1 "#$% - - PDF document

1

• Data Mining Lecture 1

2

• !!! "# $ %&'
• ()*+,,-).,-

+)/.+,*))00&,-- )-)

#1!!

2 3 4(1!! $+5-,+,1!!

Data Mining Lecture 1 3

• ,-,

/-+--6 789--(

• .-

29--&8

• &:-+)

2--,+) ;(<8(),,)- 2+,(,--,=''>

Data Mining Lecture 1 4

• Data Mining Lecture 1

5

• -,,-.

(% .,+,?! (- -!

Data Mining Lecture 1 6

!

• --
• )
• @-A

,

• ,(

,

• ,

,

• 2,

,

• A

,

• 7+

,

,

• 3.)&.-

,

2

• Data Mining Lecture 1

8

"#$%

.7)-B 7-B %C*#-(-B %(,) 2,,B ,((-) D-

Data Mining Lecture 1 9

&

C,,);-<*?( .--*.,-( @(? *() +,,-**,,+( ,),*),.,-)! E8,%.-F;),<%#

0.)G'?--(;8) <+)#'-& -+,#! 2+ '''(-,,'-,)-

• ,-.=3!7-(-

)--+.!, ;$,,,).=3,6<

Data Mining Lecture 1 10

'(!)

%8,+, 2--,,,-

• +%@,),--

(--( %@ --* 7-*-+.(#*,-6

Data Mining Lecture 1 11

!*+% 5(-+ ,,--*-

7+-& # +- @# --

• .+-

*(, .3

3.-+H-.(- !!F, !*+% %,,---

• 4@

,, ,# ) 8- (, +)(-

$-,A (+,(*- %),

,()-- )(

3

Data Mining Lecture 1 13

",

• $((-- --,).-
• ,))#*#-.(,(
• (-.,)H-,,

500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000 3,500,000 4,000,000 1995 1996 1997 1998 1999

The Data Gap

Total new disk (TB) since 1995 Number of analysts

From: R. Grossman, C. Kamath, V. Kumar, “Data Mining for Scientific and Engineering Applications”

Data Mining Lecture 1 14

*

%#*,--.)-+

=((8(.,-&., ,.,)#* -,,)(, ((-,-+

2,.-;-<

*,--.)-+%% #*,-8- ,)- ,)---(. +,,!

7-B

%-.A) =8),,5 ,!

Data Mining Lecture 1 15

• I)

7,,-(- I5

I)

3,)-(- /;--<A),

%

C,

%

/,

C

3 +(-+

C

EHH) /+(-+

Data Mining Lecture 1 16

.!-

%+C5$3 C523I)

*)(--=2,,(( ''*-.!%#B E-,,=2*.+(* '!'''#! E-,,=2*.+(!

%I

7=2-,(,( (''B,(

• (),,+)(=2(*.,+)

!, 7-(=2(A,)+) .)(B, 7(A,)+A( (((BA

Data Mining Lecture 1 17

!*/

! %+,)--

#,)- #, #+#,),,# F#,)- E.--* A,),.,) E-:-

• C2,

$8 (, -,(

7+ @,):@&(%/2-,) 2)

Data Mining Lecture 1 18

!012

7-(,)B

• -,),)--

,,,,+,,(),-

$#

E-, (;-,<* ,.,-+!

%.

.(,D+#!

&#,)

2 &,+*-, 3-+-(

4

Data Mining Lecture 1 19

!032

(,

• -()*)(+)*
• , ,(
• ()A
• ();+<-(-((
• A(-*(*,,

*

3.-)(

.,&-,) ,)(-,-)- .

Data Mining Lecture 1 20

!

E,-.,

(,*,)-- ,,)., &,-,)(,&- ,)!

F,

H---

• #-

,-,&+-

• ),).#

Data Mining Lecture 1 21

4012

2,

*-,)-,,--. ,-(-!

2

,-+,--,((-,+.- ,)-,-(),

=8,

• (,*-

,, ),--)

• ,-(,(
• ((

Data Mining Lecture 1 22

%-,

2,,-( )+,#*A- .-2,J )!

%,(-

$,,,-,-(,,- (-)*#!2,)H-.( 8-! @$,-(--(,,* (A&,,,,)+,- +#,,,-,,(-!

F,

2,)KL(,#-

• ,)!

