Step $1$: COVID$-19$ code

We use the COVID$-19$ prediction code that we discussed in class $($covid$19$ClassExample$\_$MLPRegress.py$).$

Use only US$,$ Spain, Italy, and Canada for test.

Use $5-$fold cross validation for validation purposes on the training dataset

Using the $5-$fold cross validation compare the following problem settings:

S$1$: Two hidden layers with $15$ neurons in each layer $($current setting$)$

S$2$: Two hidden layers with $10$ neurons in each layer

S$3$: Two hidden layers with $10$ neurons in the first layer and $15$ neurons in the second layer

S$4$: Two hidden layers with $15$ neurons in the first layer and $10$ neurons in the second layer

S$5$: Three hidden layers with $15$ neurons in each layer

S$6$: Three hidden layers with $10$ neurons in each layer

S$7$: One hidden layer with $10$ neurons

S$8$: One hidden layer with $15$ neurons

Compare the above settings using the score $($R$2)$ metric and select the best setting

Use the neural network with the best setting to predict the new cases of COVID$-19$ for US$,$ Spain, Italy, and Canada. Measure their performance in terms of average absolute error. Draw a bar chart to compare the corresponding errors.

Report all the performances of cross validation in a proper table.