LSTM.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "LSTM.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5V5iK9VkPMMn"
      },
      "source": [
        "### 2. Aplicación: Prediciendo las acciones de Netflix\n",
        "***"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Twyya2zdErLv"
      },
      "source": [
        "#### Cargamos librerias básicas para el manejo de datos, gráficas y cálculos en python"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PBv8ZTpBYUNH"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eXR1NqlwUl23"
      },
      "source": [
        "Ubicamos en nuestro espacio de trabajo las bases de datos que usaremos"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1UTV94s3JuFY",
        "colab": {
          "resources": {
            "http://localhost:8080/nbextensions/google.colab/files.js": {
              "data": "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",
              "ok": true,
              "headers": [
                [
                  "content-type",
                  "application/javascript"
                ]
              ],
              "status": 200,
              "status_text": ""
            }
          },
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "outputId": "9e8a8282-03a6-41d1-ab69-4ea7157b1dc5"
      },
      "source": [
        "from google.colab import files\n",
        "uploaded = files.upload()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "     <input type=\"file\" id=\"files-94a2600c-8049-4048-ba74-1b9bd9427e7e\" name=\"files[]\" multiple disabled\n",
              "        style=\"border:none\" />\n",
              "     <output id=\"result-94a2600c-8049-4048-ba74-1b9bd9427e7e\">\n",
              "      Upload widget is only available when the cell has been executed in the\n",
              "      current browser session. Please rerun this cell to enable.\n",
              "      </output>\n",
              "      <script src=\"/nbextensions/google.colab/files.js\"></script> "
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Saving NFLX1.csv to NFLX1 (1).csv\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vaRf0wjqFuR5"
      },
      "source": [
        "Importamos el  Dataset de entrenamiento el cual consta de 1677 datos con información diaria de las acciones de Netflix desde el 2 de enero del 2014 hasta el 31 de agosto del 2020"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pTi9mNe1YUNK"
      },
      "source": [
        "nt_completa= pd.read_csv(r'NFLX1.csv')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yjLUHuwZNv70"
      },
      "source": [
        "Un vistazo de la estructura de la base de datos nos muestra lo siguiente"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3H8ra_XDYUNN",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "f09fa601-0263-4c41-d3e4-000c99f53f0d"
      },
      "source": [
        "nt_completa.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Date</th>\n",
              "      <th>Open</th>\n",
              "      <th>High</th>\n",
              "      <th>Low</th>\n",
              "      <th>Close</th>\n",
              "      <th>Adj Close</th>\n",
              "      <th>Volume</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2014-01-02</td>\n",
              "      <td>52.401428</td>\n",
              "      <td>52.511429</td>\n",
              "      <td>51.542858</td>\n",
              "      <td>51.831429</td>\n",
              "      <td>51.831429</td>\n",
              "      <td>12325600</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2014-01-03</td>\n",
              "      <td>52.000000</td>\n",
              "      <td>52.495716</td>\n",
              "      <td>51.842857</td>\n",
              "      <td>51.871429</td>\n",
              "      <td>51.871429</td>\n",
              "      <td>10817100</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2014-01-06</td>\n",
              "      <td>51.889999</td>\n",
              "      <td>52.044285</td>\n",
              "      <td>50.475716</td>\n",
              "      <td>51.367142</td>\n",
              "      <td>51.367142</td>\n",
              "      <td>15501500</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2014-01-07</td>\n",
              "      <td>49.684284</td>\n",
              "      <td>49.698570</td>\n",
              "      <td>48.152859</td>\n",
              "      <td>48.500000</td>\n",
              "      <td>48.500000</td>\n",
              "      <td>36167600</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2014-01-08</td>\n",
              "      <td>48.104286</td>\n",
              "      <td>49.425713</td>\n",
              "      <td>48.074287</td>\n",
              "      <td>48.712856</td>\n",
              "      <td>48.712856</td>\n",
              "      <td>20001100</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "         Date       Open       High        Low      Close  Adj Close    Volume\n",
              "0  2014-01-02  52.401428  52.511429  51.542858  51.831429  51.831429  12325600\n",
              "1  2014-01-03  52.000000  52.495716  51.842857  51.871429  51.871429  10817100\n",
