Data Mining(Knowledge Discovery in Database)

1
Data Mining(Knowledge Discovery in Database)
2000. 4.20
???? ??? ???
???? 4?? ???
????? ?????
2
?????? ????? Data Mining
??? ??, ??, New ??? ?? ??? ??? ??? ??? ??? New
??? ??, ??? ?? ??? ???
??? ??,?? (??)
???,??? ????
Capture
Sharing
New ??? ??? ?? ??
New ??? ??, ??
Leverage
Creating
Data Mining
???? Knowledge Integration
??System Knowledge Embedding
??,??
????
????
3
??
??? ??? ??? Fortune ? 500? ?? 80? 2000??? ???
???? ?? ??? ??? ??? ??. ?? ?? ???? D/W?? ???
???? ??? ???? ? ??? ???? ?? ??? 3-5? ??? ???
??? ?? ???? ??? ??? ?? 5? ?? ?? ??? ??? ??? ???
????, ???? ?? 5? ??? ??? 10? ?? ? ???? ?? ??. -
1997?? Gartner Group ??? 1999 2003 ?? ?? ???
??? ??? ?? ??? ?? ? ??? ?????? CIO? ?? 10? ?
??? ???? 7?? ?? - Gartner Group ?? ??
4
??
??? ???, ????? ???, ???? ??? ????? ??? ???? ????
???? ?? ??? ???

• ??? ??? ??? ??
• ?? ??? ??? ?? ?? ??? ??
• ? ?? ?? ?? ?? ???
• ?? Database Data?? ??
• ? ??, ?? ??

??? ???, ?? ?? ??? ??
Companys Core Competence
Database (Raw Data)
D/W
5
KDD(Data Mining)? ??
The nontrivial process of identifying valid,
novel, potentially useful, and ultimately
understandable patterns in data(Large
Database). ? ??? Data Mining a process of
kinding patterns and regularities in sets of
data. ? Goals of KDD (or Data Mining) Acquiring
an insight on data(???? ?? ??) Getting previously
unknown or hidden pattern(??? ?? ??) Extracting
useful information from data(??? ??? ??) KDD(or
Data Mining) AI statistics database
techniques ? ??? Knowledge Discovery in
Databases (KDD) Data Mining Knowledge
Extraction Data Pattern Processing Data
Archaeology Information Harvesting Siftware Data
Dredging
6
Data Mining ???
Discovery
Data Mining
Knowledge
Summarizing
Testing /Verifying
Cleansing
Data Warehouse

• Knowledge Discovery in Database(KDD)?? ????? ???
• - Data Mining? KDD?? ?? Operation
• Data ??,?? ??? KDD? 80??

7
KDD Process(1)
Domain ??
??? ?? ?? ??, ???(????)? ?? ??
Data ??
?? ??? ?? ?????, ??? ??? ???? ?? ??? ??
Data ??
??? ??? ??, ?? ???? ?? ?? ? ??
??? ?? ??
???? ?? ???? ??? ? ?? ?? ? ??? ??? ??.
???? ??
??? ??? ??? ? ??, ??? ? ??? ??? ??? ????(NN,
ID3, etc.) ??.
Data ??
??? ?? ?? ??? ? ??? ????? ?? ??? ??
8
KDD Process(2)
??? ???
??? ??(???? ??)? ??? ??? ?? ? ??? ???? ??? ????,
?? ???? ?? ??? ??
??? ??
??? ??? ?? ???? ?? ?? ? ????? ??? ??
??? ??
?? ???? ?? ?? ??? ???? ??? ?? ? ??? ??, ?? ??
??? ?? ??? ??, ??? ??? ??
? Process? ??? ???? ??? ??????? ??? ??? ??
??? ?? ????.
9
Techniques of Data Mining
Prediction
Classification
- Class Model ??, Similarity Search
using some variables or fields in the database
to predict unknown or future values of
other variables of interest
- ??? ?? ??? ?? - ????, ?? ??, ?? ??
Regression
- ??? ?? ?? - ?? ??, ?? ??
Time Series Analysis
- ??? ??? ????? ??? - Neural Network, ID3
Clustering
Description
Association Analysis
- ??? ?? ?? ??? ??
finding human- interpretable patterns describing
the data
Sequence Analysis
- ??? ?? ? ??? ???? ??
-?? ??? ?? ? ?? ?? - statistical techniques
Deviation Detection
-???? ??? ?? ? -Histogram, Scatter Plot Display
Visualization
10
??? ??? ????? ??
?? The science of learning from data(?????? ???
??) Design of experiments, analysis and inference
from experimental or observational data
Data Mining? ??? ??
??
???
??? ???
?? ???
observational experimental data
observational
??? ?
small datasets
large datasets
??
probabilistic inference based on population
models
Inference based on various models
11
Machine Learning? ??? ????? ??
?? ??? ??? ???? ??
??
?? ??
??? ???
???(??)
improving the performance of an agent
finding understandable knowledge
??? ??
ML typically (but not always) looked at smaller
data sets(training set)
very large, real-world databases (efficiency? ??)
??? ??
???? ????, ? ???, ??? (Training Set)
???? ??, Noisy??, ????? ?? ??? ?? ?? ??.
??????
?? Application? ??? ????? ??, Learning? ???? ?? ??
?? ???? ??????? ??
??
?? ?? ??( learning from examples, reinforcement
learning, learning with teacher, etc.
?????? learning(learning from examples)
? KDD is that part of ML which is concerned with
finding understandable knowledge in large sets
of real-world examples.
12
CBR vs. Data Mining

