{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "b8f36fba-e6c1-4f65-ae88-35e795d8e89e",
   "metadata": {},
   "source": [
    "Load the _Wine Quality Dataset (Combined)_ data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "96634688-0fce-478d-a570-edad08bd37bc",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "data = pd.read_csv(\"../WineQuality.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b7b0e979-1712-4422-9baa-4076c58828c9",
   "metadata": {},
   "source": [
    "Sample the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "642be8ad-0912-452d-b914-47189e60dd71",
   "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>Unnamed: 0</th>\n",
       "      <th>fixed acidity</th>\n",
       "      <th>volatile acidity</th>\n",
       "      <th>citric acid</th>\n",
       "      <th>residual sugar</th>\n",
       "      <th>chlorides</th>\n",
       "      <th>free sulfur dioxide</th>\n",
       "      <th>total sulfur dioxide</th>\n",
       "      <th>density</th>\n",
       "      <th>pH</th>\n",
       "      <th>sulphates</th>\n",
       "      <th>alcohol</th>\n",
       "      <th>quality</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2732</td>\n",
       "      <td>7.4</td>\n",
       "      <td>0.170</td>\n",
       "      <td>0.29</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.047</td>\n",
       "      <td>23.0</td>\n",
       "      <td>107.0</td>\n",
       "      <td>0.99390</td>\n",
       "      <td>3.52</td>\n",
       "      <td>0.65</td>\n",
       "      <td>10.4</td>\n",
       "      <td>6</td>\n",
       "      <td>White Wine</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2607</td>\n",
       "      <td>5.3</td>\n",
       "      <td>0.310</td>\n",
       "      <td>0.38</td>\n",
       "      <td>10.5</td>\n",
       "      <td>0.031</td>\n",
       "      <td>53.0</td>\n",
       "      <td>140.0</td>\n",
       "      <td>0.99321</td>\n",
       "      <td>3.34</td>\n",
       "      <td>0.46</td>\n",
       "      <td>11.7</td>\n",
       "      <td>6</td>\n",
       "      <td>White Wine</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1653</td>\n",
       "      <td>4.7</td>\n",
       "      <td>0.145</td>\n",
       "      <td>0.29</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.042</td>\n",
       "      <td>35.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>0.99080</td>\n",
       "      <td>3.76</td>\n",
       "      <td>0.49</td>\n",
       "      <td>11.3</td>\n",
       "      <td>6</td>\n",
       "      <td>White Wine</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3264</td>\n",
       "      <td>6.9</td>\n",
       "      <td>0.260</td>\n",
       "      <td>0.29</td>\n",
       "      <td>4.2</td>\n",
       "      <td>0.043</td>\n",
       "      <td>33.0</td>\n",
       "      <td>114.0</td>\n",
       "      <td>0.99020</td>\n",
       "      <td>3.16</td>\n",
       "      <td>0.31</td>\n",
       "      <td>12.5</td>\n",
       "      <td>6</td>\n",
       "      <td>White Wine</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4931</td>\n",
       "      <td>6.4</td>\n",
       "      <td>0.450</td>\n",
       "      <td>0.07</td>\n",
       "      <td>1.1</td>\n",
       "      <td>0.030</td>\n",
       "      <td>10.0</td>\n",
       "      <td>131.0</td>\n",
       "      <td>0.99050</td>\n",
       "      <td>2.97</td>\n",
       "      <td>0.28</td>\n",
       "      <td>10.8</td>\n",
       "      <td>5</td>\n",
       "      <td>White Wine</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",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32480</th>\n",
       "      <td>2838</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.255</td>\n",
       "      <td>0.22</td>\n",
       "      <td>2.7</td>\n",
       "      <td>0.043</td>\n",
       "      <td>46.0</td>\n",
       "      <td>153.0</td>\n",
       "      <td>0.99238</td>\n",
       "      <td>3.75</td>\n",
       "      <td>0.76</td>\n",
       "      <td>11.3</td>\n",
       "      <td>6</td>\n",
