Week5_Excercise.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Weel 5 Excercise: PCA and MNIST"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#the following code is pasted from [1]:\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import math\n",
    "\n",
    "# TensorFlow and tf.keras\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "\n",
    "import matplotlib.style #Some style nonsense\n",
    "import matplotlib as mpl #Some more style nonsense\n",
    "\n",
    "\n",
    "#Set default figure size\n",
    "#mpl.rcParams['figure.figsize'] = [12.0, 8.0] #Inches... of course it is inches\n",
    "mpl.rcParams[\"legend.frameon\"] = False\n",
    "mpl.rcParams['figure.dpi']=200 # dots per inch\n",
    "\n",
    "#not from [1]:\n",
    "\n",
    "from IPython.display import clear_output\n",
    "from scipy import linalg"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Part 1 : Train a neural network on the MNIST training dataset that achieves acceptable performance on the test dataset (>95% accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load MNIST data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "mnist = keras.datasets.mnist   #load MNIST data\n",
    "(train_images, train_labels), (test_images, test_labels) = mnist.load_data() #obtain test and training images"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create Nueral Network: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#create sequential nn:\n",
    "model = keras.Sequential([\n",
    "    keras.layers.Flatten(input_shape=(28,28)), #input layer of 28x28 nuerons\n",
    "    keras.layers.Dense(128,activation='relu'), #hidden layer of 128 nuerons\n",
    "    keras.layers.Dense(10) #output layer of 10 nuerons\n",
    "])\n",
    "\n",
    "#sequential to compile model:\n",
    "model.compile(loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True), optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "#Normalise pixel values:\n",
    "train_images=train_images/255.0\n",
    "test_images=test_images/255.0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Train Nueral Network with MNIST Training Images:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n",
      "600/600 [==============================] - 2s 3ms/step - loss: 0.3372 - accuracy: 0.9054 - val_loss: 0.1913 - val_accuracy: 0.9445\n",
      "Epoch 2/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.1544 - accuracy: 0.9556 - val_loss: 0.1278 - val_accuracy: 0.9622\n",
      "Epoch 3/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.1088 - accuracy: 0.9686 - val_loss: 0.1035 - val_accuracy: 0.9696\n",
      "Epoch 4/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0836 - accuracy: 0.9754 - val_loss: 0.0877 - val_accuracy: 0.9728\n",
      "Epoch 5/30\n",
      "600/600 [==============================] - 2s 2ms/step - loss: 0.0677 - accuracy: 0.9799 - val_loss: 0.0818 - val_accuracy: 0.9756\n",
      "Epoch 6/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0557 - accuracy: 0.9837 - val_loss: 0.0761 - val_accuracy: 0.9765\n",
      "Epoch 7/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0450 - accuracy: 0.9878 - val_loss: 0.0757 - val_accuracy: 0.9762\n",
      "Epoch 8/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0390 - accuracy: 0.9892 - val_loss: 0.0747 - val_accuracy: 0.9773\n",
      "Epoch 9/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0328 - accuracy: 0.9905 - val_loss: 0.0690 - val_accuracy: 0.9778\n",
      "Epoch 10/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0270 - accuracy: 0.9927 - val_loss: 0.0728 - val_accuracy: 0.9789\n",
      "Epoch 11/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0227 - accuracy: 0.9938 - val_loss: 0.0777 - val_accuracy: 0.9755\n",
      "Epoch 12/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0185 - accuracy: 0.9952 - val_loss: 0.0774 - val_accuracy: 0.9778\n",
      "Epoch 13/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0162 - accuracy: 0.9960 - val_loss: 0.0798 - val_accuracy: 0.9764\n",
      "Epoch 14/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0139 - accuracy: 0.9967 - val_loss: 0.0735 - val_accuracy: 0.9787\n",
      "Epoch 15/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0114 - accuracy: 0.9973 - val_loss: 0.0774 - val_accuracy: 0.9775\n",
      "Epoch 16/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0095 - accuracy: 0.9979 - val_loss: 0.0754 - val_accuracy: 0.9794\n",
      "Epoch 17/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0077 - accuracy: 0.9985 - val_loss: 0.0816 - val_accuracy: 0.9789\n",
      "Epoch 18/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0068 - accuracy: 0.9987 - val_loss: 0.0830 - val_accuracy: 0.9766\n",
      "Epoch 19/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0082 - accuracy: 0.9980 - val_loss: 0.0859 - val_accuracy: 0.9786\n",
      "Epoch 20/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0053 - accuracy: 0.9991 - val_loss: 0.0848 - val_accuracy: 0.9784\n",
