Training Datat Please code each answer in python The training dataset for this problem corresponds to the images of handwritten digits '0' and '1' that come prepackaged with the sklearn package. There are a total of 360 images of digits '0' and '1'; we will divide them into 300 images for training purposes, while we will evaluate the performance of k-NN classification on the remaining 60 nnages, which we will refer to as the test" set! You can use the following code to obtaining the training and the test sets. from sklearn.datasets import load digits images, labels load_digits(2, return X_ y-True,) # Labeled training set training_imagesimages[:300] training-labels labels [ :300] # Labeled test set test_imagesimages [300:] test-labels-labels [3001] k-NN Classification Using 2-D Features 1. Carry out principal component analysis (PCA) of images in the training set and compute two-dimensional PCA features of training images. Display the two-dimensional features of training images as points on a two-dimensional scatter plot. Color all points corresponding to digits '0' as red and all points corresponding to digits '1' as 2. Classify each image in the test set by first transforming it to the two-dimensional PCA domain using the principal components obtained above and then using k-NN classification with k-5 and the distance metric being |I 2 Display the two-dimensional features of test images as points on a two-dimensional scatter plot Color all points that are correctly classified as blue and all points that are incorrectly classified as black. 3. Compute and display the average classification error for the test set defined as 1 1G# here. yi denotes the true label of the i-th image, i denotes the label returned by k-NN, and N 60 in this particular problem. Principled splitting up of a given dataset into training, testing, and validation sets is an important aspect of machine learning that will be covered in a future lecture. Training Datat Please code each answer in python The training dataset for this problem corresponds to the images of handwritten digits '0' and '1' that come prepackaged with the sklearn package. There are a total of 360 images of digits '0' and '1'; we will divide them into 300 images for training purposes, while we will evaluate the performance of k-NN classification on the remaining 60 nnages, which we will refer to as the test" set! You can use the following code to obtaining the training and the test sets. from sklearn.datasets import load digits images, labels load_digits(2, return X_ y-True,) # Labeled training set training_imagesimages[:300] training-labels labels [ :300] # Labeled test set test_imagesimages [300:] test-labels-labels [3001] k-NN Classification Using 2-D Features 1. Carry out principal component analysis (PCA) of images in the training set and compute two-dimensional PCA features of training images. Display the two-dimensional features of training images as points on a two-dimensional scatter plot. Color all points corresponding to digits '0' as red and all points corresponding to digits '1' as 2. Classify each image in the test set by first transforming it to the two-dimensional PCA domain using the principal components obtained above and then using k-NN classification with k-5 and the distance metric being |I 2 Display the two-dimensional features of test images as points on a two-dimensional scatter plot Color all points that are correctly classified as blue and all points that are incorrectly classified as black. 3. Compute and display the average classification error for the test set defined as 1 1G# here. yi denotes the true label of the i-th image, i denotes the label returned by k-NN, and N 60 in this particular problem. Principled splitting up of a given dataset into training, testing, and validation sets is an important aspect of machine learning that will be covered in a future lecture