stock price prediction with lstm.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "6734dc27",
   "metadata": {},
   "source": [
    "Herkesi Merhabalar. bu yazımızda zaman serilerin de en çok kullanılan modellerden biri olan lstm model ile ileriye yönelik tahminler yapmaya çalışacağız. Şöyle bir fark ile ki lstm model ile bir gün değil ileriye yönelik 7 gün tahmini yapmaya çalışacağız. \n",
    "Data seti olarak yahoo finance ile çektiğimiz bir bankanın 2000 yılından itibaren bu zamana kadar olan verisini kullanacağız."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "d6a896cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.metrics import r2_score, mean_absolute_error, mean_squared_error \n",
    "import yfinance as yf\n",
    "import mplfinance as mpf\n",
    "\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6a3f1f04",
   "metadata": {},
   "source": [
    "# Read the Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "9c555681",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[*********************100%***********************]  1 of 1 completed\n"
     ]
    },
    {
     "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>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",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2000-05-10</th>\n",
       "      <td>6.137479</td>\n",
       "      <td>6.410256</td>\n",
       "      <td>6.001091</td>\n",
       "      <td>6.137479</td>\n",
       "      <td>-138486.203125</td>\n",
       "      <td>1284067</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-05-11</th>\n",
       "      <td>6.137479</td>\n",
       "      <td>6.273867</td>\n",
       "      <td>5.864702</td>\n",
       "      <td>6.137479</td>\n",
       "      <td>-138486.203125</td>\n",
       "      <td>608846</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-05-12</th>\n",
       "      <td>6.273867</td>\n",
       "      <td>6.410256</td>\n",
       "      <td>6.001091</td>\n",
       "      <td>6.273867</td>\n",
       "      <td>-141563.656250</td>\n",
       "      <td>2283592</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-05-15</th>\n",
       "      <td>6.137479</td>\n",
       "      <td>6.273867</td>\n",
       "      <td>6.001091</td>\n",
       "      <td>6.137479</td>\n",
       "      <td>-138486.203125</td>\n",
       "      <td>122956</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-05-16</th>\n",
       "      <td>6.001091</td>\n",
       "      <td>6.137479</td>\n",
       "      <td>5.864702</td>\n",
       "      <td>6.001091</td>\n",
       "      <td>-135408.765625</td>\n",
       "      <td>1331524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2022-07-18</th>\n",
       "      <td>8.030000</td>\n",
       "      <td>8.290000</td>\n",
       "      <td>8.030000</td>\n",
       "      <td>8.270000</td>\n",
       "      <td>8.270000</td>\n",
       "      <td>87714847</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2022-07-19</th>\n",
       "      <td>8.240000</td>\n",
       "      <td>8.540000</td>\n",
       "      <td>8.180000</td>\n",
       "      <td>8.500000</td>\n",
       "      <td>8.500000</td>\n",
       "      <td>106346801</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2022-07-20</th>\n",
       "      <td>8.540000</td>\n",
       "      <td>8.570000</td>\n",
       "      <td>8.400000</td>\n",
       "      <td>8.560000</td>\n",
       "      <td>8.560000</td>\n",
       "      <td>109726652</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2022-07-21</th>\n",
       "      <td>8.580000</td>\n",
       "      <td>8.670000</td>\n",
       "      <td>8.480000</td>\n",
       "      <td>8.580000</td>\n",
       "      <td>8.580000</td>\n",
       "      <td>85802698</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2022-07-22</th>\n",
       "      <td>8.620000</td>\n",
       "      <td>8.710000</td>\n",
       "      <td>8.590000</td>\n",
       "      <td>8.640000</td>\n",
       "      <td>8.640000</td>\n",
       "      <td>59791572</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5579 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                Open      High       Low     Close      Adj Close     Volume\n",
       "Date                                                                        \n",
       "2000-05-10  6.137479  6.410256  6.001091  6.137479 -138486.203125    1284067\n",
       "2000-05-11  6.137479  6.273867  5.864702  6.137479 -138486.203125     608846\n",
       "2000-05-12  6.273867  6.410256  6.001091  6.273867 -141563.656250    2283592\n",
       "2000-05-15  6.137479  6.273867  6.001091  6.137479 -138486.203125     122956\n",
       "2000-05-16  6.001091  6.137479  5.864702  6.001091 -135408.765625    1331524\n",
       "...              ...       ...       ...       ...            ...        ...\n",
       "2022-07-18  8.030000  8.290000  8.030000  8.270000       8.270000   87714847\n",
       "2022-07-19  8.240000  8.540000  8.180000  8.500000       8.500000  106346801\n",