      "Epoch 21/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0040 - accuracy: 0.9995 - val_loss: 0.0799 - val_accuracy: 0.9798\n",
      "Epoch 22/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0037 - accuracy: 0.9994 - val_loss: 0.0920 - val_accuracy: 0.9782\n",
      "Epoch 23/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0074 - accuracy: 0.9980 - val_loss: 0.0955 - val_accuracy: 0.9763\n",
      "Epoch 24/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0028 - accuracy: 0.9996 - val_loss: 0.0868 - val_accuracy: 0.9807\n",
      "Epoch 25/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0023 - accuracy: 0.9997 - val_loss: 0.0926 - val_accuracy: 0.9786\n",
      "Epoch 26/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0050 - accuracy: 0.9987 - val_loss: 0.0903 - val_accuracy: 0.9802\n",
      "Epoch 27/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0030 - accuracy: 0.9993 - val_loss: 0.0929 - val_accuracy: 0.9800\n",
      "Epoch 28/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0011 - accuracy: 1.0000 - val_loss: 0.0913 - val_accuracy: 0.9804\n",
      "Epoch 29/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 6.3449e-04 - accuracy: 1.0000 - val_loss: 0.0911 - val_accuracy: 0.9811\n",
      "Epoch 30/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 5.4588e-04 - accuracy: 1.0000 - val_loss: 0.0923 - val_accuracy: 0.9812\n"
     ]
    }
   ],
   "source": [
    "history=model.fit(train_images, train_labels, batch_size=100, epochs=30, validation_data=(test_images, test_labels))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    ">95% test accuracy achieved! (97.58% specifcically)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Part 2: Decompose the MNIST training dataset into its PCA components"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "vecs.shape: (784, 784)\n",
      "vals.shape: (784,)\n"
     ]
    }
   ],
   "source": [
    "#the following code is adapted from [1]\n",
    "\n",
    "X=np.reshape(train_images,(60000,784)) #reshape train imgs to 60000 of 784 individual pixels\n",
    "mu=np.mean(X,axis=0)  #calculate mean\n",
    "x=X-mu #to make the mean of x = 0\n",
    "\n",
    "rho=np.cov(x,rowvar=False) # Make the covaraince matrix by np.cov\n",
    "vals,vecs=linalg.eigh(rho) #Get the eigenvalues and vectors from the covariance matrix\n",
    "\n",
    "#vals is a 1-D array of the eigenvalues in ascending order, vecs is a columnwise array of the normalised\n",
    "# eigenvectors such that vecs[:,i] is the eigenvector correspondong to vals[i]\n",
    "print(\"vecs.shape:\",np.shape(vecs))\n",
    "print(\"vals.shape:\",np.shape(vals))\n",
    "vecs=np.flip(vecs) #eigenvectors\n",
    "vals=np.flip(vals) #eigenvalues "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "A few of the PCA components are shown below:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#the following code is taken directly from [1]\n",
    "\n",
    "print(\"A few of the PCA components are shown below:\")\n",
    "fig,ax=plt.subplots(2,3)\n",
    "for i in range(2):\n",
    "    for j in range(3):\n",
    "        ax[i,j].imshow(np.reshape(vecs[:,(i+j*3)],[28,28]), origin='lower',cmap='binary',interpolation='nearest')\n",
    "plt.subplots_adjust(left=0.05,\n",
    "                    bottom=0.2, \n",
    "                    right=0.9, \n",
    "                    top=0.9, \n",
    "                    wspace=0.1, \n",
    "                    hspace=0.2)  #Adjust the space between the subplots so there isn't acres of whitespace"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Part 3: Use the N-best PCA components to make a new testing dataset (using the 10,000 MNIST test images) and make a plot of N vs testing accuracy using the neural network trained in 1) *WITOUT RETRAINING THE NETWORK*\n",
    "\n",
    "Decompose MNIST test data into its PCA components:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "vecs.shape: (784, 784)\n",
      "vals.shape: (784,)\n"
     ]
    }
   ],
   "source": [
    "#the following code is adapted from [1]\n",
    "\n",
    "X_tst=np.reshape(test_images,(10000,784)) #reshape test imgs to 10000 of 784 individual pixels\n",
    "mu_tst=np.mean(X_tst,axis=0)  #calculate mean\n",
    "x_tst=X_tst-mu_tst #to make the mean of x = 0\n",
    "\n",
    "rho_tst=np.cov(x_tst,rowvar=False) # Make the covaraince matrix by np.cov\n",
    "vals_tst,vecs_tst=linalg.eigh(rho_tst) #Get the eigenvalues and vectors from the covariance matrix\n",
    "\n",
    "#vals is a 1-D array of the eigenvalues in ascending order, vecs is a columnwise array of the normalised\n",
    "# eigenvectors such that vecs[:,i] is the eigenvector correspondong to vals[i]\n",
    "print(\"vecs.shape:\",np.shape(vecs_tst))\n",
    "print(\"vals.shape:\",np.shape(vals_tst))\n",
    "vecs_tst=np.flip(vecs_tst) #eigenvectors\n",
