corona_whitebox.ipynb

{
    "cells": [
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {},
            "outputs": [],
            "source": "!pip install tensorflow==2.2.0rc1"
        },
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {},
            "outputs": [],
            "source": "import tensorflow as tf\nimport pandas as pd\ntf.__version__"
        },
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {},
            "outputs": [],
            "source": "# The code was removed by Watson Studio for sharing."
        },
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {},
            "outputs": [],
            "source": "import types\nimport pandas as pd\nfrom botocore.client import Config\nimport ibm_boto3\n\ndef __iter__(self): return 0\n\n\nclient = ibm_boto3.client(service_name='s3',\n    ibm_api_key_id=credentials_1['IBM_API_KEY_ID'],\n    ibm_auth_endpoint=credentials_1['IBM_AUTH_ENDPOINT'],\n    config=Config(signature_version='oauth'),\n    endpoint_url=credentials_1['ENDPOINT'])\nbucket=credentials_1['BUCKET']"
        },
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {},
            "outputs": [],
            "source": "client.download_file(Bucket=bucket,Key='cases_features.csv',Filename='cases_features.csv')"
        },
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {},
            "outputs": [],
            "source": "df = pd.read_csv('cases_features.csv')\ndf = df[df['countriesAndTerritories']=='United_States_of_America']\nwindow_size = 10"
        },
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {},
            "outputs": [],
            "source": "import tensorflow as tf\nimport numpy as np\n\nnp.random.seed(seed=42)\n\[email protected]\ndef model(x, beta, gamma):\n    return x*(gamma*((beta/gamma)-1))\n    \n\n\na = tf.Variable(np.random.randn()*10, trainable=True, dtype=tf.float64)\nb = tf.Variable(np.random.randn()*10, trainable=True, dtype=tf.float64)\nc = tf.Variable(np.random.randn()*10, trainable=True, dtype=tf.float64)\nd = tf.Variable(np.random.randn()*10, trainable=True, dtype=tf.float64)\n\nx = tf.constant(df['cases_'+str(window_size)+'_day_average'], dtype=tf.float64)\ny = tf.constant(df.cases, dtype=tf.float64)\n\n\n\nvariables = [a,b]\noptimizer = tf.optimizers.Adam(0.01)\n\nloss_object = tf.keras.losses.MeanAbsoluteError()\n\nlosses = []\nepochs = 10000\n\nwith tf.GradientTape(persistent=True) as tape:\n    for i in range(epochs):\n        y_pred = model(x,a,b)\n        loss = loss_object(y,y_pred)\n        losses.append(loss.numpy())\n        grads = tape.gradient(loss, variables)\n        optimizer.apply_gradients(zip(grads, variables))"
        },
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {},
            "outputs": [],
            "source": "print(losses[0])\nprint(losses[-1])"
        },
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {},
            "outputs": [],
            "source": "import seaborn as sns\nsns.lineplot(x=range(epochs), y=losses)"
        },
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {},
            "outputs": [],
            "source": "a/b"
        },
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {},
            "outputs": [],
            "source": "predictions = []\ndef recursive_model(i,x,predictions):\n    if i <= len(df):\n        x = model(x,a,b)\n        predictions.append(x)\n        i = i+1\n        recursive_model(i,x,predictions)\nrecursive_model(1,100,predictions)"
        },
        {
            "cell_type": "code",
            "execution_count": null,
            "metadata": {},
            "outputs": [],
            "source": "%matplotlib inline\nimport matplotlib.pyplot as plt\nindex = np.array(range(len(df)))\nplt.plot(index, df['percentage_died'], label = \"reality\")\nplt.plot(index, predictions, label = \"prediction\")\nplt.legend()\nplt.show()"
        }
    ],
    "metadata": {
        "kernelspec": {
            "display_name": "Python 3.6",
            "language": "python",
            "name": "python3"
        },
        "language_info": {
            "codemirror_mode": {
                "name": "ipython",
                "version": 3
            },
            "file_extension": ".py",
            "mimetype": "text/x-python",
            "name": "python",
            "nbconvert_exporter": "python",
            "pygments_lexer": "ipython3",
            "version": "3.6.9"
        }
    },
    "nbformat": 4,
    "nbformat_minor": 4
}