       "2022-07-20  8.540000  8.570000  8.400000  8.560000       8.560000  109726652\n",
       "2022-07-21  8.580000  8.670000  8.480000  8.580000       8.580000   85802698\n",
       "2022-07-22  8.620000  8.710000  8.590000  8.640000       8.640000   59791572\n",
       "\n",
       "[5579 rows x 6 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import yfinance as yf\n",
    "data_df = yf.download(tickers=\"AKBNK.IS\", end=\"2022-07-25\")\n",
    "data_df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "74a29bfc",
   "metadata": {},
   "source": [
    "yahoo finance apisi bize dünya üzerinde ki  bir çok ülke ki  hisse senetlerinin ve para birimlerinin verisini 1 satırlık kod ile çekmemizi sağlıyor.\n",
    "download metodunun bazı paramaetreleri var. \n",
    "* tickers paramatresine hisse senedi adını yazıyoruz.\n",
    "* start paramaetresi ile istediğimiz tarihten itibaren verileri getirebiliyoruz.\n",
    "* end parametresi ise istedğimiz tarihe kadar olan verileri çekmemize yardımcı oluyor.\n",
    "\n",
    "burada en eski tarihten itibaren olan verileri getirmesini istediğim için start paramatresine herhangi bir tarih vermiyorum"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ba5f990b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "DatetimeIndex: 5579 entries, 2000-05-10 to 2022-07-22\n",
      "Data columns (total 6 columns):\n",
      " #   Column     Non-Null Count  Dtype  \n",
      "---  ------     --------------  -----  \n",
      " 0   Open       5579 non-null   float64\n",
      " 1   High       5579 non-null   float64\n",
      " 2   Low        5579 non-null   float64\n",
      " 3   Close      5579 non-null   float64\n",
      " 4   Adj Close  5579 non-null   float64\n",
      " 5   Volume     5579 non-null   int64  \n",
      "dtypes: float64(5), int64(1)\n",
      "memory usage: 305.1 KB\n"
     ]
    }
   ],
   "source": [
    "data_df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8bc49fe0",
   "metadata": {},
   "source": [
    "elimizde akbanka ait 6 sütün 5579 satırlık veri var. \n",
    "volume int diğerleri ise float veri tipinde."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "f3c8a9cc",
   "metadata": {},
   "outputs": [
    {
     "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Open</th>\n",
       "      <td>5579.0</td>\n",
       "      <td>5.058235e+00</td>\n",
       "      <td>2.373236e+00</td>\n",
       "      <td>0.362677</td>\n",
       "      <td>3.344045e+00</td>\n",
       "      <td>5.693437e+00</td>\n",
       "      <td>6.808116e+00</td>\n",
       "      <td>9.603388e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>High</th>\n",
       "      <td>5579.0</td>\n",
       "      <td>5.125646e+00</td>\n",
       "      <td>2.397054e+00</td>\n",
       "      <td>0.374376</td>\n",
       "      <td>3.426204e+00</td>\n",
       "      <td>5.760000e+00</td>\n",
       "      <td>6.885286e+00</td>\n",
       "      <td>9.650000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Low</th>\n",
       "      <td>5579.0</td>\n",
       "      <td>4.986654e+00</td>\n",
       "      <td>2.346123e+00</td>\n",
       "      <td>0.327579</td>\n",
       "      <td>3.279736e+00</td>\n",
       "      <td>5.594845e+00</td>\n",
       "      <td>6.722372e+00</td>\n",
       "      <td>9.480000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Close</th>\n",
       "      <td>5579.0</td>\n",
       "      <td>5.052858e+00</td>\n",
       "      <td>2.369930e+00</td>\n",
       "      <td>0.362677</td>\n",
       "      <td>3.344045e+00</td>\n",
       "      <td>5.691308e+00</td>\n",
       "      <td>6.808116e+00</td>\n",
       "      <td>9.540000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Adj Close</th>\n",
       "      <td>5579.0</td>\n",
       "      <td>-3.649806e+03</td>\n",
       "      <td>1.063866e+04</td>\n",
       "      <td>-141563.656250</td>\n",
       "      <td>2.416744e+00</td>\n",
       "      <td>4.824937e+00</td>\n",
       "      <td>5.899675e+00</td>\n",
       "      <td>9.540000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Volume</th>\n",
       "      <td>5579.0</td>\n",
       "      <td>2.597412e+07</td>\n",
       "      <td>3.863686e+07</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>6.855098e+06</td>\n",
       "      <td>1.352965e+07</td>\n",
       "      <td>2.751291e+07</td>\n",
       "      <td>5.006192e+08</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            count          mean           std            min           25%  \\\n",
       "Open       5579.0  5.058235e+00  2.373236e+00       0.362677  3.344045e+00   \n",
       "High       5579.0  5.125646e+00  2.397054e+00       0.374376  3.426204e+00   \n",
       "Low        5579.0  4.986654e+00  2.346123e+00       0.327579  3.279736e+00   \n",
       "Close      5579.0  5.052858e+00  2.369930e+00       0.362677  3.344045e+00   \n",
       "Adj Close  5579.0 -3.649806e+03  1.063866e+04 -141563.656250  2.416744e+00   \n",