       "      <td>White Wine</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32481</th>\n",
       "      <td>6414</td>\n",
       "      <td>6.6</td>\n",
       "      <td>0.360</td>\n",
       "      <td>0.52</td>\n",
       "      <td>11.3</td>\n",
       "      <td>0.046</td>\n",
       "      <td>8.0</td>\n",
       "      <td>110.0</td>\n",
       "      <td>0.99660</td>\n",
       "      <td>3.07</td>\n",
       "      <td>0.46</td>\n",
       "      <td>9.4</td>\n",
       "      <td>5</td>\n",
       "      <td>White Wine</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32482</th>\n",
       "      <td>1126</td>\n",
       "      <td>6.3</td>\n",
       "      <td>0.200</td>\n",
       "      <td>0.24</td>\n",
       "      <td>1.7</td>\n",
       "      <td>0.052</td>\n",
       "      <td>36.0</td>\n",
       "      <td>135.0</td>\n",
       "      <td>0.99374</td>\n",
       "      <td>3.80</td>\n",
       "      <td>0.66</td>\n",
       "      <td>10.8</td>\n",
       "      <td>6</td>\n",
       "      <td>White Wine</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32483</th>\n",
       "      <td>2924</td>\n",
       "      <td>6.2</td>\n",
       "      <td>0.200</td>\n",
       "      <td>0.33</td>\n",
       "      <td>5.4</td>\n",
       "      <td>0.028</td>\n",
       "      <td>21.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>0.99012</td>\n",
       "      <td>3.36</td>\n",
       "      <td>0.41</td>\n",
       "      <td>13.5</td>\n",
       "      <td>7</td>\n",
       "      <td>White Wine</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32484</th>\n",
       "      <td>5462</td>\n",
       "      <td>8.1</td>\n",
       "      <td>0.280</td>\n",
       "      <td>0.46</td>\n",
       "      <td>15.4</td>\n",
       "      <td>0.059</td>\n",
       "      <td>32.0</td>\n",
       "      <td>177.0</td>\n",
       "      <td>1.00040</td>\n",
       "      <td>3.27</td>\n",
       "      <td>0.58</td>\n",
       "      <td>9.0</td>\n",
       "      <td>4</td>\n",
       "      <td>White Wine</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>32485 rows × 14 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       Unnamed: 0  fixed acidity  volatile acidity  citric acid  \\\n",
       "0            2732            7.4             0.170         0.29   \n",
       "1            2607            5.3             0.310         0.38   \n",
       "2            1653            4.7             0.145         0.29   \n",
       "3            3264            6.9             0.260         0.29   \n",
       "4            4931            6.4             0.450         0.07   \n",
       "...           ...            ...               ...          ...   \n",
       "32480        2838            5.0             0.255         0.22   \n",
       "32481        6414            6.6             0.360         0.52   \n",
       "32482        1126            6.3             0.200         0.24   \n",
       "32483        2924            6.2             0.200         0.33   \n",
       "32484        5462            8.1             0.280         0.46   \n",
       "\n",
       "       residual sugar  chlorides  free sulfur dioxide  total sulfur dioxide  \\\n",
       "0                 1.4      0.047                 23.0                 107.0   \n",
       "1                10.5      0.031                 53.0                 140.0   \n",
       "2                 1.0      0.042                 35.0                  90.0   \n",
       "3                 4.2      0.043                 33.0                 114.0   \n",
       "4                 1.1      0.030                 10.0                 131.0   \n",
       "...               ...        ...                  ...                   ...   \n",
       "32480             2.7      0.043                 46.0                 153.0   \n",
       "32481            11.3      0.046                  8.0                 110.0   \n",
       "32482             1.7      0.052                 36.0                 135.0   \n",
       "32483             5.4      0.028                 21.0                  75.0   \n",