4032

Data Mining Lecture 1 23

• @2-.- ,)H-/@2

+,#--(,.-. (/*M# # -

2)

N35-3,C+.)-.-A *,(-

7+(&2-

@(&2- ,-,7+ ,(#&,--. (-+.,)H((.( 7+#.7+H!

$

Data Mining Lecture 1 24

56')

• %-(+#&,,,*+#
• -*,-*!
• 3(.,+.,(--

#,,!

• ;%-+<!3,(#,.

,!

• #)-,#).),#)+D+)
• ,.,-((+-,,*-
• +*,8+)-)#-

(,,+D!

• ()(,* *-

,,*--.)((( )(-+!

5

Data Mining Lecture 1 25

7!!

4,$-,,8)(#)+D ,,).,,)(+-, .)(3,C+.)

'''*'O'8'O'8,

2

+(& O'(+D

• ,,+-(

),-(->*-&(A ((+D-((,(-6

From [Fayyad, et.al.] Advances in Knowledge Discovery and Data Mining, 1996 Data Mining Lecture 1 26

!86

Early Intermediate Late

• !"

#$ !$ %!% &''()

Data Mining Lecture 1 27

• 9

9%!0927 ( (-(,(--! 7 0(,8 (--.-+)%%!

Data Mining Lecture 1 28

79

7 C+-(.! 7 ,-! :7 .(! $(*(! 7 C+--,! ;57 3, (,!

Data Mining Lecture 1 29

• 5,-

,.#*,--,(,

• -,
• %, -)#>'L(((6
• %-- (

E-(,(-,) .+,-. !

• ((-

H,(,!

• ,
• %((
• 3.,- #*,-

.,H(.--!

• 0(-.-#*,-

9

Data Mining Lecture 1 30

9567"!

• 7

,,---,

• 7
• F.-()0F5
• F.,
• :7

H-

• 7
• ()-
• ;57
• ()--,)(A,)-A!
• <7

3,H

6

Data Mining Lecture 1 31

9

C.( C, P,H 5% %,) ,-% % ,.% /)% % 2,

Data Mining Lecture 1 32

:!!

Data Warehouse

Data cleaning & data integration Filtering

Databases

• Data Mining Lecture 1

33

7$9* F,,-+ %* $,-+ 2-.-%@-(

C+D&--+D&,,-+ ,-,- $&--- $8-+-,--+

• ,)-+

777

Data Mining Lecture 1 34

:

Data Mining Lecture 1 35

:

• (-,

– .-, – 3 – 3-

• ,.,-
• .+,!

,-

,!

0.-, 3

=-

+*- ,-!

H 4,H

"! .

,-!

2((),) 2, A,,)

• A,

!

Data Mining Lecture 1 36

567:!

=8,## 3-(., %,. ,()+.

7

Data Mining Lecture 1 37

4012

• H-
• 4,HH--

!!-).!*

2 ,-,)

,&-,.!,&-, Q;'!!G<:,)Q;'!!G< +)Q;3,$P< R9L(-9>'LS +)$;%P%< +)$;%P%.< R9L(-9TLS

Data Mining Lecture 1 38

4032

,(-3-

E--,(-+-- ,((- =!!,()+-,,() +-, 3-&,(,, *# 3-3-#*,.,

,,)

,,+,#*4-(*,!! ,(--+ ,+-,8H&, ,)-H,,)

Data Mining Lecture 1 39

40>2

C,,)

C,-+D-,)*, +.(- +--8+A(, (--.,)

$--.,,)

$---.,) A,-),) ,)&+-,)

C&--,,)

Data Mining Lecture 1 40

')*

• 2-) A))-(

,,(!

• &-A)&+-(-
• 2(

,)-- +).,-*- *

• (),,)(,.,.,-

) #(

• $?-?7

C+D. +--(!! (-! +D. +-?+,(-!!8- .,)+,)!

Data Mining Lecture 1 41

4$!*

E-,,,

• )(-,, B

2.!,(.!,

(,)CH

• )(-,) B

2

E,,-(, ! 4,)UA) H

Data Mining Lecture 1 42

• 2*,,-;,<

F,,- ,*# P,H %+ **,,8- ,)(-+.!

8

Data Mining Lecture 1 43

• ,+,)

F,7,-% 0- =(0

Data Mining Lecture 1 44

7

• !

"

• #

$%

• Bayes Theorem
• Regression Analysis
• EM Algorithm
• K-Means Clustering
• Time Series Analysis
• