Kfold_.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.1"
    },
    "colab": {
      "name": "Kfold_.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/SinKasula/MachineLearning/blob/master/Kfold_.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "slsbblXD6Unj",
        "colab_type": "code",
        "outputId": "1335ae41-a5c7-46de-da96-9e9dc53113e8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 122
        }
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive/')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3Aietf%3Awg%3Aoauth%3A2.0%3Aoob&scope=email%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdocs.test%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive.photos.readonly%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fpeopleapi.readonly&response_type=code\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive/\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rSxslrKGLAk-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "from datetime import datetime\n",
        "import nn\n",
        "from copy import deepcopy\n",
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "import sys\n",
        "sys.path.append('drive/My Drive/ML_Assign3/')\n",
        "import nn"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pftY004TLAlD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "path = r\"drive/My Drive/ML_Assign3/\"\n",
        "df = pd.read_csv(path+'/IN_Metro_Traffic.csv')\n",
        "df = df.reindex(np.random.permutation(df.index))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xcfbh04icabq",
        "colab_type": "code",
        "outputId": "255575bd-40ac-426c-dea8-c5a5bb9816ae",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "len(df)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "33750"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "O_BNipIYLAlH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def isHolidayBool(x):\n",
        "    if x== 'None':\n",
        "        return 0\n",
        "    else:\n",
        "        return 1\n",
        "    \n",
        "df.is_holiday= df.is_holiday.apply(lambda x : isHolidayBool(x))\n",
        "\n",
        "\n",
        "weather_type_dic = {'Clouds':1, 'Clear':0, 'Mist':2, 'Rain':5, 'Snow':10, 'Drizzle':6, 'Haze':3,\n",
        "                    'Fog':4, 'Thunderstorm':9, 'Smoke':7, 'Squall':8}\n",
        "\n",
        "df.weather_type = df.weather_type.apply(lambda x : weather_type_dic[x])\n",
        "\n",
        "\n",
        "\n",
        "def isWeekend(day):\n",
        "    if day in [6,7]:\n",
        "        return 1\n",
        "    else:\n",
        "        return 0\n",
        "    \n",
        "df[\"date_time\"] = df[\"date_time\"].apply(lambda x: datetime.strptime(x,'%Y-%m-%d %H:%M:%S'))\n",
        "df[\"year\"] = df[\"date_time\"].apply(lambda x: x.year)\n",
        "df[\"month\"] = df[\"date_time\"].apply(lambda x: x.month)\n",
        "df[\"day\"] = df[\"date_time\"].apply(lambda x: x.day)\n",
        "df[\"hour\"] = df[\"date_time\"].apply(lambda x: x.hour)\n",
        "df[\"weekday\"] = df[\"date_time\"].apply(lambda x: x.isoweekday())\n",
        "df[\"is_weekend\"] = df[\"weekday\"].apply(lambda x: isWeekend(x))\n",
        "\n",
        "df = df.drop(['date_time','weather_description'],axis = 1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WfDCmPhnLAlM",
        "colab_type": "code",
        "outputId": "b10c371d-adc2-4d48-bed3-2e0332a03157",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 433
        }
      },
      "source": [
        "df.info()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 33750 entries, 200 to 12011\n",
            "Data columns (total 19 columns):\n",
            "is_holiday             33750 non-null int64\n",
            "air_pollution_index    33750 non-null int64\n",
            "humidity               33750 non-null int64\n",