       "32484            15.4      0.059                 32.0                 177.0   \n",
       "\n",
       "       density    pH  sulphates  alcohol  quality        Type  \n",
       "0      0.99390  3.52       0.65     10.4        6  White Wine  \n",
       "1      0.99321  3.34       0.46     11.7        6  White Wine  \n",
       "2      0.99080  3.76       0.49     11.3        6  White Wine  \n",
       "3      0.99020  3.16       0.31     12.5        6  White Wine  \n",
       "4      0.99050  2.97       0.28     10.8        5  White Wine  \n",
       "...        ...   ...        ...      ...      ...         ...  \n",
       "32480  0.99238  3.75       0.76     11.3        6  White Wine  \n",
       "32481  0.99660  3.07       0.46      9.4        5  White Wine  \n",
       "32482  0.99374  3.80       0.66     10.8        6  White Wine  \n",
       "32483  0.99012  3.36       0.41     13.5        7  White Wine  \n",
       "32484  1.00040  3.27       0.58      9.0        4  White Wine  \n",
       "\n",
       "[32485 rows x 14 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "09008208-91d4-49e0-806d-71461d5c3fdc",
   "metadata": {},
   "source": [
    "Retain the useful 11 features and isolate quality alone as a label, then split them both into training/test subsets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "1eb05609-e631-4a26-8ce6-d94dd6f25911",
   "metadata": {},
   "outputs": [],
   "source": [
    "x_data = data.drop(data.columns[0], axis=1).drop([\"quality\",\"Type\"], axis=1)\n",
    "y_data = data.quality"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "040a8912",
   "metadata": {},
   "source": [
    "Split the samples into 80/20 train/test subsets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "c8dea41a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using 25988 samples for training and 6497 samples for training.\n",
      "Total of 32485 records, dataset size is 32485 rows.\n",
      "Training set has a shape of (32485, 11), labels have a shape of (32485,)\n"
     ]
    }
   ],
   "source": [
    "train_size = int(len(x_data) * 0.8)\n",
    "x_train = x_data[:train_size]\n",
    "y_train = y_data[:train_size]\n",
    "x_test = x_data[train_size:]\n",
    "y_test = y_data[train_size:]\n",
    "\n",
    "print(\"Using {} samples for training and {} samples for training.\\n\".format(len(x_train), len(x_test)) +\n",
    "        \"Total of {} records, dataset size is {} rows.\\n\".format(len(x_train) + len(x_test), len(x_data)) +\n",
    "        \"Training set has a shape of {}, labels have a shape of {}\".format(x_data.shape, y_data.shape))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c4f5ae2f",
   "metadata": {},
   "source": [
    "Create a sequential model definition, one deep RELU intermediate layer with softmax output and 10 possible values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ecd1f4fd",
   "metadata": {},
   "outputs": [],
   "source": [
    "import keras\n",
    "from keras import layers\n",
    "\n",
    "classifier_init = keras.Sequential([\n",
    "    layers.Dense(11, activation=\"relu\"),\n",
    "    layers.Dense(44, activation=\"relu\"),\n",
    "    layers.Dense(10, activation=\"softmax\")\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "693762d0",
   "metadata": {},
   "source": [
    "Compile the model with adam optimiser and sparse categorical cross-entropy loss function. Track accuracy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "775ce661",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2025-04-16 20:56:16.639749: I metal_plugin/src/device/metal_device.cc:1154] Metal device set to: Apple M1 Max\n",
      "2025-04-16 20:56:16.639783: I metal_plugin/src/device/metal_device.cc:296] systemMemory: 64.00 GB\n",
      "2025-04-16 20:56:16.639792: I metal_plugin/src/device/metal_device.cc:313] maxCacheSize: 24.00 GB\n",
      "WARNING: All log messages before absl::InitializeLog() is called are written to STDERR\n",