              "2  2014-01-06  51.889999  52.044285  50.475716  51.367142  51.367142  15501500\n",
              "3  2014-01-07  49.684284  49.698570  48.152859  48.500000  48.500000  36167600\n",
              "4  2014-01-08  48.104286  49.425713  48.074287  48.712856  48.712856  20001100"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wdUCVqpeRzt8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "f395ded9-c82d-4246-fa41-c458882cdf4b"
      },
      "source": [
        "len(nt_completa)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1677"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YU2aCdynWJPa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 153
        },
        "outputId": "b2285f3e-425a-4d32-c3ee-da8aa8a2b3b8"
      },
      "source": [
        "print(nt_completa.dtypes)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Date          object\n",
            "Open         float64\n",
            "High         float64\n",
            "Low          float64\n",
            "Close        float64\n",
            "Adj Close    float64\n",
            "Volume         int64\n",
            "dtype: object\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VKdlm_tnI2Xe"
      },
      "source": [
        "nt_completa.sort_values('Date', ascending=True, inplace=True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E6-ESY92LCVw",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "5aa394a8-66c2-4bfe-be19-da66d882d440"
      },
      "source": [
        "from matplotlib import pyplot as plt\n",
        "plt.figure()\n",
        "plt.plot(nt_completa[\"Open\"])\n",
        "plt.plot(nt_completa[\"High\"])\n",
        "plt.plot(nt_completa[\"Low\"])\n",
        "plt.plot(nt_completa[\"Close\"])\n",
        "plt.title('Histórico precio de acciones Netflix')\n",
        "plt.ylabel('Precio (USD)')\n",
        "plt.xlabel('Días')\n",
        "plt.legend(['Open','High','Low','Close'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wxb6QYlDN5pL",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 170
        },
        "outputId": "7ea2a28d-1c58-4dce-e457-3612ad9bbe1a"
      },
      "source": [
        "print(\"Verificación de que no existan NAs en la base\\n\", nt_completa.isna().sum())"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Verificación de que no existan NAs en la base\n",
            " Date         0\n",
            "Open         0\n",
            "High         0\n",
            "Low          0\n",
            "Close        0\n",
            "Adj Close    0\n",
            "Volume       0\n",
            "dtype: int64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zzP5idwKL3j4"
      },
      "source": [
        "#### Elegimos la columna de interés: **Open**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "msv4bWMuYUNR"
      },
      "source": [
        "nt_processed = nt_completa.iloc[:,1:2].values"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tORNsZDMlR8I"
      },
      "source": [
        "Para la decisión del número de covariables se dará un primer vistazo al gráfico de la función de autocovarianza para 300 rezagos "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "REly2Gg-OxXz",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "outputId": "da64fdd5-f27f-476c-85b5-6aa67d02776d"
      },
      "source": [
        "from statsmodels.graphics.tsaplots import plot_acf\n",
        "from matplotlib import pyplot\n",
        "plot_acf(nt_processed,lags=300)\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9D6WNHNEmxWR"
      },
      "source": [
        "Se puede ver en la grafica de autocorrelación que existe una alta correlación en los primeros 300 rezagos, aún así, parecen se los primeros 60 rezagos los más correlacionados por lo que se decide tomar estos como autocovariables."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ts0B9A9Ao5-d"
      },
      "source": [
        "#### Escalamiento de los datos"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "s2WJFMnHLzQT"
      },
      "source": [
        "Por regla general al trabajar con redes neuronales se deben normalizar o escalar los datos, en este caso utilizaremos el comando MinMaxScaler del paquete sklearn para escalar los datos entre 0 y 1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h2LULbPJYUNZ"
      },
      "source": [
        "from sklearn.preprocessing import MinMaxScaler\n",
        "scaler = MinMaxScaler(feature_range = (0, 1))\n",
        "\n",
        "nt_training_scaled = scaler.fit_transform(nt_processed)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2PGWJvGvNu6S"
      },
      "source": [
        "#### Creación de variables rezagadas"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1iR2mChUVphh"
      },
      "source": [
        "El objetivo de nuestro proyecto es poder predecir el valor de la acción de apertura de Netflix en el tiempo T basandonos en los datos de los días T-N para cualquier número de pasos N. Para nuestro caso se ha decidido utilizar N=60."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Jni_KICZYUNc"
      },
      "source": [
        "features_set = []\n",