       "Volume     5579.0  2.597412e+07  3.863686e+07       0.000000  6.855098e+06   \n",
       "\n",
       "                    50%           75%           max  \n",
       "Open       5.693437e+00  6.808116e+00  9.603388e+00  \n",
       "High       5.760000e+00  6.885286e+00  9.650000e+00  \n",
       "Low        5.594845e+00  6.722372e+00  9.480000e+00  \n",
       "Close      5.691308e+00  6.808116e+00  9.540000e+00  \n",
       "Adj Close  4.824937e+00  5.899675e+00  9.540000e+00  \n",
       "Volume     1.352965e+07  2.751291e+07  5.006192e+08  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_df.describe().T"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "609da5fe",
   "metadata": {},
   "source": [
    "veri setimize daha detaylı göz atıyoruz. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "761a4d05",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot  as plt\n",
    "fig = plt.figure(figsize = (20, 7))\n",
    "plt.plot(data_df[\"Close\"])\n",
    "plt.title(\"Akbank Hisse Senedi Kapanış Fiyatı\")\n",
    "plt.xlabel(\"Zaman\")\n",
    "plt.ylabel(\"Kapanış Fiyatı\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5f84e83a",
   "metadata": {},
   "source": [
    "Görüldüğü gibi 2000 yılı içerisinde hisse senedi ani bir düşüşle 1 tl nin altına kadar düşüyor. Ancak 2006- 2007 yıllarından itibaren günümüze kadar benzer şekidel hareket ettiğini gözlemliyoruz. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "617fbbd7",
   "metadata": {},
   "outputs": [
    {
     "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>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>Open</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.999555</td>\n",
       "      <td>0.999519</td>\n",
       "      <td>0.999275</td>\n",
       "      <td>0.510340</td>\n",
       "      <td>0.276211</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>High</th>\n",
       "      <td>0.999555</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.999371</td>\n",
       "      <td>0.999532</td>\n",
       "      <td>0.510102</td>\n",
       "      <td>0.282605</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Low</th>\n",
       "      <td>0.999519</td>\n",
       "      <td>0.999371</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.999586</td>\n",
       "      <td>0.510698</td>\n",
       "      <td>0.271763</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Close</th>\n",
       "      <td>0.999275</td>\n",
       "      <td>0.999532</td>\n",
       "      <td>0.999586</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.510273</td>\n",
       "      <td>0.277546</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Adj Close</th>\n",
       "      <td>0.510340</td>\n",
       "      <td>0.510102</td>\n",
       "      <td>0.510698</td>\n",
       "      <td>0.510273</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.141757</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Volume</th>\n",
       "      <td>0.276211</td>\n",
       "      <td>0.282605</td>\n",
       "      <td>0.271763</td>\n",
       "      <td>0.277546</td>\n",
       "      <td>0.141757</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Open      High       Low     Close  Adj Close    Volume\n",
       "Open       1.000000  0.999555  0.999519  0.999275   0.510340  0.276211\n",
       "High       0.999555  1.000000  0.999371  0.999532   0.510102  0.282605\n",
       "Low        0.999519  0.999371  1.000000  0.999586   0.510698  0.271763\n",
       "Close      0.999275  0.999532  0.999586  1.000000   0.510273  0.277546\n",
       "Adj Close  0.510340  0.510102  0.510698  0.510273   1.000000  0.141757\n",
       "Volume     0.276211  0.282605  0.271763  0.277546   0.141757  1.000000"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "kolerasyon = data_df.corr()\n",
    "kolerasyon"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "99dbb4a9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(kolerasyon, annot = True)\n",
    "plt.title(\"Kolerasyon Haritası\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a0d05d49",
   "metadata": {},
   "source": [
    "Görüldüğü gibi close değişkenimiziz açılış, kapanış, düşük ve yüksek değerlerle arasında ki kolerasyon 1 yani yüksek pozitif yönlü bir ilişki var."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4caee1ac",
   "metadata": {},
   "source": [
    "# Bağımlı ve Bağımsız Değişken Seçimi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "3df0a397",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2000-05-10    6.137479\n",
       "2000-05-11    6.137479\n",
       "2000-05-12    6.273867\n",
       "2000-05-15    6.137479\n",
       "2000-05-16    6.001091\n",
       "Name: Close, dtype: float64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "close = data_df[\"Close\"]\n",