      "I0000 00:00:1744826176.639810 59696050 pluggable_device_factory.cc:305] Could not identify NUMA node of platform GPU ID 0, defaulting to 0. Your kernel may not have been built with NUMA support.\n",
      "I0000 00:00:1744826176.639835 59696050 pluggable_device_factory.cc:271] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 0 MB memory) -> physical PluggableDevice (device: 0, name: METAL, pci bus id: <undefined>)\n"
     ]
    }
   ],
   "source": [
    "classifier_init.compile(optimizer=\"adam\",\n",
    "                   loss=\"sparse_categorical_crossentropy\",\n",
    "                   metrics=[\"accuracy\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "de4aee83",
   "metadata": {},
   "source": [
    "Fit the model across 20 epochs with batches of 1000 samples, using a further 80/20 split for training and validation subsets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "94db127a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2025-04-16 20:56:17.150147: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:117] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 44ms/step - accuracy: 0.0138 - loss: 32.4910 - val_accuracy: 0.3705 - val_loss: 8.4827\n",
      "Epoch 2/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.3566 - loss: 7.4516 - val_accuracy: 0.3228 - val_loss: 4.4466\n",
      "Epoch 3/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 15ms/step - accuracy: 0.2947 - loss: 4.3572 - val_accuracy: 0.3032 - val_loss: 3.1995\n",
      "Epoch 4/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.3092 - loss: 2.8685 - val_accuracy: 0.3322 - val_loss: 1.9692\n",
      "Epoch 5/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.3381 - loss: 1.8321 - val_accuracy: 0.3405 - val_loss: 1.6127\n",
      "Epoch 6/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.3474 - loss: 1.5980 - val_accuracy: 0.3621 - val_loss: 1.5295\n",
      "Epoch 7/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.3604 - loss: 1.5212 - val_accuracy: 0.3588 - val_loss: 1.4604\n",
      "Epoch 8/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.3757 - loss: 1.4557 - val_accuracy: 0.3809 - val_loss: 1.4095\n",
      "Epoch 9/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.3969 - loss: 1.3931 - val_accuracy: 0.3948 - val_loss: 1.3708\n",
      "Epoch 10/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4164 - loss: 1.3594 - val_accuracy: 0.4071 - val_loss: 1.3422\n",
      "Epoch 11/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4286 - loss: 1.3367 - val_accuracy: 0.4204 - val_loss: 1.3164\n",
      "Epoch 12/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4420 - loss: 1.2971 - val_accuracy: 0.4240 - val_loss: 1.2971\n",
      "Epoch 13/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4422 - loss: 1.2808 - val_accuracy: 0.4323 - val_loss: 1.2831\n",
      "Epoch 14/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4440 - loss: 1.2797 - val_accuracy: 0.4375 - val_loss: 1.2734\n",
      "Epoch 15/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4593 - loss: 1.2620 - val_accuracy: 0.4396 - val_loss: 1.2652\n",
      "Epoch 16/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4515 - loss: 1.2635 - val_accuracy: 0.4357 - val_loss: 1.2605\n",
      "Epoch 17/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4544 - loss: 1.2541 - val_accuracy: 0.4356 - val_loss: 1.2582\n",
      "Epoch 18/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4617 - loss: 1.2464 - val_accuracy: 0.4365 - val_loss: 1.2535\n",
      "Epoch 19/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4540 - loss: 1.2518 - val_accuracy: 0.4454 - val_loss: 1.2491\n",
      "Epoch 20/20\n",
      "\u001b[1m21/21\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4577 - loss: 1.2375 - val_accuracy: 0.4361 - val_loss: 1.2477\n"
     ]
    }
   ],
   "source": [
    "epochs_init = 20\n",
    "history_init = classifier_init.fit(x_train, y_train,\n",
    "                         epochs=epochs_init, batch_size=1000,\n",
    "                         validation_split=0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "16dca591",
   "metadata": {},
   "source": [