            "wind_speed             33750 non-null int64\n",
            "wind_direction         33750 non-null int64\n",
            "visibility_in_miles    33750 non-null int64\n",
            "dew_point              33750 non-null int64\n",
            "temperature            33750 non-null float64\n",
            "rain_p_h               33750 non-null float64\n",
            "snow_p_h               33750 non-null float64\n",
            "clouds_all             33750 non-null int64\n",
            "weather_type           33750 non-null int64\n",
            "traffic_volume         33750 non-null int64\n",
            "year                   33750 non-null int64\n",
            "month                  33750 non-null int64\n",
            "day                    33750 non-null int64\n",
            "hour                   33750 non-null int64\n",
            "weekday                33750 non-null int64\n",
            "is_weekend             33750 non-null int64\n",
            "dtypes: float64(3), int64(16)\n",
            "memory usage: 5.1 MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9Cl9Wg4kOPgJ",
        "colab_type": "code",
        "outputId": "0973cee2-c3e0-4d34-8da9-a718cdf0d308",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        }
      },
      "source": [
        "df.columns"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['is_holiday', 'air_pollution_index', 'humidity', 'wind_speed',\n",
              "       'wind_direction', 'visibility_in_miles', 'dew_point', 'temperature',\n",
              "       'rain_p_h', 'snow_p_h', 'clouds_all', 'weather_type', 'traffic_volume',\n",
              "       'year', 'month', 'day', 'hour', 'weekday', 'is_weekend'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wrdxX6_tOS8a",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_col = ['is_holiday', 'air_pollution_index', 'humidity', 'wind_speed','wind_direction', 'visibility_in_miles', 'dew_point', 'temperature','rain_p_h', 'snow_p_h', 'clouds_all', 'weather_type', 'year', 'month', 'day', 'hour', 'weekday', 'is_weekend']\n",
        "Y_col = ['traffic_volume']"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YarOsJuzLAmM",
        "colab_type": "code",
        "outputId": "f82d741f-618a-451b-a7ad-46b2512de328",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "hidden_layers = [10, 15, 20]\n",
        "iterations = [1000, 5000, 10000]\n",
        "alpha = [10, 1, 0.01]\n",
        "\n",
        "param_list = [[i, j, k] for i in hidden_layers  \n",
        "                 for j in iterations \n",
        "                 for k in alpha] \n",
        "\n",
        "len(param_list)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "27"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vOVpio9rLAmW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q4ERYh3ZLAmc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vfdLcoPDLAm4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def data_partition(df, k):\n",
        "    part = len(df)//k\n",
        "    datasets =[]\n",
        "    for i in range(k):\n",
        "\n",
        "        if i == k-1:\n",
        "            datasets.append(df[part*i:])\n",
        "        else:\n",
        "             datasets.append(df[part*i:part*(i+1)])\n",
        "    return datasets"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g23XszYrLAnA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def K_CV(data, k, X, Y, in_layers, out_layers, param_list):\n",
        "      ds = data_partition(data, k)\n",
        "      test_errors = []\n",
        "      best_params = []\n",
        "      input_layers= in_layers\n",
        "      output_layers = out_layers\n",
        "      for i in range(k):\n",
        "              testset = ds[i]\n",
        "              X_test = testset[X].as_matrix()\n",
        "              Y_test = testset[Y].as_matrix().reshape(-1,1)    \n",
        "              \n",
        "              rem = []\n",
        "              for p in range(k):\n",
        "                      if p !=i:\n",
        "                          rem.append(ds[p])\n",
        "          \n",
        "              for j in range(len(rem)):\n",
        "                      rem_copy = deepcopy(rem)\n",
        "                      validate_set = rem[j]\n",
        "                      trainset = rem_copy.pop(j)\n",
        "\n",
        "                      X_train = trainset[X].as_matrix()\n",
        "                      Y_train = trainset[Y].as_matrix().reshape(-1,1)\n",
        "                      X_validate  =  validate_set[X].as_matrix()\n",
        "                      Y_validate =  validate_set[Y].as_matrix().reshape(-1,1)\n",
        "\n",
        "                      parms_errors = []\n",
        "                      for l in range(len(param_list)):\n",
        "                              params = param_list[l]\n",
        "                              # Initializing a class\n",
        "                              cls  = nn.NeuralNet([input_layers, params[0], output_layers], params[2])\n",
        "                              cls.train(X_train, Y_train, niter= params[1])\n",
        "                             \n",
        "                              #Predicting a the output of testset\n",
        "                              Pred = cls.use(X_validate)\n",
        "                              err = np.mean(np.abs(Pred - Y_validate))\n",
        "                              #print(err)\n",
        "                              parms_errors.append(err)\n",
        "                             \n",
        "                        \n",
        "                      \n",
        "                      b_parm = param_list[parms_errors.index(min(parms_errors))]\n",
        "                      print(parms_errors)\n",
        "                      print(b_parm)\n",
        "                      best_params.append(b_parm)\n",
        "                      cls_best  = nn.NeuralNet([input_layers, b_parm[0] , output_layers], b_parm[2])\n",
        "                      cls_best.train(X_train, Y_train, niter= b_parm[1])\n",
        "                      Pred_test = cls.use(X_test)\n",
        "                      err = np.mean(np.abs(Pred_test - Y_test))\n",
        "                      test_errors.append(err)\n",
        "                      print(err)\n",
        "                      \n",
        "      print(\"The best param set is \")\n",
        "      zkl = best_params[test_errors.index(min(test_errors))]\n",
        "      print(zkl)\n",
        "      return zkl"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lAx6ysfyLAmm",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# # SAMPLE DATA\n",
        "\n",
        "X = np.linspace(-1000,1000, 500).reshape((-1, 1))\n",
        "T = ((-0.75 * X) + 750)\n",
        "data = np.hstack([X, T])\n",
        "data = pd.DataFrame(data, columns=['X', 'T'])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "DW3V7MBl8ruA",
        "outputId": "42ed34bc-413a-4136-b76e-08029f28d489",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "sample_data_params = K_CV(data= data,k=5, X= ['X'], Y= ['T'], in_layers=1, out_layers=1, param_list=param_list)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[66.7290921227239, 0.27735261843450415, 0.06292660204466075, 14.054713543174735, 0.3167955049293789, 0.07951209233817849, 48.843553305602626, 0.11735752209783186, 0.148091555160305, 20.904431470460572, 0.1195542181023086, 0.058394763311284806, 7.394554444343711, 0.4207525451316883, 0.040084328590743325, 29.7209639275954, 0.2629454596422386, 0.04945900073974599, 3.012774088475395, 0.21520404051516492, 0.058626864566233505, 28.411481122427627, 0.018581289468600062, 0.09888804667425802, 7.318327498700168, 0.13497243734889822, 0.09167125704560021]\n",