    "close.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "9c0b5714",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5579,)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "close.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "a03ffb48",
   "metadata": {},
   "outputs": [],
   "source": [
    "#close = close[:5500]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "d1806697",
   "metadata": {},
   "outputs": [],
   "source": [
    "def to_x_y(frame, time_step, b):\n",
    "    df_as_numpy = frame.values\n",
    "    X = []\n",
    "    y = []\n",
    "    for i in range(len(df_as_numpy)- time_step):\n",
    "        row = [[a for a in df_as_numpy[i:i+time_step]]]\n",
    "        label = [a for a in df_as_numpy[i+time_step: i+time_step+b ]]\n",
    "        X.append(row)\n",
    "        y.append(label)\n",
    "    return np.array(X), np.array(y)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "575bdab5",
   "metadata": {},
   "outputs": [],
   "source": [
    "time_step = 15\n",
    "b = 7\n",
    "X, y = to_x_y(close, time_step, b)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "474845ed",
   "metadata": {},
   "source": [
    "yazdığımız bu fonksiyon yardımı ile bağımı ve bağımısz değişkenimizi oluşturacağız. Bu fonksiyonun 3 parametresi var. birincisi kapanış fiyatı değerlerimiz ikincisi ise time_step yani eğitim setinde geçmişe dönük kaç günlük veri istiyoruz. 3. ise ileriye dönük kaç günlük tahmin almak istiyoruz. Biz bu çalışmada geçmişe dönüş 15 günlük veri ile ileriye dönük 7 günlük tahmin yapacağız. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "55f4c0f0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[8.31000042, 8.02000046, 8.02000046, 8.15999985, 7.94999981,\n",
       "        7.80000019, 7.90999985, 8.06000042, 8.09000015, 7.96000004,\n",
       "        7.90999985, 8.27000046, 8.5       , 8.56000042, 8.57999992]])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X[-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "c890db7a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[8.640000343322754]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y[-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "55749114",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5564"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "370b3717",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 0.],\n",
       "       [0., 0., 0.],\n",
       "       [0., 0., 0.],\n",
       "       [0., 0., 0.],\n",
       "       [0., 0., 0.]])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a = np.zeros((5,3))\n",
    "a"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "ee88360c",
   "metadata": {},
   "outputs": [],
   "source": [
    "y1 = pd.DataFrame(np.zeros((close.shape[0]-time_step, 7)), columns = [0, 1, 2, 3, 4, 5, 6])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "5b6f6fec",
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range((len(y))):\n",
    "    for j in range((len(y[i]))):\n",
    "        y1[j][i] = y[i][j]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "ca6b067e",
   "metadata": {},
   "outputs": [],
   "source": [
    "y1 = y1.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "6a2f550c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5564, 1, 15), (5564, 7))"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape, y1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "ab14c36e",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = X[:5557]\n",
    "y1 = y1[:5557]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "491d3bd4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5557, 7)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y1.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d92d1e68",
   "metadata": {},
   "source": [
    "### Train ve Test Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "8c6393c6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(4500, 1, 15) (4500, 7)\n",
      "(500, 1, 15) (500, 7)\n",
      "(557, 1, 15) (557, 7)\n"
     ]
    }
   ],
   "source": [
    "X_train, y_train = X[:4500], y1[:4500]\n",
    "X_val, y_val = X[4500:5000], y1[4500:5000]\n",
    "X_test, y_test = X[5000:], y1[5000:]\n",
    "print(X_train.shape, y_train.shape)\n",
    "print(X_val.shape, y_val.shape)\n",
    "print(X_test.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "62fa6012",
   "metadata": {},
   "source": [
    "Verimizin ilk 4500 değerini train seti, sonraki 500 değeri validation seti ve en son ki 564 değeri ise test seti olarak ayırıyoruz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "63233408",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, LSTM, Dropout\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "from keras.callbacks import ModelCheckpoint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "578adfcf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " lstm_1 (LSTM)               (None, 250)               266000    \n",
      "                                                                 \n",
