1. Explain what type of task apriori algorithm is designed to perform, in what way it is better than some other algorithm at performing this task, any why it has this advantage.
2. Create a classification tree for the circled items in the graph. The attributes are the numerical values on the x- and y-axes and the classes are white and black.
3. Compute the
(a) Manhattan, and
(b) Euclidean
distance between the vectors X = (31, 14, 27) and Y = (8, 23, 15).
4. Describe one of the problems with 1-NN classifiers and explain why a k-NN classifier with k > 1 would not have that problem.
5. Explain the difference between an eager and a lazy learner.
6. Assume that data has three numerical attributes and that there are two classes, one of which consists of the points where the third value equals the sum of the first two, so that e.g. (2, 3, 5) and (4, 5, 9) belongs to the described class and (2, 3, 7) does not. Would a decision tree be expected to classify well?
7. Is there a pure strategy to this zero-sum game where the numbers are the payoffs of player B? In other words, is there a saddle point? If so, what is the value of the game?
|
Player A |
| L |
M |
R |
Player B |
U |
10 |
5 |
1 |
| M |
8 |
9 |
3 |
| D |
7 |
4 |
2 |
8. At a caf´e, these numbers of waffles have been sold the last eight years.
2008 2009 2010 2011 2012 2013 2014 2015
875 923 913 958 1,023 978 1,015 1,110
Despite the fact that these are few data points, find a linear regression line for the number of sold waffles as a function of time and use it to predict how many waffles will be sold in year 2017.
9. Explain why scale-free networks are more sensitive to the spread of infections than random networks.