            "[20, 5000, 1]\n",
            "48.37122835358932\n",
            "[20.06036690221445, 0.19017662411378639, 0.31442423312547474, 30.446043893057258, 0.2436534747959638, 0.2703208376320606, 7.656820971434761, 0.33134821633288425, 0.1456856938078988, 41.33702455890376, 0.04444729287089331, 0.23124218295551713, 9.582915949533797, 0.3025725723424034, 0.15720690407528765, 48.126863288917946, 0.09736596164393518, 0.28637678136266914, 19.806028896702728, 0.04778706855486575, 0.2308821285478109, 47.0992714931789, 0.06880951679111036, 0.10714844386373443, 3.4663966759825438, 0.0423300974634833, 0.07854093878769959]\n",
            "[20, 10000, 1]\n",
            "195.08824884571123\n",
            "[69.94359534310361, 0.19842578383544038, 0.1395356546507651, 56.163497447589215, 0.2805501161202068, 0.07745387527027106, 5.776010043382903, 0.27861155366297735, 0.11673734192281188, 10.24484439743106, 0.22444559526643104, 0.22095576174056475, 3.9410380155041396, 0.2609947496611216, 0.07400611314770458, 24.10104358188945, 0.2645233306719621, 0.0726903923883492, 8.853816611980365, 0.22454984059961589, 0.11158711920952896, 4.31452187634928, 0.24117719629789292, 0.08154743303597116, 5.105720773161748, 0.036024605011295424, 0.047010839759261104]\n",
            "[20, 10000, 1]\n",
            "381.0330557328985\n",
            "[59.48872996483319, 0.30597658562432145, 0.10244546251962419, 10.998935874922301, 0.26839215330655336, 0.2795870703825399, 15.591583372904537, 0.09779223063614992, 0.12733072412594482, 7.21514537798517, 0.25632054248482095, 0.2839264065944006, 4.768948165288133, 0.043723200387357226, 0.2749534729052371, 4.525210380003848, 0.2567685755574882, 0.2358721708068937, 45.13267322941247, 0.02947230490087321, 0.2804203601213399, 4.268682875487649, 0.19034130439309024, 0.06867136120839618, 2.6974812606710943, 0.09363012554838335, 0.09932496720202792]\n",
            "[20, 1000, 1]\n",
            "803.8054638791377\n",
            "[17.989998222709538, 0.11727900168788664, 0.29749482487791057, 59.17591893164753, 0.2856792519688065, 0.7772979722663468, 37.794632823855075, 0.28974324027413784, 0.30542636227290587, 9.645070654126364, 0.26508685318065545, 0.1365642090413644, 11.524464274763437, 0.1013280154254153, 0.07809868612785294, 27.561316389103794, 0.0909356773414629, 0.255508258332527, 12.970859175732967, 0.11268072333218697, 0.035843974304250424, 5.393721083540661, 0.0735371679763989, 0.05716746788427145, 1.3036868268107105, 0.042236647807583266, 0.047502564389621964]\n",
            "[20, 1000, 0.01]\n",
            "40.70200050127201\n",
            "[50.051532938413715, 0.3009704352619963, 0.0741188737319726, 5.2452525780727965, 0.31182124362946867, 0.3081638553743926, 9.418770855999254, 0.3156725476360418, 0.10518273131601859, 1.798351158914951, 0.27698530744483835, 0.2832105135293159, 0.917253797871009, 0.15585463956017975, 0.23476112636317795, 0.6322542940879226, 0.07862403801089272, 0.1097852905722823, 7.096625802394143, 0.2502865806781051, 0.09400514031800754, 10.322884954105202, 0.06787658779009689, 0.051231335223781115, 2.7221542391993396, 0.1100383305327432, 0.2632095514608557]\n",
            "[20, 5000, 0.01]\n",
            "16.40291108497584\n",
            "[56.27809604374174, 0.23867875652305487, 0.3152940067158033, 57.10975822602065, 0.2946260727646728, 0.3258556918582622, 51.98664307110133, 0.10602932230495866, 0.10388822939277872, 9.670467559149579, 0.10462068296042332, 0.281805323706738, 21.132108584668284, 0.15641866281418174, 0.07489034460856715, 3.781855213806176, 0.28545347694016415, 0.2674949171223466, 5.822374463065175, 0.11905901203754922, 0.2492273518697607, 5.872422000507895, 0.12718848388516166, 0.07380725707614318, 5.53611174875616, 0.12333097167484994, 0.037501140240540846]\n",
            "[20, 10000, 0.01]\n",
            "154.98886344384343\n",