      " dropout_1 (Dropout)         (None, 250)               0         \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 7)                 1757      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 267,757\n",
      "Trainable params: 267,757\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(LSTM(250, input_shape = (1,time_step), return_sequences = False))\n",
    "model.add(Dropout(0.3))\n",
    "model.add(Dense(b, activation=  \"linear\"))\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "75fbacd7",
   "metadata": {},
   "source": [
    "model yapımızı oluşturduk.burda lstm yapısı içinde 250 tane nöron kullanmayı tercih ettim. dropoutumuz 0.3. output olarak tahmin etmek istediğimiz gün sayısı kadar nöron kullanıyoruz yani yukarıda tanımladığımız b değeri."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "e793208b",
   "metadata": {},
   "outputs": [],
   "source": [
    "cp = ModelCheckpoint(\"model/\" ,save_best_only = True)\n",
    "model.compile(loss ='mean_squared_error', optimizer = \"Adam\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "7f1dc169",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "131/141 [==========================>...] - ETA: 0s - loss: 2.9543"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 4s 20ms/step - loss: 2.7760 - val_loss: 0.2749\n",
      "Epoch 2/100\n",
      "141/141 [==============================] - ETA: 0s - loss: 0.3044"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 3s 19ms/step - loss: 0.3044 - val_loss: 0.1934\n",
      "Epoch 3/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.2215 - val_loss: 0.2181\n",
      "Epoch 4/100\n",
      "124/141 [=========================>....] - ETA: 0s - loss: 0.2061"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 2s 17ms/step - loss: 0.2053 - val_loss: 0.1657\n",
      "Epoch 5/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1830 - val_loss: 0.1769\n",
      "Epoch 6/100\n",
      "121/141 [========================>.....] - ETA: 0s - loss: 0.1669"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 3s 19ms/step - loss: 0.1684 - val_loss: 0.1436\n",
      "Epoch 7/100\n",
      "125/141 [=========================>....] - ETA: 0s - loss: 0.1575"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 2s 18ms/step - loss: 0.1580 - val_loss: 0.1391\n",
      "Epoch 8/100\n",
      "140/141 [============================>.] - ETA: 0s - loss: 0.1635"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 3s 18ms/step - loss: 0.1638 - val_loss: 0.1348\n",
      "Epoch 9/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1478 - val_loss: 0.1848\n",
      "Epoch 10/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1520 - val_loss: 0.1378\n",
      "Epoch 11/100\n",
      "140/141 [============================>.] - ETA: 0s - loss: 0.1446"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 2s 17ms/step - loss: 0.1449 - val_loss: 0.1343\n",
      "Epoch 12/100\n",
      "125/141 [=========================>....] - ETA: 0s - loss: 0.1477"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 3s 18ms/step - loss: 0.1483 - val_loss: 0.1330\n",
      "Epoch 13/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1436 - val_loss: 0.1336\n",
      "Epoch 14/100\n",
      "138/141 [============================>.] - ETA: 0s - loss: 0.1351"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 2s 18ms/step - loss: 0.1342 - val_loss: 0.1259\n",
      "Epoch 15/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1345 - val_loss: 0.1272\n",
      "Epoch 16/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1427 - val_loss: 0.1284\n",
      "Epoch 17/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1277 - val_loss: 0.1310\n",
      "Epoch 18/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1339 - val_loss: 0.1473\n",
      "Epoch 19/100\n",
      "134/141 [===========================>..] - ETA: 0s - loss: 0.1335"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 2s 16ms/step - loss: 0.1337 - val_loss: 0.1242\n",
      "Epoch 20/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1302 - val_loss: 0.1445\n",
      "Epoch 21/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1316 - val_loss: 0.1448\n",
      "Epoch 22/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1330 - val_loss: 0.1413\n",
      "Epoch 23/100\n",
      "131/141 [==========================>...] - ETA: 0s - loss: 0.1230"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 3s 18ms/step - loss: 0.1239 - val_loss: 0.1234\n",
      "Epoch 24/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1199 - val_loss: 0.1752\n",
      "Epoch 25/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1221 - val_loss: 0.1267\n",
      "Epoch 26/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1247 - val_loss: 0.1344\n",
      "Epoch 27/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1199 - val_loss: 0.1251\n",
      "Epoch 28/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1211 - val_loss: 0.1759\n",
      "Epoch 29/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1239 - val_loss: 0.1342\n",