    "The above doesn't seem to be achieving a particularly good accuracy. Let's tweak the model a bit and retrain:\n",
    "\n",
    "* add an explicit input specification\n",
    "* add more capacity to the deep layers\n",
    "* randomize coefficients\n",
    "* tweak gradient descent parameters\n",
    "* add more epochs\n",
    "* decrease batch size a bit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "af4cf9c7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 24ms/step - accuracy: 0.3593 - loss: 17.3410 - val_accuracy: 0.3596 - val_loss: 3.3322\n",
      "Epoch 2/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.3873 - loss: 2.9919 - val_accuracy: 0.3977 - val_loss: 1.9189\n",
      "Epoch 3/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4113 - loss: 1.7894 - val_accuracy: 0.4211 - val_loss: 1.3407\n",
      "Epoch 4/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4389 - loss: 1.3377 - val_accuracy: 0.4422 - val_loss: 1.2794\n",
      "Epoch 5/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 14ms/step - accuracy: 0.4576 - loss: 1.2830 - val_accuracy: 0.4526 - val_loss: 1.2479\n",
      "Epoch 6/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4629 - loss: 1.2533 - val_accuracy: 0.4617 - val_loss: 1.2289\n",
      "Epoch 7/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 14ms/step - accuracy: 0.4738 - loss: 1.2359 - val_accuracy: 0.4615 - val_loss: 1.2208\n",
      "Epoch 8/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 14ms/step - accuracy: 0.4770 - loss: 1.2270 - val_accuracy: 0.4645 - val_loss: 1.2155\n",
      "Epoch 9/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4809 - loss: 1.2223 - val_accuracy: 0.4657 - val_loss: 1.2157\n",
      "Epoch 10/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4810 - loss: 1.2186 - val_accuracy: 0.4701 - val_loss: 1.2144\n",
      "Epoch 11/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4813 - loss: 1.2167 - val_accuracy: 0.4733 - val_loss: 1.2142\n",
      "Epoch 12/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4810 - loss: 1.2151 - val_accuracy: 0.4742 - val_loss: 1.2139\n",
      "Epoch 13/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4813 - loss: 1.2138 - val_accuracy: 0.4724 - val_loss: 1.2130\n",
      "Epoch 14/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4817 - loss: 1.2128 - val_accuracy: 0.4743 - val_loss: 1.2123\n",
      "Epoch 15/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4822 - loss: 1.2120 - val_accuracy: 0.4716 - val_loss: 1.2119\n",
      "Epoch 16/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4822 - loss: 1.2115 - val_accuracy: 0.4724 - val_loss: 1.2111\n",
      "Epoch 17/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4823 - loss: 1.2104 - val_accuracy: 0.4724 - val_loss: 1.2103\n",
      "Epoch 18/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 14ms/step - accuracy: 0.4825 - loss: 1.2101 - val_accuracy: 0.4735 - val_loss: 1.2098\n",
      "Epoch 19/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4835 - loss: 1.2096 - val_accuracy: 0.4747 - val_loss: 1.2101\n",
      "Epoch 20/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4828 - loss: 1.2089 - val_accuracy: 0.4769 - val_loss: 1.2081\n",
      "Epoch 21/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4845 - loss: 1.2080 - val_accuracy: 0.4780 - val_loss: 1.2078\n",
      "Epoch 22/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4839 - loss: 1.2070 - val_accuracy: 0.4781 - val_loss: 1.2064\n",
      "Epoch 23/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4846 - loss: 1.2058 - val_accuracy: 0.4789 - val_loss: 1.2049\n",
      "Epoch 24/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4834 - loss: 1.2048 - val_accuracy: 0.4802 - val_loss: 1.2044\n",
      "Epoch 25/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4856 - loss: 1.2040 - val_accuracy: 0.4825 - val_loss: 1.2030\n",
      "Epoch 26/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4854 - loss: 1.2031 - val_accuracy: 0.4825 - val_loss: 1.2027\n",
