Problem 1) Classifying handwritten digits Using the MNIST dataset, we will build a simple threshold based classifier that classifies 0 digits from non 0s a) Create a new training, validation and test set as you did in problem 3 of Assignment 1, but this time include only the 0 and 1 digits Hint From the test set select only 0s and 1s Repeat this for the training set Then, create a validation set as done in the previous assignment b) Convert each image to one attribute by calculating the average of all the pixel values in the center 3x3 grid of the image c) Randomly select 500 of the images from the training data Plot their attribute values that you calculated in part (b) Use different colors and shapes for 0s and 1s The x axis in your figure is the image number (1 to 500) and y axis is the calculated attribute Label the axes and add legends appropriately Note that these 500 images are your validation set Problem 3 from Assignment 1 from keras datasets import mnist import matplotlib pyplot as plt import numpy as np from random import randint (x train, y train), (x test, y test) mnist load data() This is what I have so far, but I need to select all 1's and 0's not just 4 Selecting O's and I's from training set print('selecting zeros and ones from training set') x train rnd np zeros((10,x train shape 1 ,x train shape 2 )) y train rnd np arange (10) for digit in range (2) x train d x train y train digit, , x train rnd digit, , x train d randint(0,x train d shape 0 ), , img plt(x train rnd,y train rnd) print('selecting zeros and ones from testing set') x test rnd np zeros((10,x test shape 1 ,x test shape 2 )) y test rnd np arange (10) for digit in range(2) x test d x test y test digit, , x test rnd digit, , x test d randint(0,x test d shape 0 ), , img plt(x test rnd,y test rnd) Selecting O's and I's from training set print('selecting zeros and ones from training set') x train rnd np zeros((10,x train shape 1 ,x train shape 2 )) y train rnd np arange (10) for digit in range (2) x train d x train y train digit, , x train rnd digit, , x train d randint(0,x train d shape 0 ), , img plt(x train rnd,y train rnd) print('selecting zeros and ones from testing set') x test rnd np zeros((10,x test shape 1 ,x test shape 2 )) y test rnd np arange (10) for digit in range(2) x test d x test y test digit, , x test rnd digit, , x test d randint(0,x test d shape 0 ), , img plt(x test rnd,y test rnd)

The Answer is in the image, click to view ...

Answered step by step

Verified Expert Solution

Link Copied!

Question

1 Approved Answer

Posted on Sep 22, 2024

Problem 1) Classifying handwritten digits: Using the MNIST dataset, we will build a simple threshold-based classifier that classifies 0 digits from non-0s. a) Create a

Problem 1) Classifying handwritten digits: Using the MNIST dataset, we will build a simple threshold-based classifier that classifies 0 digits from non-0s. a) Create a new training, validation and test set as you did in problem 3 of Assignment 1, but this time include only the 0 and 1 digits. Hint: From the test set select only 0s and 1s. Repeat this for the training set. Then, create a validation set as done in the previous assignment. b) Convert each image to one attribute by calculating the average of all the pixel values in the center 3x3 grid of the image. c) Randomly select 500 of the images from the training data. Plot their attribute values that you calculated in part (b). Use different colors and shapes for 0s and 1s. The x axis in your figure is the image number (1 to 500) and y axis is the calculated attribute. Label the axes and add legends appropriately. Note that these 500 images are your validation set. Problem 3 from Assignment 1:

from keras.datasets import mnist import matplotlib.pyplot as plt import numpy as np from random import randint

(x_train, y_train), (x_test, y_test) = mnist.load_data() image text in transcribed This is what I have so far, but I need to select all 1's and 0's not just 4.

#Selecting O's and I's from training set print('selecting zeros and ones from training set') x_train_rnd=np.zeros((10,x_train.shape[1],x_train.shape[2])) y_train_rnd=np.arange (10) for digit in range (2): x_train_d=x_train[y_train=-digit,:,:] x_train_rnd[digit,:,:]=x_train_d[randint(0,x_train_d.shape[0]),:,:] img_plt(x_train_rnd,y_train_rnd) print('selecting zeros and ones from testing set') x_test_rnd=np.zeros((10,x_test. shape[1],x_test. shape[2])) y_test_rnd=np.arange (10) for digit in range(2) : x_test_d=x_test[y_test=digit,:,:] x_test_rnd[digit, :,:]=x_test_d[randint(0,x_test_d.shape[0]),:,:] img_plt(x_test_rnd,y_test_rnd) #Selecting O's and I's from training set print('selecting zeros and ones from training set') x_train_rnd=np.zeros((10,x_train.shape[1],x_train.shape[2])) y_train_rnd=np.arange (10) for digit in range (2): x_train_d=x_train[y_train=-digit,:,:] x_train_rnd[digit,:,:]=x_train_d[randint(0,x_train_d.shape[0]),:,:] img_plt(x_train_rnd,y_train_rnd) print('selecting zeros and ones from testing set') x_test_rnd=np.zeros((10,x_test. shape[1],x_test. shape[2])) y_test_rnd=np.arange (10) for digit in range(2) : x_test_d=x_test[y_test=digit,:,:] x_test_rnd[digit, :,:]=x_test_d[randint(0,x_test_d.shape[0]),:,:] img_plt(x_test_rnd,y_test_rnd)