Q$\mathrm{.}1.$ Consider the dataset below that categorizes dogs into two classes: "danger" and "safe", based on their features.

Test Data

$1.$ Use the ID$3$ algorithm to construct a decision train on the training data above.

$2.$ Make sure that you use log$2($base $2)$ in your calculations of entropy.

$3.$ In no more than $2$ pages. Show your calculations. Take a note of the calculated entropy, weighted means of entropy, and information gains.

$4.$ If there is a tie between any two attributes in information gain, we consider them from left to right as they appear in the table. For example if splitting on Color, and on Body Shape give the same smallest possible gain, we choose to split on Color.

$5.$ In your calculations keep the computed values within $4$ decimal digits. For example, say an entropy was found to be $\backslash (\backslash$mathbf$\{$E$\}=3\mathrm{.}9753711\backslash ).$ We drop any digit after the $\backslash (4^\{\backslash$text $\{$b $\}\}\backslash )$ decimal digit with no rounding so $\backslash ($ E $\backslash )$ becomes $\backslash (\backslash$mathrm$\{$E$\}=\backslash )3\mathrm{.}9753.$ Make sure that you apply this rule before using any computed value in any further computation. This rule applies to any calculated entropy, weighted mean of entropy, and gain.

$6.$ If you used Python to perform any calculations, compile a list of the main commands you used in a separate page, to appear as page number three.

$7.$ On one separate page draw the entire decision tree. In total no more than $4$ pages for this problem. $($Calculations $2$ pages, code $1$ page, tree $1$ page$)$

$8.$ Answer any questions about your decision tree and work in Blackboard.