            "[66.44871387540033, 0.2798272504967052, 0.10889523451238275, 26.395514557868964, 0.24907715078183998, 0.2931977707992654, 13.989403006787677, 0.32553457733305835, 0.11345599044632636, 9.06132662803469, 0.11921444628440028, 0.04887122713913939, 6.5930588872332745, 0.2696097586963829, 0.11555359994104662, 24.743229280060554, 0.05400534953879983, 0.2992787434760038, 0.29996925421406556, 0.10190480310997245, 0.1279535903827103, 4.168187648039423, 0.12571783505878897, 0.10237850880741405, 1.4011382694459231, 0.23050511833403817, 0.2236716857610594]\n",
            "[15, 1000, 0.01]\n",
            "247.5163263094554\n",
            "[19.77538528628824, 0.5154623238579847, 0.28007989116984844, 8.828156928015419, 0.1591430458569698, 0.32575142586152195, 33.436598637063824, 0.1265040693750848, 0.09461540809988037, 61.02925022687461, 0.23021232905567784, 0.24551888255964058, 6.041372893129121, 0.2383185917236142, 0.06787368061629649, 8.114556638567363, 0.10199406675254749, 0.07272690973977433, 5.241870542307493, 0.20951880600889125, 0.11637262029727935, 4.501615642508927, 0.11692584207553637, 0.026754228263500865, 3.3165566114822855, 0.14537794733854753, 0.07351202074154571]\n",
            "[20, 5000, 0.01]\n",
            "178.9482911864843\n",
            "[68.22752895446592, 0.31747075423612275, 0.28350890201718926, 22.77221944610641, 0.29957360331658944, 0.08035054957550415, 11.780820077657772, 0.2924808031008638, 0.23120943886533837, 8.102013282879, 0.09081085621461511, 0.09430831118263881, 1.8165325165722674, 0.5143071402161036, 0.2592559841429056, 5.412096936168258, 0.07650987349927277, 0.22477652243137755, 1.5364689366417565, 0.22972113622464804, 0.04795385939365815, 31.30748866768061, 0.01656002099981265, 0.14156916287936724, 1.0398983984486176, 0.08986655819575957, 0.2037914557729391]\n",
            "[20, 5000, 1]\n",
            "13.014182407859296\n",
            "[61.31243924214462, 0.1333278165529714, 0.05630540672849236, 44.59059692909796, 0.10203282821255356, 0.2958584627721109, 34.8632115564106, 0.06898071486308652, 0.06422879472494174, 6.930802890655929, 0.14142126360293786, 0.2818643033756223, 43.38539539002663, 0.11451800141992464, 0.09045621091744466, 3.118804896502979, 0.257603459047416, 0.07306723359482248, 5.550342684150933, 0.0812145531732483, 0.32138972590070636, 5.973849322854596, 0.057952514808549724, 0.23460541537059554, 2.6973336537731605, 0.06679523817256779, 0.24527316798630922]\n",
            "[10, 1000, 0.01]\n",
            "15.102052641034351\n",
            "[21.298109791793525, 0.0927362081453164, 0.199326781134653, 5.936710182440473, 0.31005623052870507, 0.3105953219366059, 40.30688963348364, 0.32352128516824974, 0.16299171472555984, 37.64354958727573, 0.26621821471793067, 0.11913002399245556, 46.70498995806868, 0.07133540516650286, 0.33868270447716425, 1.922343122280927, 0.276839771269952, 0.2599141755460421, 6.436579782536386, 0.05136613610595276, 0.21350585415311044, 5.333487634209946, 0.18068012214634932, 0.10443224124441158, 2.270755705081704, 0.07283952374715157, 0.04804823511988601]\n",
            "[20, 10000, 0.01]\n",
            "163.6231442789184\n",
            "[68.55671878052952, 0.1570296767756372, 0.28470390461594663, 14.550370982688435, 0.30686845182203343, 0.14805563331005941, 24.62565081272914, 0.2491724213736188, 0.14497124752021365, 9.410315206683608, 0.266893188492038, 0.2512052796226453, 29.551027569062608, 0.1265038107042392, 0.28139177160066536, 5.99461527055563, 0.21321000880734345, 0.24831764499023848, 13.664092009209305, 0.23444161075496595, 0.037948860039377905, 1.4916639980387822, 0.04940312243391873, 0.10188107162608731, 19.563015478319596, 0.07701877141547811, 0.11356813666198605]\n",
            "[20, 1000, 0.01]\n",
            "465.49460083065014\n",
