Problem $2[5$ pts$,$ Peer Review$]$ : Building an Random Forest Classifier to classify MNIST digits $3$ and $8.$

Remember that training the random forest algorithms involves the following steps:

for k$=1$ to K:

$$ a$)$ build kth tree of depth d

$$ b$)$ Return the kth tree trained on the subset of dataset with the random feature splits

Predicting the classification result on new data involves returning the majority vote by all the trees in the random forest.

Part A $[5$ points, Peer Review$]$: Complete the create$\_$tree function to build a new tree trained on a subset of data. Within this function a decision tree classifier is built and trained on the subset of data with the subset of features. Answer the Peer Review question for this section.