?? ???? (Data Mining Techniques ) : An Overview

1
?? ???? (Data Mining Techniques ) An Overview

• ????????
• ? ??
• jkim_at_kordic.re.kr

2
????? ?? ?? ??
???
????
??

• ????
• Point of Sale
• ATM
• ????
• ????
• ??
• ????
• ????
• ??????

• A?? ???? 80? B??? ????
• ????? ??? ???? 6??? ??
• A??? ?? ??? B??? 2?
• ?? ??? ??? ??

• ????? ?
• ??? ??
• ??? ?? ??? ?
• ????? ????? ?
• ??? ?? ???? ?
• ??? ?

3
Data Mining ?? ?

• ??? ??????
• ??? ??? ????
• ???? ?? ??? ????
• ??? ??????? ????
• ???
• ??? ????? ??, ??, ??, ??,??? ???

4
? ?

• ???? ??? ?? ??? ??
• ?????? ???? ?? ??
• ??? ??? ???
• ??, ??? ?? (???)
• ??? ??? ??
• ????? ?? ??
• ????(Machine Learning) ??? ??
• Knowledge Discovery, Knowledge Extraction,
  Machine Learning, Data/Pattern Analysis

5
Data Mining ??

• 1) ??? ??
• - ??? ??? ?? ??
• 2) ??? ??
• - ?? ?? ??? ?? ?? ??
• - ???, ???, ???,
• 3) ?? ??
• 4) ?? ??
• - ??? (??), ?? ??
• - ??, ???

6
Data Mining ??
Select
Transform
Mine
Assimilate
????
????
????
?? ? ??
DATABASE
??? ???
Extracted Data
Selected Data
Assimilated data
Transformed Data
Visualization
???
7
? ? ? ?
Target Audience
Customers Purchased frozen orange juice in 12oz
cans

• Purchase History
• Point of Sale Data
• Survey data

60? ??? ??? ?? ??
Royal Customers (buys the same brand 80 of time)
8
Data Mining?? ??

• u
• ??? ??, ??? ??? ???
• u
• ??? ??????? ??? ???
• u
• ??? ?? ??? ???
• u

9
??? ??? ??? ???
Data Mining?? ??

• Association(??? ??)
• Characterization(????)
• Classification(??)
• Summarization(??)
• Clustering(???)
• Sequential Pattern Discovery(??????)
• Trend(?? ??)
• Deviation Detection(??????)

10
??? DB? ??? ???
Data Mining????

• Relational DB
• transactional DB
• Object-oriented DB
• Spatial DB
• Temporal DB
• Textual vs Multimedia
• Hetrogeneous,

11
?? ??? ???
Data Mining ????

• ????? ????? ???
• ??? ??, rule induction
• ????/??? ??
• Statistical Classification(supervised learning)
• Clustering Techniques(unsupervised learning)
• Time Series Analysis
• ???? ??
• ????? ??? functional mapping? ??
• ??? ?? algorithm? ??

12
??? ?? ??

• Transaction DB? ????
• RULE ??? ??
• A gt B support, confidence
• support (A and B) / (total transactions)
• confidence (A and B) / (A)
• ? milk gt bread 7, 70
• ?? 1 ??????? ????
• ?? 2 ??? ??????
• ???? ??? ???? ?? ??

13
??? ?? ??
Association Rules with Maximum support of 50
?? ??
14
Data Classification

• ?????? ??? ??? ??
• ????? Class-label ? feature set?? ??
• ????(Supervised Learning)
• ????? ??? ??, ??? ??
• ??? ??? ??? ? ??? ?? ??
• ?? Credit Approval, ?? ??
• ? ??? ????? ????? ?? ???? ?? ?? ??
• Decision Tree, ???, ??? ???

15
Classification Example
Classifier
Class 1 ??? ??
Class 2 ??? ??
Class 3 ??? ??
16
Decision Tree Classifier

• ?????? Decision Tree ???? ??
• ID3 algorithm

17
Neural Network Classifier

• ??? ?????? ??? ???? ??
• ??? Neuron? ????? ???
• ?? ???? ??
• Error-back-propagation ????????
• ??? Functional Mapping? ?? ???

