CS5590 Foundations of Machine Learning

$ 24.99
Category:

Description

Assignment 3

The test accuracies corresponding to the given width and depth are shown below in the table. Batch Size – 128 Epochs – 10
n- number of input features

Width—->
Depth n/4 n/2 n 2n 4n
2 86.05 86.5 85 85.5 87.99
4 87.5 83.99 86.05 80 85.05
8 85.5 86 77.99 80.5 50
16 50 50 50 50 50
32 50 50 50 50 50

As observed the highest accuracy was obtained with 4 n width and 2 layers which is 87.99%. It is comparable to logistic regression from Assignment 1 with hyperparameter C=1000. The accuracy with 2 n width and 4 depth and 8 depth respectively i.e., 80 % and 80.5 % are almost equal to the logistic regression with C=1000 which was 81%. The accuracy with n width and 8 depth i.e., 77.99 % is almost equal to the logistic regression with C=100 i.e . 77.5% .

Reviews

There are no reviews yet.

Be the first to review “CS5590 Foundations of Machine Learning”

Your email address will not be published. Required fields are marked *