Consider a data set that has $5\%$ missing values for one of the numerical descriptive features. The data

set has $3200$ instances, and there are also outliers for this feature. These outliers have not been removed.

Which of the following approaches should be followed with reference to this descriptive feature if a neural

network is trained on this data set? Just provide the letter$($s$)$ of the correct answer$($s$).(2)$

$($a$)$ Remove the feature from the data set

$($b$)$ Impute using the mean over all of the available values

$($c$)$ Impute using the median over all available values

$($d$)$ Do nothing

$2.$ Consider the plot below.