      "Epoch 30/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1180 - val_loss: 0.1339\n",
      "Epoch 31/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1187 - val_loss: 0.1273\n",
      "Epoch 32/100\n",
      "130/141 [==========================>...] - ETA: 0s - loss: 0.1199"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 2s 17ms/step - loss: 0.1193 - val_loss: 0.1234\n",
      "Epoch 33/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1155 - val_loss: 0.1431\n",
      "Epoch 34/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1206 - val_loss: 0.1290\n",
      "Epoch 35/100\n",
      "127/141 [==========================>...] - ETA: 0s - loss: 0.1185"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 2s 17ms/step - loss: 0.1173 - val_loss: 0.1217\n",
      "Epoch 36/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1278 - val_loss: 0.1316\n",
      "Epoch 37/100\n",
      "128/141 [==========================>...] - ETA: 0s - loss: 0.1182"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 3s 18ms/step - loss: 0.1174 - val_loss: 0.1206\n",
      "Epoch 38/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1135 - val_loss: 0.1402\n",
      "Epoch 39/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1130 - val_loss: 0.1681\n",
      "Epoch 40/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1290 - val_loss: 0.1593\n",
      "Epoch 41/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1202 - val_loss: 0.1214\n",
      "Epoch 42/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1102 - val_loss: 0.1281\n",
      "Epoch 43/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1081 - val_loss: 0.1220\n",
      "Epoch 44/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1071 - val_loss: 0.2252\n",
      "Epoch 45/100\n",
      "131/141 [==========================>...] - ETA: 0s - loss: 0.1131"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 2s 17ms/step - loss: 0.1116 - val_loss: 0.1203\n",
      "Epoch 46/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1247 - val_loss: 0.1448\n",
      "Epoch 47/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1093 - val_loss: 0.1211\n",
      "Epoch 48/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1153 - val_loss: 0.1448\n",
      "Epoch 49/100\n",
      "126/141 [=========================>....] - ETA: 0s - loss: 0.1124"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 2s 16ms/step - loss: 0.1164 - val_loss: 0.1202\n",
      "Epoch 50/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1137 - val_loss: 0.1365\n",
      "Epoch 51/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1094 - val_loss: 0.1262\n",
      "Epoch 52/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1142 - val_loss: 0.1213\n",
      "Epoch 53/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1057 - val_loss: 0.1261\n",
      "Epoch 54/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1177 - val_loss: 0.2088\n",
      "Epoch 55/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1073 - val_loss: 0.1772\n",
      "Epoch 56/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1189 - val_loss: 0.1252\n",
      "Epoch 57/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1107 - val_loss: 0.1347\n",
      "Epoch 58/100\n",
      "126/141 [=========================>....] - ETA: 0s - loss: 0.1026"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 3s 18ms/step - loss: 0.1036 - val_loss: 0.1187\n",
      "Epoch 59/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1079 - val_loss: 0.1422\n",
      "Epoch 60/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1126 - val_loss: 0.1683\n",
      "Epoch 61/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1100 - val_loss: 0.1679\n",
      "Epoch 62/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1094 - val_loss: 0.1197\n",
      "Epoch 63/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1059 - val_loss: 0.1224\n",
      "Epoch 64/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1077 - val_loss: 0.1736\n",
      "Epoch 65/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1006 - val_loss: 0.1363\n",
      "Epoch 66/100\n",
      "129/141 [==========================>...] - ETA: 0s - loss: 0.1014"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 2s 17ms/step - loss: 0.1031 - val_loss: 0.1182\n",
      "Epoch 67/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1131 - val_loss: 0.1253\n",
      "Epoch 68/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1101 - val_loss: 0.1244\n",
      "Epoch 69/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1133 - val_loss: 0.1259\n",
      "Epoch 70/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1099 - val_loss: 0.1208\n",
      "Epoch 71/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1126 - val_loss: 0.1717\n",
      "Epoch 72/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1153 - val_loss: 0.1464\n",
      "Epoch 73/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1074 - val_loss: 0.1238\n",