        "labels = []\n",
        "for i in range(60, 1677):\n",
        "    features_set.append(nt_training_scaled[i-60:i, 0])\n",
        "    labels.append(nt_training_scaled[i, 0])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5Cli_z9mbPD8"
      },
      "source": [
        "Lo que hicimos previamente fue crear dos listas(features_set y labels). Para ello creamos un ciclo que comienza desde el registro 61 y almacena los 60 registros previos en features_set y el registro 61 en la lista \"labels\".\n",
        "\n",
        "Esto se puede obsevar más claramente con el siguiente conjunto de datos"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f5_V-u6xP5xN",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 425
        },
        "outputId": "573c9542-59af-496e-ab18-a8d78743e1af"
      },
      "source": [
        "features_set[0:2]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[array([0.01489511, 0.01412811, 0.01391793, 0.00970352, 0.00668465,\n",
              "        0.00806034, 0.00705041, 0.00515611, 0.00797026, 0.00621244,\n",
              "        0.00441367, 0.00541268, 0.00545636, 0.00471119, 0.02051523,\n",
              "        0.01949984, 0.0205125 , 0.01885567, 0.02480607, 0.02589243,\n",
              "        0.02458498, 0.02720261, 0.02544205, 0.02481426, 0.02570682,\n",
              "        0.02762296, 0.03198749, 0.03213215, 0.03344779, 0.03096665,\n",
              "        0.03407013, 0.03404283, 0.0325989 , 0.03223861, 0.03451778,\n",
              "        0.03363613, 0.03760215, 0.03921259, 0.03701257, 0.0384565 ,\n",
              "        0.03519743, 0.03785328, 0.03869124, 0.03895056, 0.03871308,\n",
              "        0.03718454, 0.03559322, 0.03314754, 0.03479346, 0.03152347,\n",
              "        0.03164902, 0.03006043, 0.02937805, 0.02932346, 0.03107856,\n",
              "        0.02545297, 0.01822242, 0.01658742, 0.01604151, 0.01309634]),\n",
              " array([0.01412811, 0.01391793, 0.00970352, 0.00668465, 0.00806034,\n",
              "        0.00705041, 0.00515611, 0.00797026, 0.00621244, 0.00441367,\n",
              "        0.00541268, 0.00545636, 0.00471119, 0.02051523, 0.01949984,\n",
              "        0.0205125 , 0.01885567, 0.02480607, 0.02589243, 0.02458498,\n",
              "        0.02720261, 0.02544205, 0.02481426, 0.02570682, 0.02762296,\n",
              "        0.03198749, 0.03213215, 0.03344779, 0.03096665, 0.03407013,\n",
              "        0.03404283, 0.0325989 , 0.03223861, 0.03451778, 0.03363613,\n",
              "        0.03760215, 0.03921259, 0.03701257, 0.0384565 , 0.03519743,\n",
              "        0.03785328, 0.03869124, 0.03895056, 0.03871308, 0.03718454,\n",
              "        0.03559322, 0.03314754, 0.03479346, 0.03152347, 0.03164902,\n",
              "        0.03006043, 0.02937805, 0.02932346, 0.03107856, 0.02545297,\n",
              "        0.01822242, 0.01658742, 0.01604151, 0.01309634, 0.01344845])]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rAlg6Yw6Q7Hp",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "602cab7d-4ed4-4292-f562-1f1c3b55f2fd"
      },
      "source": [
        "labels[0:2]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.013448452558329535, 0.01078441825646731]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hPaIv49nb651"
      },
      "source": [
        "Para poder usar dichos datos para nuestro entrenamientro es necesario convertir nuestra lista como una matriz numpy con el siguiente código"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "agwS_9Q7YUNe"
      },
      "source": [
        "features_set, labels = np.array(features_set), np.array(labels)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-5QCIpPeqEU6"
      },
      "source": [
        "#### División de la variable entre entrenamiento y prueba"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wncl-eTbsuug"
      },
      "source": [
        "Para nuestro caso hemos decidido utilizar el 90% de los datos como entrenamiento y el 10/ como validación"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vo9IDn8noLVv"
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "features_train, features_test, labels_train, labels_test=train_test_split(features_set,labels,test_size=0.1, train_size=0.9)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "76VvhmsltsSc"
      },
      "source": [
        "Con lo que obtenemos que las siguientes dimensiones para el entrenamiento y prueba"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-UTLadjCo7QY",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 85
        },
        "outputId": "e282a2e1-cccd-4952-de5c-a0b4e4ff300f"
      },
      "source": [
        "print(features_train.shape)\n",
        "print(features_test.shape)\n",
        "\n",
        "print(labels_train.shape)\n",
        "print(labels_test.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(1455, 60)\n",
            "(162, 60)\n",
            "(1455,)\n",
            "(162,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wvxLkvKGOEGj"
      },
      "source": [