    "vals_tst=np.flip(vals_tst) #eigenvalues "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot 'Fraction of Explained Variance' against 'PCA index'. It shows you how much each principle component contributes to the total variance of the image (ie quality of image):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#the following code is adapted from [1]\n",
    "fig,ax=plt.subplots()\n",
    "cumsum_tst=np.cumsum(vals_tst.real) #cumulative sum of eigenvalues\n",
    "cdf_tst=cumsum_tst/cumsum_tst[-1] #in units of final eigenvalue\n",
    "index_tst=np.arange((np.shape(vals_tst)[0])) #index range\n",
    "\n",
    "ax.plot(index_tst,cdf_tst,\"-o\", markersize=3)\n",
    "ax.set_xlabel(\"PCA Index\")\n",
    "ax.set_ylabel(\"Fraction of explained variance\")\n",
    "ax.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the plot, we can see that 0-150 PCA index contributes to ~95% of the variance. Therefore, using the first 100 components should give good test accuracy. Therefore, for Part 3, 4 and 5, I will be using 150 PCA as N to plot Test Accuracy vs N. This can also be done using all 784 components but this will take more time, use more computing power and will only be very slightly more accurate. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "New test images shape: (10000, 28, 28)\n",
      "Original MNIST test images shape: (10000, 28, 28)\n"
     ]
    }
   ],
   "source": [
    "#using 150:\n",
    "P_tst=np.dot(x_tst,vecs_tst) #dot product of test image pixel values and eigenvectors\n",
    "newImage_tst=(np.dot(P_tst[:,0:150],vecs_tst.T[0:150,:]))+mu #new testing data set\n",
    "\n",
    "#reshape to original:\n",
    "test_n = np.reshape(newImage_tst, (10000, 28, 28))\n",
    "                      \n",
    "#we can see how this new test dataset is the same shape as the original MNIST test data\n",
    "print(\"New test images shape:\", np.shape(test_n))\n",
    "print(\"Original MNIST test images shape:\",np.shape(test_images))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Make plot of N vs Test Accuracy for the each of the N-best PCA components used:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "index=150 #max index we will use\n",
    "\n",
    "test_acc_n = [] #initialise empty array of test accuracies\n",
    "for i in range(index+1): #generate new test data and run nn with all index up to 150\n",
    "    \n",
    "    newImage_tst=(np.dot(P_tst[:,0:i],vecs_tst.T[0:i,:]))+mu #new test images\n",
    "    test_n = np.reshape(newImage_tst, (10000, 28, 28)) #reshape\n",
    "    \n",
    "    test_loss, test_acc = model.evaluate(test_n,  test_labels, verbose=2) #run nn\n",
    "    test_acc_n.append(test_acc)\n",
    "    \n",
    "    clear_output(wait=True)\n",
    "\n",
    "\n",
    "#plot:\n",
    "fig,ax=plt.subplots()\n",
    "ax.plot(range(index+1),test_acc_n,\"-o\", markersize=3)\n",
    "ax.set_ylabel(\"Testing Accuracy\")\n",
    "ax.set_xlabel(\"N (PCA Index used)\")\n",
    "ax.set_title(\"Testing Accuracy vs N for Nueral Netowrk Trained under MNIST Training Data\")\n",
    "ax.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Comments:\n",
    "\n",
    "It is clear from the plot that the higher N is used, the greater the test accuracy. However, there is a point wher the test accuracy doesnt increase much more. Here, that point is ~N=40 where ~95% accracy is achieved and is sufficient."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Part 4 & 5: Now using the 100-best PCA components make a new training dataset (using the 60,000 MNIST training images) and retrain your neural network from 1) and Repeat step 3 (using your new neural network from 4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Generate new training images using he 100-best PCA components:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "New training images shape: (60000, 28, 28)\n",
      "Original MNIST training images shape: (60000, 28, 28)\n"
     ]
    }
   ],
   "source": [
    "#using the 100 best:\n",
    "P_train=np.dot(x,vecs) #dot product of test image pixel values and eigenvectors\n",
    "newImage100=(np.dot(P_train[:,0:100],vecs.T[0:100,:]))+mu #new training data set (N=100)\n",
    "\n",
    "#reshape to original:\n",
    "train100 = np.reshape(newImage100, (60000, 28, 28))\n",
    "                      \n",
    "#we can see how this new test dataset is the same shape as the original MNIST training data\n",
    "print(\"New training images shape:\", np.shape(train100))\n",
    "print(\"Original MNIST training images shape:\",np.shape(train_images))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Train Nueral Network using the new training images:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0573 - accuracy: 0.9828\n",
      "Epoch 2/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0241 - accuracy: 0.9918\n",
      "Epoch 3/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0166 - accuracy: 0.9945\n",
      "Epoch 4/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0125 - accuracy: 0.9962\n",
      "Epoch 5/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0112 - accuracy: 0.9968\n",
