TU Wien:Machine Learning VU (Musliu)/Exam 2019-06-27

Prüfung 2019-06-27: True/False questions:

• Perceptron with Softmax solves XOR-Problem
• Lazy learners good for big data
• D-Separation used in neighbourhood searches in a Bayes net
• Naive Bayes, Lapace correction must be used
• Paired T-tests used at folds verification??

2D plots with dots. Plot the decision boundary of Decision tree, SVM linear, MLP and Knn(1)

Decision Tree: k features >> N samples. A) If all k are binary, how many leaves are possible, what is max depth? B) If k are continuous, how many max leaves and max depth?

Describe gradient descent in linear regression

Describe optimization of hyper parameters with Bayes optimization

Dataset with Marians and Humans. Features Height, number of legs, smell, color. Task a) Generate 1R. Calculate Accuracy and Precision b) Generate covering algorithms to sepaerate Marians from humans