Download Mining Serial Episode Rules with Time Lags over Multiple Data

Efficient Mining of High Utility Itemsets from
Large Datasets
Alva Erwin
Department ofComputing
Raj P. Gopalan, and N.R. Achuthan
Department of Mathematics and Statistics
Curtin University of Technology Kent St. Bentley Western Australia
PAKDD08
1
Outline
 Introduction
 Preliminaries
 Method – Compressed Transaction Utility-Prol
 Experiments
 Conclusions
2
Introduction
 The goal of frequent itemset mining is to find items that
co-occur in a transaction database above a user given
frequency threshold, without considering the quantity or
weight such as profit of the items.
 Quantity and weight are significant for addressing real
world decision problems that require maximizing the
utility in an organization.
 TwoPhase based on Apriori is suitable for sparse data sets
with short patterns, CTU-Mine based on the pattern
growth is suitable for dense data.
3
Definition


4
u(3 4, t1) =$60
u(3 4, t3)=$60
u(3 4) = $120 ,
Definition
 Transaction Utility :
 Transaction weighted Utility:
twu(X) 
 tu(T
Tq D  X  Tq
 tu(1) = 80
twu(3 4)=$190
5
q
)
Compressed Transaction Utility-Prol
6
GlobalItem
index
1
2
3
4
5
-
Original item id
5
1
2
4
3
6
Profit
5
10 150
35
25
2
Quantity
60
12
5
4
2
TWU
98
7
96 810 595 422 9
4
9
4
99<min_Utility(129.9)
Compressed Utility Pattern-Tree
 Parallel projection of transaction database
7
CUP-tree
 Traverse index 1 (110) from 5, 2 (310) from (2,3,4),
 3 (195) from 2, and 4 (190)from (3,5)
8
ProCUP-tree
 index 1 (110) from 5, cause 110<min_Utility(129.9)
 2 (310) from (2,3,4),3 (195) from 2, and 4 (190)from (3,5)
9
ProCUP-tree

GlobalItem index
1
2
3
4
5
Original item id
5
1
2
4
3
ProItem index
--
1
2
3
--
Profit
5
10
150
35
25
Quantity
60
12
4
5
4
TWU
987
964
810
595
422
 oriUtility*itemQuantity + proUtility*proQuantity = Utility
 35*2+25*2=120, 150*1+25*1=175, 10*5+25*3=125
 High_Utility_Itemset = (3,2) (3,2,1)
10
Experiments
11
Conclusion
 CTU-Pro algorithm to mine the complete set of high utility
itemsets from both sparse and relatively dense datasets with short
or longer high utility patterns.
 The algorithm adapts to large data by constructing parallel
subdivisions on disk that can be mined independently.
12