• Data mining- ??? ??? ??
• CBR - ?? ??? ??, ??, ???, ??, ??? ?? ??? ??? ???
  ??
• Case based reasoning (CBR)? ??
• Rule base reasoning(RBR)? ?? ?? ??? ?? ?? ????
  ??? ?? ?? ??? ???
• Ex) acquiring cases, case vocabulary, retrieval
  knowledge, adaptation knowledge
• Case Filtering, acquiring solving strategy or
  background knowledge
• ? ??? ???? ??
• ??? ????? CBR? ??
• ??? ??? ??? ??? ??? ??? ???? ??? ?? data sets?
  ??? ??. ??? ??????? ???? ??? ??? ??? ???? ?? ????
  ?? ?? ??? ???? ???? ??? ???? ?? ??? ?? ? ?? ???
• ? CBR? ?? ?? ??? ??
• ????? ?? ??? ???
• ?? ??? ???
• ???? ?? ??? ??? ??? ???.
• ??? ?? ? ??? ??? ??.
• ??? ???? ?? ???? ??? ??? ?? ??? ?? ? ??? ??? ??
• ??/?? ??

13
Applications
??? ??? ??, ?? ?? ?? ??? ?? ??(genetic sequence
analysis)
Medicine/Biology
Scientific Discovery
superconductivity research
?? ??, ?? ?? ??, ?? ??, target mailing
Marketing/Sales
?? ?? ??, ?? ??, ?? ??, ?? ?? ??
Finance/Banking
????? ?? ??? ???? ??, ?? ????? ?? ?? ?, ?? ??
???? ??? ??? ??, ??? ?? ??
Engineering
Security
?? ??
14
?? ??
Improve product quality and customer service via
a data mining
LTV Steel Corp. In Cleveland
?? ? 3? ?? ????(http//www.ltvsteel.com/) - with
4 billion in revenue(1996)
? ??1 ?? ??? ?? ???? ??? 3-5?? ?????? ?? ???
???? ?? ? Data Mining? ?? ??? 0.03 - 0.05 ???
??? 10 million? ??. ? ??2 ? ??(East
West)? ??? ?(7-8 2-3) ?? ? ???? ??? ??? ???
??? ? West? ? 2??? ??? ????? East? ?? ??
??. ? 16.8 million??
15
Data mining ?? ??
???
?? ??
??????
KDD
?? ??
?? ??
?? ??
???(??) ??
16
KDD ?? Tool
Product
Company
URL
Clementine
Integral Solutions, Ltd.
http//www.isl.co.uk/clem.html
Darwin
Thinking Machines, Corp.
http//www.think.com/html/products/products.htm
DataCruncher
DataMind
http//www.datamindcorp.com
DBMiner
DBMiner Technology Inc.
http//fas.sfu.ca/0h/cs/research/groups/DB/DBMiner
/
Enterprise Miner
SAS Institute
http//www.sas.com/software/components/miner.html
GainSmatrs
Urban Science
http//www.urbanscience.com/main/gainpage.htm
Intelligent Miner
IBM
http//www.software.ibm.com/data/iminer/
MineSet
Silicon Graphics, Inc.
http//www.sgi.com/Products/software/MineSet/
Model 1
Group 1/Unica Technologies
http//www.unica-usa.com/model1.htm
ModelQuest
AbTech Corp.
http//www.abtech.com
PRW
Unica Technologies, Inc.
http//www.unica-usa.com/prodinfo.htm
CART
Salford Systems
http//www.salford-systems.com
NeuroShell
Ward Systems Group, Inc.
http//www.wardsystems.com/neuroshe.htm
OLPARS
PAR Government Systems
mailto//olpars_at_partech.com
Quest
IBM
http//www.almaden.ibm.com/cs/quest/
Scenario
Cognos
http//www.cognos.com/busintell/products/index.htm
l
See5
RuleQuest Research
http//www.rulequest.com/see5-info.html
S-Plus
MathSoft
http//www.mathsoft.com/splus/
Wiz Why
WizSoft
http//www.wizsoft.com/why.html
17
Research Issues and Challenges

• Research Issues
• - ??? ???? ???,??? ??? ??
• - ???? ??? ?? ??? ??
• - ??? ??? ??? ??
• - ?? ???, ?? ???, ??? ??? ??? ?? ??? ??
• - ?? ?? ??? ??? ? ? ??
• - ???? ??? ??? ??
• - ?? ?? ??? ? ? ???? ??
• - ??? ?? ???, multimedia ??? ?? ??
• - Web ? document??? Mining
• ?? ? ???? ? ??
• ??? ???? ????? ?? ???? ???? ?? ??? ?? ??? ?? ??
  ??
• ????? ?? ???? ??? ?? ? ??
• ??? ???? ??? ??? ??? ?? ???? ???? ?? ???? ????
• ???? ??? ??? ??? ????? ?? ??? ?? ???? ????? ???.
  
• ???? ?? ??, ???? ?? ?? ??, ??? ??? ??
• ??? ?? ??? ??? ??
• ?? ??(?? ?? ??)? ??

18
Conclusion
-????? ??? ???? ??? ??? ????? ??? ? ?? ??? ???
??? ?????? ??? ?? ?? ????? ? ?? ??? ??? ? ???
??? - ??? Warehouse? DSS? ???? ??? ? ?? ??? -
??? ???? ?? ????? ?? ?????? 10?? ???? ??. - ???
??? ??? ???? ???? ??(??, ?? ??? 80) - ??? ???
????? ? ?? ???? ?????, ??? ??? ?? ? ?? ? ????
??? ??? ??? ???? ??? ??, ?????? ???? ??
???? ???? ?????? ?? ??? ??? ??? ? ??. ?? ??
???(???????) ??? ???? ???? ???? ??? ??. Masato
Koda(University of Tsukuba)