      "Epoch 74/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1038 - val_loss: 0.1450\n",
      "Epoch 75/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1067 - val_loss: 0.1431\n",
      "Epoch 76/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1025 - val_loss: 0.1469\n",
      "Epoch 77/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1004 - val_loss: 0.1783\n",
      "Epoch 78/100\n",
      "128/141 [==========================>...] - ETA: 0s - loss: 0.1110"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 2s 17ms/step - loss: 0.1095 - val_loss: 0.1178\n",
      "Epoch 79/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1078 - val_loss: 0.1204\n",
      "Epoch 80/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1021 - val_loss: 0.1181\n",
      "Epoch 81/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1085 - val_loss: 0.1187\n",
      "Epoch 82/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1057 - val_loss: 0.1537\n",
      "Epoch 83/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1031 - val_loss: 0.1455\n",
      "Epoch 84/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1002 - val_loss: 0.1226\n",
      "Epoch 85/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1004 - val_loss: 0.1202\n",
      "Epoch 86/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1002 - val_loss: 0.1187\n",
      "Epoch 87/100\n",
      "127/141 [==========================>...] - ETA: 0s - loss: 0.1039"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as lstm_cell_1_layer_call_fn, lstm_cell_1_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n",
      "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x000001F290423760> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "141/141 [==============================] - 3s 20ms/step - loss: 0.1027 - val_loss: 0.1174\n",
      "Epoch 88/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1041 - val_loss: 0.1184\n",
      "Epoch 89/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1067 - val_loss: 0.1185\n",
      "Epoch 90/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1097 - val_loss: 0.1401\n",
      "Epoch 91/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1027 - val_loss: 0.1207\n",
      "Epoch 92/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.0992 - val_loss: 0.1205\n",
      "Epoch 93/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1048 - val_loss: 0.1755\n",
      "Epoch 94/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1051 - val_loss: 0.1327\n",
      "Epoch 95/100\n",
      "141/141 [==============================] - 0s 3ms/step - loss: 0.1039 - val_loss: 0.1267\n",
      "Epoch 96/100\n",
      "141/141 [==============================] - 1s 3ms/step - loss: 0.0992 - val_loss: 0.1199\n",
      "Epoch 97/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1126 - val_loss: 0.1246\n",
      "Epoch 98/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1016 - val_loss: 0.1244\n",
      "Epoch 99/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1021 - val_loss: 0.1357\n",
      "Epoch 100/100\n",
      "141/141 [==============================] - 1s 4ms/step - loss: 0.1026 - val_loss: 0.2033\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train, y_train, validation_data = (X_val, y_val), epochs = 100, callbacks = [cp], batch_size = 32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "d6980aae",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "dictry = {\"Train Loss\":history.history[\"loss\"], \"Validation Loss\": history.history[\"val_loss\"]}\n",
    "plt.plot(dictry[\"Train Loss\"], label = \"Train Loss\", color = \"red\")\n",
    "plt.plot(dictry[\"Validation Loss\"], label = \"Validation Loss\", color = \"blue\")\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "777ccafe",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import load_model\n",
    "model = load_model(\"model/\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4dffe17f",
   "metadata": {},
   "source": [
    "kaydedilen modeli load ediyoruz."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "311eb3dd",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test_pred = model.predict(X_test)\n",
    "y_train_pred = model.predict(X_train)\n",
    "y_val_pred = model.predict(X_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2fc840db",
   "metadata": {},
   "source": [
    "Öncelikle Test setimizin ilk 7 günlük ardından test data üzerinden tahmin ettiğimiz ilk 7 günlük değerleri inceliyoruz."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "7b1e8e22",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5.78999996, 5.78999996, 5.67999983, 5.69000006, 5.73999977,\n",
       "       6.03999996, 5.92000008])"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "9dd57d57",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5.6547904, 5.6783485, 5.681908 , 5.698029 , 5.692752 , 5.68041  ,\n",
       "       5.7114553], dtype=float32)"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test_pred[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5d102049",
