Information Technology Reference
In-Depth Information
(1)
TB
(
D
) = 1, when the workload across all processors are the same.
(2)
TB
(
D
) = 0, when the workload is concentrated on one processor.
(3) 0<
TB
(
D
)<1, in all the other case.
The data skewness metric and workload balance factor are not independent.
Theoretically, each one of them could have values range from 0 and 1. However,
some combinations of their values are not admissable.
Association rules mining has become a mature field of research with diverse
branches of specialization. The fundamentals of association mining are now well
established and, with some important exceptions. Current research appears to
focus on the specialization of fundamental association mining algorithms, many
areas of which are still emerging. These include fields such as, measures of
interest, the inclusion of domain knowledge and semantics, quantitative mining,
disassociation mining, privacy mining, incremental mining, iterative and
interactive or guided mining, and higher order mining.
Exercises
1. Illustrate what the support, confidence, minsup, minconf and large item set
are.
2. The following table shows four transactions with the misup being 60%, please
find out one frequent item set
TID
Date
Items_bought
1
01/05/2005
I
1
, I
2
, I
4
, I
6
2
02/05/2005
I
1
, I
2
, I
3
, I
4
, I
5
3
03/05/2005
I
1
, I
2
, I
3
, I
5
4
04/05/2005
I
1
, I
2
, I
4
3. Try to describe classical Apriori algorithm.
4. Try to do association rule mining based on the distributed system.
5. What are the positive and negative relations? Can you give an example?
6. Summarize how to effectively mine the following rule, “A free commodity
possibly triggers 200 total shopping in identical transaction” (agrees the price
of each kind of commodity not negative)