        "#### Cambio de forma de los datos como input de la red LSTM"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vp_5GBnRc55M"
      },
      "source": [
        "Por último, para moder entrenar nuestros datos, será necesario expresar nuestra matriz en formato de tres dimensiones ya que este es el formato aceptado por las redes LSTM.\n",
        "*   Primera dimensión: Número de filas o registros de nuestra base de entrenamiento (1455 en nuestro caso)\n",
        "*   Segunda dimensión: Número de pasos de tiempo previos (60) \n",
        "*   Tercera dimensión: Número de features, como sólo estamos trabajando **Open** nuestro indicador es 1\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zZvzf5WyYUNi"
      },
      "source": [
        "features_train = np.reshape(features_train, (features_train.shape[0], features_train.shape[1], 1))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OBgwZphnwzD2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 850
        },
        "outputId": "9a2d9390-a2b4-4e50-851c-c57c727e870e"
      },
      "source": [
        "features_train"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[[0.10049077],\n",
              "        [0.10177094],\n",
              "        [0.1011213 ],\n",
              "        ...,\n",
              "        [0.10773225],\n",
              "        [0.10639478],\n",
              "        [0.10813349]],\n",
              "\n",
              "       [[0.57120555],\n",
              "        [0.57395693],\n",
              "        [0.5918791 ],\n",
              "        ...,\n",
              "        [0.48539697],\n",
              "        [0.478423  ],\n",
              "        [0.47888154]],\n",
              "\n",
              "       [[0.03188923],\n",
              "        [0.03023239],\n",
              "        [0.03119593],\n",
              "        ...,\n",
              "        [0.04593271],\n",
              "        [0.04554785],\n",
              "        [0.04394288]],\n",
              "\n",
              "       ...,\n",
              "\n",
              "       [[0.59571959],\n",
              "        [0.58586048],\n",
              "        [0.58532549],\n",
              "        ...,\n",
              "        [0.50905118],\n",
              "        [0.5142291 ],\n",
              "        [0.49840868]],\n",
              "\n",
              "       [[0.10110766],\n",
              "        [0.10220766],\n",
              "        [0.10826997],\n",
              "        ...,\n",
              "        [0.10566872],\n",
              "        [0.11140076],\n",
              "        [0.11159182]],\n",
              "\n",
              "       [[0.01309634],\n",
              "        [0.01344845],\n",
              "        [0.01078442],\n",
              "        ...,\n",
              "        [0.03744657],\n",
              "        [0.03520016],\n",
              "        [0.03469792]]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 64
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NDvWLuvQeryL"
      },
      "source": [
        "####Entrenando el LSTM"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YO4CLg3RexmL"
      },
      "source": [
        "Una vez hemos adecuado nuestra base para el entrenamiento podemos proceder a la creación de nuestro modelo.\n",
        "Nuestro LSTM será un modelo secuencial con cuatro capas seguido de una capa densa.\n",
        "\n",
        "Así que empezamos importando las librerias para la construcción del modelo"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YfU4wxVaYUNo"
      },
      "source": [
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "from keras.layers import LSTM\n",
        "from keras.layers import Dropout"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "R52-OFZefwI6"
      },
      "source": [
        "El primer paso será inicializar la clase secuencial al modelo y posteriormente añadirle las capas LSTM y la capa densa al modelo"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PbeprAWvYUNr"
      },
      "source": [
        "model = Sequential()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4gFT5ZmXjHu9"
      },
      "source": [
        "\n",
        "\n",
        "*   units=\"Número de neuronas que deseamos en la capa\"\n",
        "*   return_sequences=\"True\" si se le añadiran más capas al modelo\n",
        "*   input_shape[1,]=el primer parametro es el número de pasos de tiempo N\n",
        "*   input_shape[,2]=el segundo parametro es el número de indicadores\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fUdOUgjqYUNt"
      },
      "source": [
        "model.add(LSTM(units=50, return_sequences=True, input_shape=(features_train.shape[1], 1)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "s8ADgFs_kwEK"
      },
      "source": [
        "Añadimos una capa \"Dropout\" la cual se utilizá para evitar el sobre ajuste de nuestro modelo"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z_ReDpGuYUNw"
      },
      "source": [
        "model.add(Dropout(0.2))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JgKccpUxlGtX"
      },
      "source": [