      "Epoch 6/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0103 - accuracy: 0.9967\n",
      "Epoch 7/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0098 - accuracy: 0.9968\n",
      "Epoch 8/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0091 - accuracy: 0.9974\n",
      "Epoch 9/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0072 - accuracy: 0.9979\n",
      "Epoch 10/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0074 - accuracy: 0.9980\n",
      "Epoch 11/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0074 - accuracy: 0.9979\n",
      "Epoch 12/30\n",
      "600/600 [==============================] - 2s 3ms/step - loss: 0.0062 - accuracy: 0.9982\n",
      "Epoch 13/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0071 - accuracy: 0.9979\n",
      "Epoch 14/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0069 - accuracy: 0.9980\n",
      "Epoch 15/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0049 - accuracy: 0.9987\n",
      "Epoch 16/30\n",
      "600/600 [==============================] - 2s 3ms/step - loss: 0.0052 - accuracy: 0.9985\n",
      "Epoch 17/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0031 - accuracy: 0.9994\n",
      "Epoch 18/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0043 - accuracy: 0.9989\n",
      "Epoch 19/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0087 - accuracy: 0.9973\n",
      "Epoch 20/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0042 - accuracy: 0.9988\n",
      "Epoch 21/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0028 - accuracy: 0.9994\n",
      "Epoch 22/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0049 - accuracy: 0.9985\n",
      "Epoch 23/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0025 - accuracy: 0.9995\n",
      "Epoch 24/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0012 - accuracy: 0.9999\n",
      "Epoch 25/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0052 - accuracy: 0.9982\n",
      "Epoch 26/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0096 - accuracy: 0.9966\n",
      "Epoch 27/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0032 - accuracy: 0.9992\n",
      "Epoch 28/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0024 - accuracy: 0.9994\n",
      "Epoch 29/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0017 - accuracy: 0.9996\n",
      "Epoch 30/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0049 - accuracy: 0.9984\n"
     ]
    }
   ],
   "source": [
    "history=model.fit(train100, train_labels, batch_size=100, epochs=30)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the newly trained nueral network with the test images from before:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_acc_n = []\n",
    "for i in range(index+1):\n",
    "    newImage_tst=(np.dot(P_tst[:,0:i],vecs_tst.T[0:i,:]))+mu\n",
    "    test_n = np.reshape(newImage_tst, (10000, 28, 28)) #reshape\n",
    "    \n",
    "    test_loss, test_acc = model.evaluate(test_n,  test_labels, verbose=2)\n",
    "    test_acc_n.append(test_acc)\n",
    "    \n",
    "    clear_output(wait=True)\n",
    "\n",
    "fig,ax=plt.subplots()\n",
    "ax.plot(range(index+1),test_acc_n,\"-o\", markersize=3)\n",
    "ax.set_ylabel(\"Testing Accuracy\")\n",
    "ax.set_xlabel(\"N (PCA Index used)\")\n",
    "ax.set_title(\"Testing Accuracy vs N for Nueral Netowrk Trained under 100-best PCA Components MNIST Training Data\")\n",
    "ax.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Part 6: Finally repeat steps 4 and 5 using a network trained with the best 20 components"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Generate new training images using he 20-best PCA components:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "New training images shape: (60000, 28, 28)\n",
      "Original MNIST training images shape: (60000, 28, 28)\n"
     ]
    }
   ],
   "source": [
    "#using the first 20:\n",
    "newImage20=(np.dot(P_train[:,0:20],vecs.T[0:20,:]))+mu #new training data set\n",
    "\n",
    "#reshape to original:\n",
    "train20 = np.reshape(newImage20, (60000, 28, 28))\n",
    "                      \n",
    "#we can see how this new test dataset is the same shape as the original MNIST training data\n",
    "print(\"New training images shape:\", np.shape(train20))\n",
    "print(\"Original MNIST training images shape:\",np.shape(train_images))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Train Nueral Network using the new training images:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n",
      "600/600 [==============================] - 2s 3ms/step - loss: 0.2014 - accuracy: 0.9498\n",
      "Epoch 2/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.1256 - accuracy: 0.9642\n",
      "Epoch 3/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.1081 - accuracy: 0.9675\n",
      "Epoch 4/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0971 - accuracy: 0.9709\n",
