$3\mathrm{.}3$ Construct a decision tree based on the following training samples that obtained from $3$D printing. Please select the best attribute to split the root node $($only the first split$).(20$ points$)$

Sample index Surface roughness Shape deformation level Quality $($Label$)$

$1>100$ High Bad

$2<30$ Low Good

$330<=<=100$ Low Good

$430<=<=100$ High Bad

$5<30$ High Good

$6>100$ High Bad

$7<30$ Low Good

$8<30$ Low Good

$9>100$ High Bad

$1030<=<=100$ High Good

$11<30$ High Good

$1230<=<=100$ Low Good