1. We are given the below transaction database and 3 as the support threshold.
t1 | {1,3,4}
t2 | {1,4,5}
t3 | {2,4}
t4 | {1,4,6}
t5 | {1,6}
t6 | {2,3}
t7 | {1,4,6}
t8 | {2,3}
Perform the Park-Chen-Yu algorithm, supposing that we use the hash functions h1(a,b)=(a+b) mod 3 and h2(a,b)=(a+b) mod 5.
Which hash function seems to be of better use?
What would be the case if we increased the suppor threshold (for instance to 5)?
The frequency of which item pairs would be necessary in order to determine set F_2 (assuming a frequency threshold of 3 and 5)?

2. Decide which item sets are frequent, closed frequent and maximal frequent in the example in the lecture slides.

3. Determine the FP trees for the example transaction database in the lecture slides. Apply a representation of buckets which orders the imets in them
 a) in a decreasing order of their frequency,
 b) in a increasing order of their frequency.
 Based on these trees build a conditionated FP tree!