Assignment
1. Consider the data shown in the following table where each pattern is
represented by 4 binary features f1, f2, f3, and f4. Suggest:
(a) a vector-space representation
(b) a logic based representation
(c) Which of the above two representations is suited well for the data?
Why?
pattern no f1 f2 f3 f4 Class label
1 1 0 1 0 1
2 0 1 0 1 2
3 1 0 1 0 1
4 1 0 1 0 1
5 0 1 0 1 2
2. Consider the table below of 4 transactions in a supermarket where we
consider only 4 items, i1, i2, i3, and i4. There are two classes. Obtain
the frequent patterns by considering items with frequency 2 and above.
Is it possible to use the resultant FP-Tree in classification?
Transaction no i1 i2 i3 i4 Class label
1 1 0 1 0 inexpensive
2 0 1 0 1 expensive
3 1 0 1 0 inexpensive
4 0 1 0 1 expensive
3. Consider a two-class problem where the classes are labeled pen drive
and laptop. Suggest a set of features that could be used to discriminate
between these two classes of objects. What is the corresponding vector-
space representation?
4. For what kind of datasets does the FP-Tree be large in size? Provide
such a dataset.
5. Given a dataset, will the corresponding MST be unique? Why is it so?
If not provide a counterexample.
6. Consider a dataset in a d-dimensional space. What is the worst-case
size of a K-d tree built on such a dataset?
Docsity.com