                "upload_time": "2019-03-01T11:28:23",
                "url": "https://files.pythonhosted.org/packages/d7/31/2157850a793a2842f72c10fb12c6a8e4293a5280475e41c6647bb02cf18f/ace_nn-0.1.tar.gz"
            }
        ],
        "0.2": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "22d9b1bdab87e39d0f525baf9c204755",
                    "sha256": "c51c7c2b0853c6945173def0170eb93b28275ffee8de81ee66ef6c7ea338f519"
                },
                "downloads": -1,
                "filename": "ace_nn-0.2.tar.gz",
                "has_sig": false,
                "md5_digest": "22d9b1bdab87e39d0f525baf9c204755",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 3793,
                "upload_time": "2019-03-01T11:31:13",
                "url": "https://files.pythonhosted.org/packages/b1/cc/69418eba970547c4fed55a8e7bfdd0ec83dd5646a3a59db8e1cf35c172f8/ace_nn-0.2.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "22d9b1bdab87e39d0f525baf9c204755",
                "sha256": "c51c7c2b0853c6945173def0170eb93b28275ffee8de81ee66ef6c7ea338f519"
            },
            "downloads": -1,
            "filename": "ace_nn-0.2.tar.gz",
            "has_sig": false,
            "md5_digest": "22d9b1bdab87e39d0f525baf9c204755",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 3793,
            "upload_time": "2019-03-01T11:31:13",
            "url": "https://files.pythonhosted.org/packages/b1/cc/69418eba970547c4fed55a8e7bfdd0ec83dd5646a3a59db8e1cf35c172f8/ace_nn-0.2.tar.gz"
        }
    ]
}