        "Realizamos los mismos pasos para crear otras tres capas"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lNI5QcEUYUNy"
      },
      "source": [
        "model.add(LSTM(units=50, return_sequences=True))\n",
        "model.add(Dropout(0.2))\n",
        "\n",
        "model.add(LSTM(units=50, return_sequences=True))\n",
        "model.add(Dropout(0.2))\n",
        "\n",
        "model.add(LSTM(units=50))\n",
        "model.add(Dropout(0.2))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pQ7obrONlO4z"
      },
      "source": [
        "y finalmente añadimos una capa densa y como solo deseamos predecir un valor dados los 60 valores previos, el valor de units=1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ok_m989sYUN1"
      },
      "source": [
        "model.add(Dense(units = 1))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FaOK3Lw0lx0n"
      },
      "source": [
        "#### Compilación del modelo"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jaqtH7KTl5-g"
      },
      "source": [
        "El paso previo al entrenamiento es compilar el modelo"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l0qeU87uYUN3"
      },
      "source": [
        "model.compile(optimizer = 'adam', loss = 'mean_squared_error')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hDOWE6p2mADa"
      },
      "source": [
        "En este caso usamos la funcion de perdida del cuadrado medio del error y el optimizador escogido para reducir dicha función de perdida es \"adam\""
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9n6CcvtLnD6t"
      },
      "source": [
        "#### Entrenamiento"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3Gn2GK9SnJDN"
      },
      "source": [
        "Por último, para entrenar el modelo utilizamos la función \"fit\" con los siguientes parámetros\n",
        "\n",
        "Este proceso puede tradar unos minutos dependiendo la maquina que lo ejecute"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wufDlRkmYUN6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "99a3ee44-ee15-4d9b-9702-fe76b0672e50"
      },
      "source": [
        "history=model.fit(features_train, labels_train, epochs = 100, batch_size = 32,validation_split=0.1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/100\n",
            "41/41 [==============================] - 6s 143ms/step - loss: 0.0201 - val_loss: 0.0053\n",
            "Epoch 2/100\n",
            "41/41 [==============================] - 5s 113ms/step - loss: 0.0047 - val_loss: 0.0018\n",
            "Epoch 3/100\n",
            "41/41 [==============================] - 5s 113ms/step - loss: 0.0035 - val_loss: 0.0016\n",
            "Epoch 4/100\n",
            "41/41 [==============================] - 5s 111ms/step - loss: 0.0033 - val_loss: 0.0018\n",
            "Epoch 5/100\n",
            "41/41 [==============================] - 5s 112ms/step - loss: 0.0025 - val_loss: 0.0015\n",
            "Epoch 6/100\n",
            "41/41 [==============================] - 4s 109ms/step - loss: 0.0026 - val_loss: 0.0019\n",
            "Epoch 7/100\n",
            "41/41 [==============================] - 5s 111ms/step - loss: 0.0026 - val_loss: 0.0016\n",
            "Epoch 8/100\n",
            "41/41 [==============================] - 4s 108ms/step - loss: 0.0024 - val_loss: 0.0017\n",
            "Epoch 9/100\n",
            "41/41 [==============================] - 5s 110ms/step - loss: 0.0022 - val_loss: 0.0014\n",
            "Epoch 10/100\n",
            "41/41 [==============================] - 4s 108ms/step - loss: 0.0022 - val_loss: 0.0021\n",
            "Epoch 11/100\n",
            "41/41 [==============================] - 5s 113ms/step - loss: 0.0025 - val_loss: 0.0014\n",
            "Epoch 12/100\n",
            "41/41 [==============================] - 4s 109ms/step - loss: 0.0021 - val_loss: 0.0028\n",
            "Epoch 13/100\n",
            "41/41 [==============================] - 5s 110ms/step - loss: 0.0026 - val_loss: 0.0013\n",
            "Epoch 14/100\n",
            "41/41 [==============================] - 5s 111ms/step - loss: 0.0019 - val_loss: 0.0015\n",
            "Epoch 15/100\n",
            "41/41 [==============================] - 5s 115ms/step - loss: 0.0020 - val_loss: 0.0012\n",
            "Epoch 16/100\n",
            "41/41 [==============================] - 5s 112ms/step - loss: 0.0020 - val_loss: 0.0011\n",
            "Epoch 17/100\n",
            "41/41 [==============================] - 5s 113ms/step - loss: 0.0025 - val_loss: 0.0012\n",
            "Epoch 18/100\n",
            "41/41 [==============================] - 5s 115ms/step - loss: 0.0021 - val_loss: 0.0029\n",
            "Epoch 19/100\n",
            "41/41 [==============================] - 5s 115ms/step - loss: 0.0024 - val_loss: 0.0021\n",
            "Epoch 20/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 0.0019 - val_loss: 0.0012\n",
            "Epoch 21/100\n",
            "41/41 [==============================] - 5s 112ms/step - loss: 0.0019 - val_loss: 0.0010\n",
            "Epoch 22/100\n",
            "41/41 [==============================] - 5s 112ms/step - loss: 0.0017 - val_loss: 0.0010\n",