      "Epoch 5/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0903 - accuracy: 0.9720\n",
      "Epoch 6/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0839 - accuracy: 0.9737\n",
      "Epoch 7/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0800 - accuracy: 0.9748\n",
      "Epoch 8/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0759 - accuracy: 0.9758\n",
      "Epoch 9/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0723 - accuracy: 0.9766\n",
      "Epoch 10/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0705 - accuracy: 0.9768\n",
      "Epoch 11/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0665 - accuracy: 0.9789\n",
      "Epoch 12/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0659 - accuracy: 0.9784\n",
      "Epoch 13/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0630 - accuracy: 0.9792\n",
      "Epoch 14/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0602 - accuracy: 0.9807\n",
      "Epoch 15/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0598 - accuracy: 0.9801\n",
      "Epoch 16/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0581 - accuracy: 0.9804\n",
      "Epoch 17/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0563 - accuracy: 0.9814\n",
      "Epoch 18/30\n",
      "600/600 [==============================] - ETA: 0s - loss: 0.0559 - accuracy: 0.98 - 1s 2ms/step - loss: 0.0557 - accuracy: 0.9818\n",
      "Epoch 19/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0547 - accuracy: 0.9820\n",
      "Epoch 20/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0529 - accuracy: 0.9822\n",
      "Epoch 21/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0529 - accuracy: 0.9824\n",
      "Epoch 22/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0512 - accuracy: 0.9823\n",
      "Epoch 23/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0486 - accuracy: 0.9839\n",
      "Epoch 24/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0491 - accuracy: 0.9837\n",
      "Epoch 25/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0471 - accuracy: 0.9843\n",
      "Epoch 26/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0482 - accuracy: 0.9834\n",
      "Epoch 27/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0467 - accuracy: 0.9839\n",
      "Epoch 28/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0452 - accuracy: 0.9853\n",
      "Epoch 29/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0442 - accuracy: 0.9855\n",
      "Epoch 30/30\n",
      "600/600 [==============================] - 1s 2ms/step - loss: 0.0446 - accuracy: 0.9852\n"
     ]
    }
   ],
   "source": [
    "history=model.fit(train20, train_labels, batch_size=100, epochs=30)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the newly trained nueral network with the test images from before:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_acc_n = []\n",
    "for i in range(index+1):\n",
    "    newImage_tst=(np.dot(P_tst[:,0:i],vecs_tst.T[0:i,:]))+mu\n",
    "    test_n = np.reshape(newImage_tst, (10000, 28, 28)) #reshape\n",
    "    \n",
    "    test_loss, test_acc = model.evaluate(test_n,  test_labels, verbose=2)\n",
    "    test_acc_n.append(test_acc)\n",
    "    \n",
    "    clear_output(wait=True)\n",
    "\n",
    "fig,ax=plt.subplots()\n",
    "ax.plot(range(index+1),test_acc_n,\"-o\", markersize=3)\n",
    "ax.set_ylabel(\"Testing Accuracy\")\n",
    "ax.set_xlabel(\"N (PCA Index used)\")\n",
    "ax.set_title(\"Testing Accuracy vs N for Nueral Netowrk Trained under 20-best PCA Components MNIST Training Data\")\n",
    "ax.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conclusion:\n",
    "\n",
    "Using 100-best PCA components gives virtually the same result as using the unmodified MNIST training data. This shows that our PCA training data mimics the MNIST training data very well. Therefore, Principle Component Analysis (PCA) is a very accurate technique in machine learning.\n",
    "\n",
    "Using 20-best PCA components gives the worst result because it decreases in accuracy around the 20 PCA component (for the TEST data, not training data). So, using less PCA components means that accuracy will decrease. Therefore using more PCA compenents for the training dataset will increase testing accuracy. However, from our discussion of the very first plot in this notebook, there is a point where increasing the PCA components used further, will have a negligible improvement in the testing accuracy.\n",
    "\n",
    "# BIG MISTAKE:\n",
    "\n",
    "### Didn't make, train and run a NEW nueral network for each part. Because of this the same NN used was ran over the wrong data from the beginning. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Bibliography:\n",
    "\n",
    "[1]Nichol. R 2021. \"Week5_PCA\". [Online] UCL Moodle resource. Available from https://moodle.ucl.ac.uk/course/view.php?id=6793 [accessed 9th November 2021] "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}