The objective of this classification task is to determine the specific forest type at a particular

geographic location, based on a set of $27$ input attributes. There are four classes of forest

types: Sugi $(\text{'}$s$\text{'}),$ Hinoki $(\text{'}$h$\text{'}),$ mixed deciduous $(\text{'}$d$\text{'}),$ and non$-$forest land $(\text{'}$o$\text{'}).$ The first

nine input attributes correspond to remote sensing measurements from multiple spectral

bands, and the remaining attributes correspond to interpolated spectral information. The data

set is partitioned into a training set and test set.

You can implement a decision tree model using the Python package scikit$-$learn, and

visualize the model using a suitable package, e$.$g$.,$ python$-$graphviz.

You should submit a report to summarize your work. The following tasks are to be performed:

$($a$)$ Construct multiple decision trees based on the default training set$/$test set partition

using different parameter settings. Compare the structures and classification

performances of these decision trees. $(25\%)$

$($b$)$ Exchange the training and test set and repeat the tasks in $($a$).(25\%)$