            "Epoch 23/100\n",
            "41/41 [==============================] - 5s 111ms/step - loss: 0.0017 - val_loss: 0.0012\n",
            "Epoch 24/100\n",
            "41/41 [==============================] - 5s 111ms/step - loss: 0.0018 - val_loss: 0.0011\n",
            "Epoch 25/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 0.0016 - val_loss: 0.0010\n",
            "Epoch 26/100\n",
            "41/41 [==============================] - 5s 113ms/step - loss: 0.0016 - val_loss: 0.0010\n",
            "Epoch 27/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 0.0016 - val_loss: 0.0013\n",
            "Epoch 28/100\n",
            "41/41 [==============================] - 5s 114ms/step - loss: 0.0017 - val_loss: 0.0014\n",
            "Epoch 29/100\n",
            "41/41 [==============================] - 5s 111ms/step - loss: 0.0018 - val_loss: 0.0011\n",
            "Epoch 30/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 0.0017 - val_loss: 9.7470e-04\n",
            "Epoch 31/100\n",
            "41/41 [==============================] - 5s 113ms/step - loss: 0.0017 - val_loss: 9.9768e-04\n",
            "Epoch 32/100\n",
            "41/41 [==============================] - 5s 118ms/step - loss: 0.0016 - val_loss: 9.4102e-04\n",
            "Epoch 33/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 0.0015 - val_loss: 9.5386e-04\n",
            "Epoch 34/100\n",
            "41/41 [==============================] - 5s 120ms/step - loss: 0.0014 - val_loss: 0.0010\n",
            "Epoch 35/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 0.0015 - val_loss: 0.0017\n",
            "Epoch 36/100\n",
            "41/41 [==============================] - 5s 113ms/step - loss: 0.0014 - val_loss: 0.0011\n",
            "Epoch 37/100\n",
            "41/41 [==============================] - 5s 113ms/step - loss: 0.0016 - val_loss: 0.0010\n",
            "Epoch 38/100\n",
            "41/41 [==============================] - 5s 113ms/step - loss: 0.0014 - val_loss: 0.0012\n",
            "Epoch 39/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 0.0014 - val_loss: 9.2323e-04\n",
            "Epoch 40/100\n",
            "41/41 [==============================] - 5s 124ms/step - loss: 0.0013 - val_loss: 8.5957e-04\n",
            "Epoch 41/100\n",
            "41/41 [==============================] - 5s 115ms/step - loss: 0.0015 - val_loss: 9.2181e-04\n",
            "Epoch 42/100\n",
            "41/41 [==============================] - 5s 115ms/step - loss: 0.0015 - val_loss: 8.6040e-04\n",
            "Epoch 43/100\n",
            "41/41 [==============================] - 5s 115ms/step - loss: 0.0013 - val_loss: 9.6356e-04\n",
            "Epoch 44/100\n",
            "41/41 [==============================] - 5s 119ms/step - loss: 0.0013 - val_loss: 8.8938e-04\n",
            "Epoch 45/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 0.0015 - val_loss: 9.0255e-04\n",
            "Epoch 46/100\n",
            "41/41 [==============================] - 5s 119ms/step - loss: 0.0013 - val_loss: 8.9382e-04\n",
            "Epoch 47/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 0.0014 - val_loss: 0.0011\n",
            "Epoch 48/100\n",
            "41/41 [==============================] - 5s 123ms/step - loss: 0.0013 - val_loss: 8.5453e-04\n",
            "Epoch 49/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 0.0013 - val_loss: 7.9926e-04\n",
            "Epoch 50/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 0.0012 - val_loss: 7.5333e-04\n",
            "Epoch 51/100\n",
            "41/41 [==============================] - 5s 125ms/step - loss: 0.0011 - val_loss: 0.0011\n",
            "Epoch 52/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 0.0015 - val_loss: 0.0014\n",
            "Epoch 53/100\n",
            "41/41 [==============================] - 5s 119ms/step - loss: 0.0012 - val_loss: 7.5729e-04\n",
            "Epoch 54/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 0.0013 - val_loss: 8.9185e-04\n",
            "Epoch 55/100\n",
            "41/41 [==============================] - 5s 115ms/step - loss: 0.0013 - val_loss: 7.0923e-04\n",
            "Epoch 56/100\n",
            "41/41 [==============================] - 5s 118ms/step - loss: 0.0013 - val_loss: 8.7379e-04\n",
            "Epoch 57/100\n",
            "41/41 [==============================] - 5s 119ms/step - loss: 0.0011 - val_loss: 7.5400e-04\n",
            "Epoch 58/100\n",
            "41/41 [==============================] - 5s 119ms/step - loss: 0.0012 - val_loss: 6.7907e-04\n",
            "Epoch 59/100\n",
            "41/41 [==============================] - 5s 118ms/step - loss: 0.0012 - val_loss: 7.5551e-04\n",
            "Epoch 60/100\n",
            "41/41 [==============================] - 5s 114ms/step - loss: 0.0012 - val_loss: 7.5623e-04\n",
            "Epoch 61/100\n",
            "41/41 [==============================] - 5s 118ms/step - loss: 0.0011 - val_loss: 6.4023e-04\n",
            "Epoch 62/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 0.0011 - val_loss: 0.0012\n",