18
Neural Network Classifier
Input layer
Hidden layer
Output layer
19
Sequential Pattern Discovery

• Transaction ????? ??? ?? ??
• ??
• ??????? ?? ?? ??
• ???? ?? ??
• ?? ??? ?? ?? ??, ??
• ??? ??? ?? ??, ??
• ???
• ??? ??? ??
• Hidden Markov Model for doubly stochastic process
  modeling

20
Sequential Pattern Example
Sequential Pattern in DataBase
21
Similar Time Series
Matching Curve Found
22
Clustering(???)

• ??? ???? ? ???? ??
• ????? ??? ???
• Unsupervised Learning Algorithms
• Symbolic
• Neural Network based (Kohonen Feature Map)
• ??
• ???? ??? ??? - ?? ??? ??
• ??? ???, ????? ?? ?? ????

23
Clustering Example
24
Symbolic Clustering
Similarity 2
Similarity 2
Diff3
Diff2.83
Diff3
Similarity 3
Total Score for this cluster partition average
similarity average difference

2.33 2.94 5.27
25
Data Mining Interface

• Interactive Mining
• GUI? ?? Task? ??
• Data Mining Query Language
• find association rules
• related to gpa, birth_place, family_income
• from student
• where major CS and birth_place Seoul
• with support threshold 0.05
• with confidence threshold 0.7

26
Kohonens Feature Map

• ???? ??? ??? ??
• ??? ??? ??? ???? ???? ??
• ??? Feature Map??? ? ??? ??
• ???? ??
• Feature Map ?? ??? ?? Difference
• ????? ?? ??
• 1) ??? ?? X? ?? ? ?? ?? N? ??
• 2) N? ? N? ???? ????? X? ???? ??
• 3) ?? ??? ??? ??? ??? ?? ?? ??

27
???? ??? ?? ??

• ?????? ?
• ??? ????? ??
• ? ??? ???? ?
• ?? ??? ??? ???? ??? ????? ?
• ?? ??? ?? ??, ??? ???? ??
• ?? ??? ??? ????? ?
• ?? ??
• ??? ??????? ????? ???? ???????
• ?? ??, ????, ?? ??, ?? ??

28
??? ?? ??
Scoring???
???? ???? ???? ??? ??
?? ?? ?? ?? ????
????
? ??? ????
29
??????? ??? Overview
???? ???? ???? ?????
???? DB
Credit ???
Decision Tree
??? ??
???? ??
?? ??? Scoring (Neural Network
Scoring ???
Credit ?? ? ???? ??
30
?? ?? ???? ????

• LG?????
• ???? ????? ??? ??
• ?? ???? ???? ???? ?? ?? ??
• ????? ?? ??
• ????, ????, ??? ??, ??? ??
• ??? ???? Fraud Score ??
• 1995? LG???? ???? 14???? ??
• ?? ??? ??

31
IBM Intelligent Miner
32
?? ??

• Mining Business Databases, Brachman, et al.,
  CACM, Vol39, No11, 1996
• Mining Scientific Data, Fayyad, et al., CACM,
  Vol39, No11, 1996
• Quest(IBM Almaden)
• http//www.almaden.ibm.com/cs/quest
• DBMiner(Simon Fraser Univ.)
• http//db.cs.sfu.ca/DBMiner
• KDD(GTE)
• http//info.gte.com/kdd/index.html
• International Conference on Knowledge Discovery
  and Data Mining
• Advances in Knowledge Discovery and Data Mining,
  MIT press, 1996

33
? ?

• ??? ?? ?? gt ??, ??? ?? ??
• ??????? ??? ??
• ????, DB ??? ?? ??
• ???? ??? ??? ?? ??? ??
• ?? ?? ??? ?? ?
• ??? ?????? ?? ?? ??
• Hot Research Item