$2.$ Consider the training dataset shown in the table. The last attribute Cheat is the class attribute. For a binary split of these data using attribute Refund, compute the gain in purity using Gini index.

$\backslash$begin$\{$tabular$\}\{|$l$|$l$|$l$|$l$|$l$|\}$

$\backslash$hline Tid & Refund & Marital & $\backslash$multicolumn$\{2\}\{|$c$|\}\{$ Taxable $\}\backslash \backslash$

$\backslash$hline & & Status & Income & Cheat $\backslash \backslash$

$\backslash$hline $1$ & Yes & Single & $125$ K & No $\backslash \backslash$

$2$ & No & Married & $100$ K & No $\backslash \backslash$

$3$ & No & Single & $70$ K & No $\backslash \backslash$

$4$ & Yes & Married & $120$ K & No $\backslash \backslash$

$5$ & No & Divorced & $95$ K & Yes $\backslash \backslash$

$6$ & No & Married & $60$ K & No $\backslash \backslash$

$7$ & Yes & Divorced & $220$ K & No $\backslash \backslash$

$8$ & No & Single & $85$ K & Yes $\backslash \backslash$

$9$ & No & Married & $75$ K & No $\backslash \backslash$

$10$ & No & Single & $90$ K & Yes $\backslash \backslash$

$\backslash$hline

$\backslash$end$\{$tabular$\}$

Training Data