            "Epoch 63/100\n",
            "41/41 [==============================] - 5s 118ms/step - loss: 0.0011 - val_loss: 8.3026e-04\n",
            "Epoch 64/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 0.0013 - val_loss: 9.7587e-04\n",
            "Epoch 65/100\n",
            "41/41 [==============================] - 5s 118ms/step - loss: 0.0012 - val_loss: 7.4616e-04\n",
            "Epoch 66/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 0.0013 - val_loss: 0.0012\n",
            "Epoch 67/100\n",
            "41/41 [==============================] - 5s 119ms/step - loss: 0.0011 - val_loss: 6.7632e-04\n",
            "Epoch 68/100\n",
            "41/41 [==============================] - 5s 122ms/step - loss: 0.0012 - val_loss: 6.4289e-04\n",
            "Epoch 69/100\n",
            "41/41 [==============================] - 5s 122ms/step - loss: 0.0012 - val_loss: 5.9626e-04\n",
            "Epoch 70/100\n",
            "41/41 [==============================] - 5s 113ms/step - loss: 0.0011 - val_loss: 5.7341e-04\n",
            "Epoch 71/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 0.0011 - val_loss: 5.5581e-04\n",
            "Epoch 72/100\n",
            "41/41 [==============================] - 5s 118ms/step - loss: 0.0012 - val_loss: 5.6354e-04\n",
            "Epoch 73/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 0.0010 - val_loss: 9.7551e-04\n",
            "Epoch 74/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 0.0011 - val_loss: 6.4844e-04\n",
            "Epoch 75/100\n",
            "41/41 [==============================] - 5s 115ms/step - loss: 0.0014 - val_loss: 0.0011\n",
            "Epoch 76/100\n",
            "41/41 [==============================] - 6s 158ms/step - loss: 0.0011 - val_loss: 5.8189e-04\n",
            "Epoch 77/100\n",
            "41/41 [==============================] - 7s 163ms/step - loss: 9.9644e-04 - val_loss: 0.0010\n",
            "Epoch 78/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 9.5185e-04 - val_loss: 6.1684e-04\n",
            "Epoch 79/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 9.8770e-04 - val_loss: 5.1210e-04\n",
            "Epoch 80/100\n",
            "41/41 [==============================] - 5s 115ms/step - loss: 0.0012 - val_loss: 5.1974e-04\n",
            "Epoch 81/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 0.0011 - val_loss: 5.4689e-04\n",
            "Epoch 82/100\n",
            "41/41 [==============================] - 5s 118ms/step - loss: 9.9074e-04 - val_loss: 5.1928e-04\n",
            "Epoch 83/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 9.0392e-04 - val_loss: 5.2439e-04\n",
            "Epoch 84/100\n",
            "41/41 [==============================] - 5s 120ms/step - loss: 9.6835e-04 - val_loss: 5.5748e-04\n",
            "Epoch 85/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 0.0010 - val_loss: 6.2285e-04\n",
            "Epoch 86/100\n",
            "41/41 [==============================] - 5s 119ms/step - loss: 0.0011 - val_loss: 5.7730e-04\n",
            "Epoch 87/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 0.0010 - val_loss: 8.0736e-04\n",
            "Epoch 88/100\n",
            "41/41 [==============================] - 5s 118ms/step - loss: 0.0011 - val_loss: 6.2022e-04\n",
            "Epoch 89/100\n",
            "41/41 [==============================] - 5s 125ms/step - loss: 8.5145e-04 - val_loss: 4.6806e-04\n",
            "Epoch 90/100\n",
            "41/41 [==============================] - 5s 119ms/step - loss: 8.8789e-04 - val_loss: 4.7447e-04\n",
            "Epoch 91/100\n",
            "41/41 [==============================] - 5s 125ms/step - loss: 8.6393e-04 - val_loss: 5.9427e-04\n",
            "Epoch 92/100\n",
            "41/41 [==============================] - 5s 117ms/step - loss: 0.0010 - val_loss: 5.5157e-04\n",
            "Epoch 93/100\n",
            "41/41 [==============================] - 5s 119ms/step - loss: 8.5601e-04 - val_loss: 5.5441e-04\n",
            "Epoch 94/100\n",
            "41/41 [==============================] - 5s 124ms/step - loss: 9.0957e-04 - val_loss: 0.0012\n",
            "Epoch 95/100\n",
            "41/41 [==============================] - 5s 120ms/step - loss: 0.0010 - val_loss: 4.8883e-04\n",
            "Epoch 96/100\n",
            "41/41 [==============================] - 5s 122ms/step - loss: 9.4639e-04 - val_loss: 0.0011\n",
            "Epoch 97/100\n",
            "41/41 [==============================] - 5s 118ms/step - loss: 0.0011 - val_loss: 8.4118e-04\n",
            "Epoch 98/100\n",
            "41/41 [==============================] - 5s 116ms/step - loss: 9.0101e-04 - val_loss: 4.7507e-04\n",
            "Epoch 99/100\n",
            "41/41 [==============================] - 5s 115ms/step - loss: 8.9532e-04 - val_loss: 4.3806e-04\n",
            "Epoch 100/100\n",
            "41/41 [==============================] - 5s 118ms/step - loss: 8.3032e-04 - val_loss: 4.4589e-04\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "86kUx7Lhv5Jb"
      },
      "source": [