      "Epoch 27/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 14ms/step - accuracy: 0.4862 - loss: 1.2027 - val_accuracy: 0.4822 - val_loss: 1.2021\n",
      "Epoch 28/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4865 - loss: 1.2016 - val_accuracy: 0.4849 - val_loss: 1.2008\n",
      "Epoch 29/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 14ms/step - accuracy: 0.4871 - loss: 1.2008 - val_accuracy: 0.4865 - val_loss: 1.1995\n",
      "Epoch 30/30\n",
      "\u001b[1m37/37\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4867 - loss: 1.1996 - val_accuracy: 0.4848 - val_loss: 1.1989\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "tf.random.set_seed(42);\n",
    "tf.keras.utils.set_random_seed(42);\n",
    "\n",
    "classifier_new = keras.Sequential([\n",
    "    layers.Input((11,)),\n",
    "    layers.Dense(128, activation=\"relu\"),\n",
    "    layers.Dense(128, activation=\"relu\"),\n",
    "    layers.Dense(10, activation=\"softmax\")\n",
    "])\n",
    "\n",
    "classifier_new.compile(optimizer=keras.optimizers.Adam(learning_rate=0.0001),\n",
    "                   loss=\"sparse_categorical_crossentropy\",\n",
    "                   metrics=[\"accuracy\"])\n",
    "\n",
    "epochs_new = 30\n",
    "history_new = classifier_new.fit(x_train, y_train,\n",
    "                         epochs=epochs_new, batch_size=500,\n",
    "                         validation_split=0.3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e0519464",
   "metadata": {},
   "source": [
    "This looks better, but how can you be sure? Visualise it!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "c236c890",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "val_loss = history_new.history[\"val_loss\"]\n",
    "trn_loss = history_new.history[\"loss\"]\n",
    "val_accuracy = history_new.history[\"val_accuracy\"]\n",
    "trn_accuracy = history_new.history[\"accuracy\"]\n",
    "\n",
    "fig, loss = plt.subplots()\n",
    "loss.plot(range(0, epochs_new), val_loss, \"b-\", label=\"Validation Loss\")\n",
    "loss.plot(range(0, epochs_new), trn_loss, \"r-\", label=\"Training Loss\")\n",
    "loss.set_ylabel(\"Loss\")\n",
    "h1, l1 = loss.get_legend_handles_labels()\n",
    "\n",
    "accr = loss.twinx()\n",
    "accr.plot(range(0, epochs_new), val_accuracy, \"g\", label=\"Validation Accuracy\")\n",
    "accr.plot(range(0, epochs_new), trn_accuracy, \"y\", label=\"Training Accuracy\")\n",
    "accr.set_ylabel(\"Accuracy\")\n",
    "h2, l2 = accr.get_legend_handles_labels()\n",
    "\n",
    "fig.legend(h1 + h2, l1 + l2, loc=(0.5, 0.5))\n",
    "\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.xticks(range(0, epochs_new, 5))\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d2a84be1",
   "metadata": {},
   "source": [
    "Let's compare that with the previous run."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "d5731bbc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "val_loss = history_init.history[\"val_loss\"]\n",
    "trn_loss = history_init.history[\"loss\"]\n",
    "val_accuracy = history_init.history[\"val_accuracy\"]\n",
    "trn_accuracy = history_init.history[\"accuracy\"]\n",
    "\n",
    "fig, loss = plt.subplots()\n",
    "loss.plot(range(0, epochs_init), val_loss, \"b-\", label=\"Validation Loss\")\n",
    "loss.plot(range(0, epochs_init), trn_loss, \"r-\", label=\"Training Loss\")\n",
    "loss.set_ylabel(\"Loss\")\n",
    "h1, l1 = loss.get_legend_handles_labels()\n",
    "\n",
    "accr = loss.twinx()\n",
    "accr.plot(range(0, epochs_init), val_accuracy, \"g\", label=\"Validation Accuracy\")\n",
    "accr.plot(range(0, epochs_init), trn_accuracy, \"y\", label=\"Training Accuracy\")\n",
    "accr.set_ylabel(\"Accuracy\")\n",
    "h2, l2 = accr.get_legend_handles_labels()\n",
    "\n",
    "fig.legend(h1 + h2, l1 + l2, loc=(0.1, 0.7))\n",
    "\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.xticks(range(0, epochs_init, 5))\n",
    "plt.show()\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "tf-2",
   "language": "python",
   "name": "tf-2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}