$1.$ In the "perturb the current feature subset" step of the SA algorithm, if the random percentage is $\backslash (4\backslash \%\backslash )$ and the number of features is $200,$ what is the number of features that will be added or deleted from the current subset of features?

$2.$ Calculate $\backslash (\backslash$operatorname$\{$Pr$\}[\backslash )[$accept$]$ for iteration $4,$ where the accuracy for iteration $3$ was is $0\mathrm{.}820$ and the new accuracy is $\backslash (0\mathrm{.}790($c$=1)\backslash ).$

$3.$ Determine the Status $($Improved$,$ Accept, or Reject$)$ for the table to the right.

if performance is better than the previous subset then Accept new subset; Improved else

Calculate acceptance probability;

if random uniform variable $\backslash (>\backslash )$ probability then

Reject new subset; Reject else

Accept new subset; Accept end

end

In$-$Class Problem

Note:

$-$ Treat each row as a separate iteration

$-$ The Old Acc is the previous New Acc regardless of the Status