Please download

MNIST_Subset.csv

Each row of this file represents an image from a subset of the MNIST data set with some added noise. If you are asked to create a plot, you are expected to include the plot in your final report. You do not need axis labels for plots created in this problem. a. Plot the first five images. What numbers displayed in these images? Are the images clear or noisy? (Hint: The reshape function will allow you to transform the individual rows into 28 by 28 matrices.) b. Using an SVD with k : 2, denoise the images. Plot the first five denoised images, along with the original noisy images. How do the shapes of the digits in the denoised images compare to the original noisy images? How do the edges of the image compare to the original noisy images? 0. Using an SVD with k = 15, denoise the images. Plot the first five denoised images, along with the original noisy images. How do the shapes of the digits in the denoised images compare to the original noisy images? How do the edges of the image compare to the original noisy images? Additionally, compare your denoised images using k = 15 to your results under k = 2. d. Using an SVD with k = 250, denoise the images. Plot the first five denoised images, along with the original noisy images. How do the shapes of the digits in the denoised images compare to the original noisy images? How do the edges of the image compare to the original noisy images? Additionally, compare your denoised images using k : 250 to your results under 71: : 2 and k : 15. Submit all of your responses to Question 1 as a single PDF document (including code). AttributeError Traceback (most recent call last) Input In [ 2], in () 9 # Assuming 'data' is the given MNIST data set with added noise 10 for i in range (5) : - - -> 11 img = data [i] . reshape (28, 28) 12 pit . imshow(img, cmap= ' gray' ) 13 pit . show( ) AttributeError: 'tuple' object has no attribute 'reshape'