            "[43.93941514608202, 0.2585686770171787, 0.3058012681076332, 19.70016442100783, 0.34220824565132146, 0.10893227463297421, 16.926846993447715, 0.11609409352539729, 0.1135531523428756, 19.876954902173896, 0.27845757510498403, 0.1301300705831295, 6.2678559409881975, 0.13851120518933385, 0.25136106292210914, 4.30394925699463, 0.25674281480532274, 0.024391068929078302, 7.013648590368027, 0.26438933611635773, 0.03751275578126979, 9.440163105251848, 0.028554290414464275, 0.2602748846049576, 7.617982783664527, 0.05721757795081885, 0.09867761482403921]\n",
            "[15, 10000, 0.01]\n",
            "237.05895901688461\n",
            "[6.600839578183473, 0.3188902487171708, 0.33425386582338545, 4.921448592287902, 0.2721066317514999, 0.11768045731428402, 3.854715587264875, 0.11031972169144638, 0.13581325657026697, 42.23466500725224, 0.2614425265433465, 0.24204931716691816, 10.785187766941768, 0.09638405862827767, 0.07365491844283951, 4.684828311038802, 0.06683479862203967, 0.26900982623726466, 50.706810986203465, 0.04980330250042471, 0.2001606202405219, 38.305905229217984, 0.2143366323806174, 0.043876835785573574, 2.316656421037882, 0.23722365911221005, 0.09271185596565829]\n",
            "[20, 5000, 0.01]\n",
            "29.275979202709994\n",
            "[13.196773506787947, 0.28595895291137213, 0.2917253260030776, 6.909233557841141, 0.435148361769548, 0.32324812548309295, 51.64815696605112, 0.26386299365437355, 0.3152845928642209, 5.414074572457736, 0.0891949153199016, 0.08164228336693156, 11.005776420385237, 0.1262472730079682, 0.2533156318692072, 6.502456674213184, 0.24031648142040696, 0.036214911933929554, 6.215228646823878, 0.17938629325912644, 0.11401552977761056, 35.91344246728519, 0.09909941206317953, 0.029528061336231134, 6.607336295143925, 0.24757195674851984, 0.23334320357349725]\n",
            "[20, 5000, 0.01]\n",
            "14.762918237215175\n",
            "[17.765409352886365, 0.3009507433440308, 0.3185508995234295, 14.156885105105202, 0.29708459590935266, 0.07549508346085532, 34.1685380113615, 0.15518211564268541, 0.31794509713788327, 10.905141160183296, 0.25021735972735315, 0.24702253869443438, 8.242581777766738, 0.03553134973662509, 0.04520321293174447, 4.153563438775489, 0.27184334223249607, 0.12207098097831931, 10.675660957783352, 0.10817110916968431, 0.2243236221856887, 12.681783370155877, 0.04629645712627507, 0.09331386226905351, 6.095319032354321, 0.07196801686851131, 0.19544284769989417]\n",
            "[15, 5000, 1]\n",
            "436.41550266639666\n",
            "[14.858134166776631, 0.3301090977285003, 0.31290513008540644, 44.2267104798891, 0.3095668036794552, 0.3095520268156088, 5.868531093539789, 0.10580901119702389, 0.2971134446896281, 51.80171909633399, 0.2573456161171373, 0.26257190945127834, 0.933641326563303, 0.2604560131379276, 0.1158256714588299, 3.38394414469839, 0.051241693185112354, 0.2795828276553334, 4.709318765916922, 0.2582920293140228, 0.2010736236297407, 4.047119913718311, 0.0949697940131432, 0.052381633372641546, 4.839898273402562, 0.03117022367008076, 0.09689575194848203]\n",
            "[20, 10000, 1]\n",
            "480.24280285071035\n",
            "[18.741354288882317, 0.2983335629514329, 0.2776481947140462, 14.611812719625505, 0.4309302539977375, 0.3260233904751226, 7.823623850000619, 0.25942562788559714, 0.33285659038380117, 12.051889571935167, 0.27955872797089, 0.2799002892034275, 21.909383855427677, 0.040856205450885454, 0.27395788186830033, 4.583960607331721, 0.11441744784066828, 0.09948753755605139, 3.237079664753062, 0.25112308968570574, 0.08396512438019613, 8.925728108967846, 0.12099764816331572, 0.24818241761385137, 2.8846141789074933, 0.25430764774757564, 0.24022236900949792]\n",
            "[15, 5000, 1]\n",
            "96.32879732965107\n",
            "[41.26057060803961, 0.26837227650341844, 0.15273542658360667, 14.816607913167932, 0.26448529832837947, 0.16307143904237875, 16.595586235519914, 0.13909720377083146, 0.29122993909789274, 69.11481785868187, 0.06122211005308998, 0.05937430828404331, 4.699899048439286, 0.13078673032877647, 0.28524650809961144, 14.14423571297673, 0.22597395695098782, 0.2180834416911307, 56.32866241929988, 0.25955887408466366, 0.10093395211480413, 4.6708020495040925, 0.3168521755590069, 0.07379702053113818, 3.026494147155758, 0.07148239328542899, 0.21294036365841806]\n",
