Transcript Mining Fuzzy Multiple-Level Association Rules from Quantitative Data

Mining Fuzzy Multiple-Level
Association Rules from
Quantitative Data
Author: TZUNG-PEI HONG
KUEI-YING LIN
BEEN-CHIAN CHIEN
Advisor: Dr. Hsu
Graduate: Yan Pin Huang
ADSL
Wednesday, October 01, 2003
Content








Motivation
Objective
Introduction
Notation
The Multiple-Level Fuzzy Data-Mining Algorithm
Experimental Results
Conclusion
Personal opinion
Motivation


Machine-learning and data-mining techniques have
been developed to turn data into useful taskoriented knowledge.
Most algorithms for mining association rules identify
relationships among transactions using binary
values and find rules at a single-concept level.
Objective



This paper proposes a fuzzy multiple-level mining
algorithm for extracting knowledge implicit in
transactions stored as quantitative values.
The proposed algorithm adopts a top-down
progressively deepening approach to finding large
itemsets.
It integrates fuzzy-set concepts, data-mining
technologies and multiple-level taxonomy to find
fuzzy association rules from transaction data sets.
Introduction

Proposed a method for mining association
rules from data sets using quantita-tive and
categorical attributes.

R. Srikant and R. Agrawal, “Mining quantitative
association rules in large relational tables,” in The
1996 ACM SIGMOD International Conference on
Management of Data.
Introduction(cont.)

Fuzzy set theory is being used more and
more fre-quently in intelligent systems
because of its simplicity and similarity to
human reasoning [15].
Mining at Multiple Concept Levels

They divided the min-ing process into two
phases.


Candidate itemsets were generated and counted
by scanning the transaction data.
Association rules were induced from the large
itemsets found in the first phase
Mining at Multiple Concept Levels
Notation(cont.)
Notation(cont.)
The Multiple-Level Fuzzy
Data-Mining Algorithm
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)

Step 1. Each item
name is first encoded
using the predefined
taxonomy. Results are
shown in Table 2.
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)
1**
11*
111
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)



Step 2. All transactions shown in Table 1 are then
encoded using the above coding scheme
Step 3. k is initially set at 1, where k is used to store the
level number being processed.
Step 4. All the items in the transactions are first
grouped on level one.
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)

Step 5. The quantitative values of the items on level
1 are represented using fuzzy sets.
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)



(1,1)(6,0) 帶入 y=ax+b 求解a,b (a=-1/5 b=6/5)
Function: y= -1/5x+6/5
(1**,5)帶入member function求得 y=0.2
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)


Step 6. The scalar
cardinality of each
fuzzy region in the
transactions is
calculated as the count
value.
Its scalar
cardinality=(0.8+0.8+
0.0+0.2+0.0+0.0) =1.8
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)

Step 7. The fuzzy
region with the highest
count among the three
possible regions for
each item is found.
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)

Step 8. The count of any region selected in Step 7 is
checked against the predefined minimum support
value α. (α =2.1)
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)
1.2
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)


Step 12. r is set at 2, where r is used to store the
number of items kept in the current itemsets.
Step 13. Since is null, k =k + 1=2 and Step 4 is done.
The results for level 2 are shown in Table 10.The
results for level 3 are shown in Table 11.Since there are
no items on level 4, Step 17 is done.
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)
Step
17. The association rules are constructed for each
large itemset using the following substeps.
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)
The Multiple-Level Fuzzy
Data-Mining Algorithm(cont)

Step 18. The confidence values of the possible
association rules are checked against the
predefined confidence threshold λ. (λ =0.7)
6. Experimental Results
They were implemented in C on a Pentium-III 700
Personal Computer.
The number of levels was set at 3. 64 purchased items
(terminal nodes) on level 3, 16generalized items on level
2, and 4 generalized items on level 1.
Experimental Results(cont)
Experimental Results(cont)
Discussion and Conclusions



Proposed a fuzzy multiplelevel data-mining
algorithm that can process transaction data with
quantitative values and discover interesting patterns
among them.
This method achieves better time complexity since
only the most important fuzzy term is used for each
item.
This proposed algorithm does not find association
rules for items on the same paths in given hierarchy
trees.
Discussion and Conclusions


We will therefore attempt to dynamically
adjust the membership functions in the
proposed mining algorithm
We will also attempt to design specific datamining models for various problem domains.
Personal opinion





Find association rules for items on the same paths in
given hierarchy trees.
Find a method that can dynamically adjust the
membership functions.
Fuzzy SOM.
Fuzzy clustering.
The strategy of using fuzzy set in Ant Colony
algorithm.
Personal opinion(cont.)
Personal opinion(cont.)