   "metadata": {},
   "outputs": [],
   "source": [
    "Görüldüğü gibi gerçek değerlerle tahmin ettiğimiz değerler birbirine oldukça yakın. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "7216bca5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(y_test, color = \"red\")\n",
    "plt.plot(y_test_pred, color = \"blue\")\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "114c2095",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(y_val, color = \"red\")\n",
    "plt.plot(y_val_pred, color = \"blue\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "2864b2a6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(y_train[2500:], color = \"red\")\n",
    "plt.plot(y_train_pred[2500:], color = \"blue\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4dbdc412",
   "metadata": {},
   "source": [
    "görüldğü gibi hem train hem validation hemde test setlerinde tahmin ve gerçek değerler birbirine çok yakın.."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a5f7b0b7",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test[:1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "fa373574",
   "metadata": {},
   "outputs": [],
   "source": [
    "r2_test = r2_score(y_test, y_test_pred)\n",
    "mae_test = mean_absolute_error(y_test, y_test_pred)\n",
    "mse_test = mean_squared_error(y_test, y_test_pred)\n",
    "rmse_test = np.sqrt(mean_squared_error(y_test, y_test_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "d68961cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "r2_train = r2_score(y_train, y_train_pred)\n",
    "mae_train = mean_absolute_error(y_train, y_train_pred)\n",
    "mse_train = mean_squared_error(y_train, y_train_pred)\n",
    "rmse_train = np.sqrt(mean_squared_error(y_train, y_train_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "c82b9817",
   "metadata": {},
   "outputs": [],
   "source": [
    "r2_val = r2_score(y_val, y_val_pred)\n",
    "mae_val = mean_absolute_error(y_val, y_val_pred)\n",
    "mse_val = mean_squared_error(y_val, y_val_pred)\n",
    "rmse_val = np.sqrt(mean_squared_error(y_val, y_val_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "1e09205b",
   "metadata": {},
   "outputs": [
    {
     "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>R2 Score</th>\n",
       "      <th>Mean Absolute Error</th>\n",
       "      <th>Mean Squared Error</th>\n",
       "      <th>Root Mean Squared Error</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Train Data</th>\n",
       "      <td>0.991394</td>\n",
       "      <td>0.159115</td>\n",
       "      <td>0.052109</td>\n",
       "      <td>0.228273</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Validation Data</th>\n",
       "      <td>0.866956</td>\n",
       "      <td>0.252589</td>\n",
       "      <td>0.117440</td>\n",
       "      <td>0.342695</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Test Data</th>\n",
       "      <td>0.927637</td>\n",
       "      <td>0.219324</td>\n",
       "      <td>0.108362</td>\n",
       "      <td>0.329184</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 R2 Score  Mean Absolute Error  Mean Squared Error  \\\n",
       "Train Data       0.991394             0.159115            0.052109   \n",
       "Validation Data  0.866956             0.252589            0.117440   \n",
       "Test Data        0.927637             0.219324            0.108362   \n",
       "\n",
       "                 Root Mean Squared Error  \n",
       "Train Data                      0.228273  \n",
       "Validation Data                 0.342695  \n",
       "Test Data                       0.329184  "
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scores = { \n",
    "    \"Train Data\" : {\"R2 Score\" : r2_train,\n",
    "                      \"Mean Absolute Error\" : mae_train,\n",
    "                      \"Mean Squared Error\" : mse_train,\n",
    "                    \"Root Mean Squared Error\" : rmse_train},\n",
    "        \"Validation Data\" : {\"R2 Score\" : r2_val,\n",
    "                      \"Mean Absolute Error\" : mae_val,\n",
    "                      \"Mean Squared Error\" : mse_val,\n",
    "                    \"Root Mean Squared Error\" : rmse_val},\n",
    "        \"Test Data \" : {\"R2 Score\" : r2_test,\n",
    "                      \"Mean Absolute Error\" : mae_test,\n",
    "                      \"Mean Squared Error\" : mse_test,\n",
    "                    \"Root Mean Squared Error\" : rmse_test} \n",
    "        }\n",
    "Scores = pd.DataFrame(scores)\n",
    "Scores.T"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "30890761",
   "metadata": {},
   "source": [
    "R2 scorlarımız ve Hata metriclerimiz tabloda. Modelimizde gayet iyi çalışıyor. Şu unutulmamalıdır ki modelimizin ileri doğru tahmin gün sayısı arttıkça  tarin ve validation setlerinde hatalarımız artacaktır."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}