            "[15, 1000, 0.01]\n",
            "13.600246872815164\n",
            "The best param set is \n",
            "[20, 5000, 1]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8vuqQ2KOLAnp",
        "colab_type": "code",
        "outputId": "42740c16-27dd-45ae-8a2f-0c0296bc8a3f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 332
        }
      },
      "source": [
        "## Actual data hyper parameters\n",
        "df_params = K_CV(data= df, k=5, X= X_col, Y= Y_col, in_layers= len(X_col), out_layers=1,\n",
        "                 param_list = param_list)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[764.98, 512.91, 526.84, 1645.35, 431.49, 431.16, 1134.5, 369.48, 369.09, 1668.62, 493.12, 388.77, 1660.83, 336.69, 319.92, 909.69, 316.86, 336.44, 903.75, 423.05, 419.19, 1561.72, 307.18, 305.77, 1595.73, 322.31, 298.61, 1749.01, 478.15, 540.42, 1683.63, 385.76, 476.26, 1656.58, 348.39, 362.62, 1604.31, 504.58, 422.5, 1364.67, 353.96, 459.63, 1599.27, 309.32, 360.9, 1606.84, 401.69, 410.78, 1652.9, 308.41, 318.0, 1360.13, 300.72, 289.74, 1708.72, 480.05, 554.05, 1663.73, 472.56, 370.59, 1107.76, 370.37, 370.66, 1540.49, 485.46, 524.88, 1632.78, 329.28, 332.0, 1688.99, 320.13, 312.07, 1541.52, 408.98, 420.25, 1530.78, 342.36, 301.07, 1092.23, 299.52, 299.18, 1740.77, 486.3, 544.4, 1337.58, 389.51, 380.31, 951.86, 440.97, 349.51, 1574.01, 485.85, 437.63, 1603.23, 315.14, 327.41, 1595.09, 304.45, 336.95, 1568.24, 401.46, 372.99, 1569.33, 294.44, 288.56, 1576.55, 292.46, 285.45]\n",
            "[20, 5000, 0.01]\n",
            "363.79439575558996\n",
            "[1607.47, 536.51, 505.65, 1683.67, 524.87, 423.05, 1693.02, 385.29, 395.31, 1621.69, 462.88, 450.97, 1588.52, 335.33, 315.47, 1272.12, 324.32, 345.51, 1406.71, 485.14, 378.98, 1571.93, 288.42, 295.05, 1411.81, 272.94, 346.21, 1700.05, 415.02, 486.6, 1736.83, 485.92, 390.56, 1695.46, 449.6, 347.83, 1656.12, 463.06, 469.23, 1147.69, 323.31, 321.93, 1729.43, 309.54, 316.54, 1615.93, 375.96, 405.68, 1608.81, 295.25, 301.16, 1614.24, 299.22, 303.71, 1652.27, 545.93, 522.39, 1679.28, 489.38, 387.48, 1698.96, 388.28, 486.24, 1627.94, 471.99, 492.96, 1705.48, 332.73, 326.86, 1453.63, 324.86, 305.79, 958.13, 397.56, 414.25, 1118.76, 298.59, 296.99, 1544.95, 297.52, 296.52, 1692.82, 503.35, 553.15, 1682.96, 481.51, 446.74, 1509.11, 481.04, 348.92, 1511.71, 408.9, 472.0, 1527.71, 343.65, 311.71, 1654.33, 299.31, 316.89, 1574.14, 509.74, 379.11, 1606.75, 291.82, 289.77, 1574.08, 291.6, 284.58]\n",
            "[20, 5000, 1]\n",
            "358.254320007423\n",
            "[1718.91, 502.07, 455.76, 1721.89, 376.23, 388.7, 1669.51, 451.31, 480.56, 1571.75, 518.25, 527.48, 1574.86, 296.73, 302.95, 1508.0, 350.93, 311.34, 1558.53, 492.71, 431.52, 1573.08, 292.54, 292.86, 1577.13, 284.6, 284.82, 1698.49, 485.44, 528.29, 1684.84, 400.92, 383.29, 1618.69, 374.09, 385.74, 1591.26, 481.92, 518.94, 1567.86, 338.2, 316.59, 1574.16, 353.5, 360.27, 1431.18, 433.88, 527.16, 1573.91, 294.6, 315.93, 1589.11, 285.01, 294.94, 1695.06, 411.41, 530.5, 1529.67, 442.14, 480.43, 1627.81, 386.35, 364.06, 1663.37, 469.84, 553.3, 1568.23, 344.55, 350.32, 1646.91, 317.98, 333.28, 1582.4, 551.66, 388.78, 1453.21, 305.08, 304.98, 1546.32, 305.55, 296.87, 1710.51, 502.08, 408.94, 1725.92, 380.99, 384.8, 1543.68, 410.57, 374.97, 1554.51, 511.36, 490.02, 1620.44, 307.19, 461.12, 1615.64, 337.25, 305.84, 1571.17, 495.69, 424.52, 1572.29, 288.7, 286.03, 1596.32, 290.3, 279.46]\n",
            "[20, 5000, 0.01]\n",
            "351.86048638068166\n",
