Part I: Select a dataset

Select a data set to be used to show your problem. You can choose something from the UCI

repository.

Part II: Train a multi$-$layer perceptron for this problem.

Use a one hidden layer perceptron as your base classifier to train an AdaBoost ensemble learner.

Report your results.

Part III: Train a random decision forest where each decision in the forest is in turn is a trainable

perceptron.

This time instead of using a comparison decision at each node, train a perceptron for each node to

make the best decision. Report your results.