        "Para graficar la función de perdida ejecutamos el siguiente código"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ahqJx13ua7tF",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "e1210034-d7e4-4b7e-f368-60f311ddad8e"
      },
      "source": [
        "plt.figure()\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.title('Model loss')\n",
        "plt.ylabel('Loss')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['Train', 'Test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "biEaQKqgwFE0"
      },
      "source": [
        "En la grafica anterior se nota el rapido descenso de la función de perdida, lo que nos sugiere que el modelo se esta ajustando bien a los datos y por ende, el error cuadratico medio está disminuyendo"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "v0Lwj141p591"
      },
      "source": [
        "#### Testeo de nuestro modelo LSTM"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_Eyj6OpVp-AN"
      },
      "source": [
        "Hemos entrenado exitosamente nuestro modelo, nuestro siguiente paso será observar que tan bien se ajusta a datos que no estuvieron presentes  en el entrenamiento. En nuestro caso nuestros datos de testeo serán el valor de la apertura de las acciones de Netflix los 162 días posteriores a los datos de entrenamiento.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n2W8YwL8rWPS"
      },
      "source": [
        "Como requerimos de los 60 registros previos para poder predecir en nuestro modelo, el siguiente paso es añadir nuestros 60 datos previos al primer valor de testeo"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wVe8u7Y1YUOB"
      },
      "source": [
        "test_inputs =nt_training_scaled[len(nt_training_scaled) - len(features_test) - 60:]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S_Z_xXgT-ZlH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "410133f0-a38e-4141-ee85-a59453403f3c"
      },
      "source": [
        "len(test_inputs)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "222"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ks0hG5-9vfau"
      },
      "source": [
        "Creamos nuestro conjunto de 60 datos para los test_feature, los convertimos en matrices y los escribimos en su formato de tercera dimensión"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cRkS_ZweYUOF"
      },
      "source": [
        "test_features = []\n",
        "for i in range(60,222):\n",
        "    test_features.append(test_inputs[i-60:i, 0])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gZHLv8_SYUOI"
      },
      "source": [
        "test_features = np.array(test_features)\n",
        "test_features = np.reshape(test_features, (test_features.shape[0], test_features.shape[1], 1))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iaVnLOguv5iC"
      },
      "source": [
        "#### Predicciones"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MSrIuSbDwBZ_"
      },
      "source": [
        "Ya que hemos preprocesado nuestros datos de testeo los podemos usar para realizar predicciones, basta con hacer uso de la función predict en el modelo que creamos."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pK6Aop6IYUOL"
      },
      "source": [
        "predictions = model.predict(test_features)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GQOcc_MkwhHf"
      },
      "source": [
        "Como los datos estuvieron escalados al entrar al modelo, su predicción también estará escalada por lo que ejecutamos la función inverse_transform como sigue"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QC10fs_qYUOO"
      },
      "source": [
        "predictions = scaler.inverse_transform(predictions)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BGUZ5C8mAJEt"
      },
      "source": [
        "label_test1=nt_processed[1515:1677]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qmi_i10MxAkP"
      },
      "source": [
        "Finalmente miremos qué tan bien está prediciendo nuestro modelo"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yxt5sbwbYUOQ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 404
        },
        "outputId": "e66a7030-9bdf-43ab-aeaa-c46a692a6b06"
      },
      "source": [
        "plt.figure(figsize=(10,6))\n",
        "plt.plot(label_test1, color='blue', label='Actual Netflix Stock Price')\n",
        "plt.plot(predictions , color='red', label='Predicted Netflix Stock Price')\n",
        "plt.title('Netflix Stock Price Prediction')\n",
        "plt.xlabel('Date')\n",
        "plt.ylabel('Netflix Stock Price')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sUk1nr8mjKcs",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "bddb00eb-0036-4154-c111-5bbaa7089e82"
      },
      "source": [
        "error=mean_squared_error(label_test1,predictions)\n",
        "print(\"El error es\", error, predictions.shape, label_test1.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "El error es 212.66488348209447 (162, 1) (162, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZgaG4GmVbPEQ"
      },
      "source": [
        "Podemos apreciar que el modelo logra capturar gran parte del comportamiento del modelo "
      ]
    }
  ]
}