            "[1692.53, 549.8, 507.62, 1614.12, 482.89, 350.04, 1680.29, 376.84, 480.88, 1619.27, 497.6, 494.68, 1569.67, 338.47, 309.12, 1625.12, 293.0, 299.56, 1564.01, 498.43, 399.43, 1558.0, 291.45, 291.23, 1570.44, 275.99, 312.7, 1590.38, 485.94, 520.37, 1667.08, 382.08, 389.41, 1575.51, 384.4, 458.77, 1649.69, 413.35, 516.39, 1561.15, 321.78, 459.02, 1590.13, 336.77, 339.68, 1565.99, 415.64, 476.84, 1562.01, 300.34, 329.11, 1113.69, 301.29, 299.5, 1743.06, 504.63, 508.91, 1719.45, 466.05, 378.13, 887.7, 485.68, 387.12, 896.08, 487.69, 411.03, 1607.43, 328.63, 308.73, 1650.85, 345.28, 329.9, 1377.3, 463.89, 421.12, 1600.62, 322.21, 300.58, 1347.35, 302.76, 286.12, 1731.8, 491.58, 481.09, 1727.28, 356.12, 397.26, 1617.07, 441.93, 356.47, 1607.12, 526.2, 495.49, 1566.99, 318.26, 300.23, 1676.14, 303.4, 324.18, 1044.96, 388.67, 495.69, 1654.89, 291.57, 301.6, 1602.6, 282.84, 286.2]\n",
            "[20, 5000, 1]\n",
            "354.0771714620143\n",
            "[1711.51, 522.3, 513.86, 1666.67, 365.61, 373.75, 1571.09, 351.14, 372.92, 1578.56, 418.34, 389.64, 1668.65, 360.44, 312.35, 1623.96, 299.32, 298.56, 1557.35, 495.62, 506.62, 1048.41, 284.47, 283.98, 1580.35, 282.74, 288.44, 1584.27, 514.04, 549.62, 1606.9, 358.44, 408.22, 1660.02, 379.89, 385.34, 1016.55, 391.59, 410.85, 1210.59, 326.58, 468.94, 1558.73, 327.36, 322.18, 1353.9, 404.94, 440.13, 1600.99, 303.1, 298.5, 1584.13, 296.8, 292.82, 1758.08, 554.36, 527.06, 1618.09, 464.63, 487.25, 1717.54, 384.19, 382.77, 1655.3, 408.81, 452.84, 1606.25, 369.54, 321.88, 1627.07, 313.95, 361.28, 1606.12, 390.19, 545.93, 1607.24, 307.13, 300.03, 1663.76, 301.43, 298.88, 1713.69, 419.62, 491.94, 1675.44, 366.59, 370.52, 1672.74, 365.34, 378.48, 1611.58, 411.23, 425.96, 1613.97, 365.41, 328.8, 1662.44, 317.27, 334.7, 1591.99, 493.18, 393.97, 1573.28, 291.24, 311.19, 1584.41, 322.92, 293.97]\n",
            "[20, 5000, 1]\n",
            "351.080492623097\n",
            "The best param set is \n",
            "[20, 5000, 1]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mp6YTZh2cBv5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vUUoGY-S5med",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "9pHtqoziKaXY",
        "colab": {}
      },
      "source": [
        "# clsf1  = nn.NeuralNet([18, 1, 1])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "1tKSTst2KYb4",
        "colab": {}
      },
      "source": [
        "# _test = datasets[0]\n",
        "# _train = df[~df.index.isin(datasets[0].index)]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "gRURsCFZKXDU",
        "colab": {}
      },
      "source": [
        "# X_train = _train.drop('traffic_volume', axis=1)\n",
        "# Y_train = _train['traffic_volume']\n",
        "\n",
        "# X_test = _test.drop('traffic_volume', axis=1)\n",
        "# Y_test = _test['traffic_volume']\n",
        "\n",
        "# X_train_mat = X_train.as_matrix()\n",
        "# X_test_mat = X_test.as_matrix()\n",
        "# Y_train_mat = Y_train.as_matrix()\n",
        "# Y_test_mat = Y_test.as_matrix()\n",
        "\n",
        "# Y_train_mat = Y_train_mat.reshape(-1,1)\n",
        "# print(Y_train_mat.shape)\n",
        "# Y_test_mat = Y_test_mat.reshape(-1,1)\n",
        "# print(Y_test_mat.shape)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pIEY2VQ4LAoU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# X_train.shape\n",
        "# Y_train.shape"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W2q85I2TLAoW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# clsf1.train(X_train_mat, Y_train_mat, _lambda = 0.001)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JbgdrvNhLAoZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Y_pred = clsf1.use(X_test_mat)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M0eaYHxVLAoc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# #np.sum(np.square(Y_pred-Y_test_mat))/len(Y_pred)\n",
        "\n",
        "# np.mean(np.abs(Y_test_mat - Y_pred)) "
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}