in python

The Extended Classifier: The basic idea is to have 10 single-digit classifiers, each built to recognize a single digit, working in parallel. Recall that all the features of all the images are presented at the same time to the classifier using the X matrix. Then, each image, put through the classifier built above, to obtain the regression prediction followed by being put through the Sigmoid (before using np. round () ), delivers a value between 0&1. The prediction value giving the confidence in the prediction. A similar approach will be used for the extended classifier. Each of the 10 single digit classifiers will return a value between 0&1 for each image presented to it. Matrix algebra offers a particularly simple way to put the single-digit classifiers in parallel: - Write a function to create and return the encoded- Y matrix, that as before, has 60K rows (one per training sample) \& has number of columns equal to the classes of images (you have 10 classes: representing digits from 0 to 9 ). Each row of this matrix will have exactly one " 1 ". - Remember, the labels are in Ytrain. In the function, and the label number corresponds to the column number in the encoded- Y matri. Use these to create the encoded- Y matrix which is numImages nunclasses =mc. - These being matrices, the right hand side of the regression equation; X must have the same dimensions as the encoded-Y. encodedY(mc)=X=(mf)(fc) Where f are the number of features (i.e. columns) of X. You will need to initialize a matrix of appropriate size