describe hmm (5p)
random forests in detail + compare to  similar  algorithms (I used 1R and  decision trees) (6p)
compare svm with perceptron  what  is common what differs
what can be used to create  a bayesian networks compare it
metalearning something model based
given dataset   train  1R, naive bayes   apply models to  another dataset   calculate  recal from result (16p!)(the  constant was necessary)
Lazy learner   and an example(knn)
feature selection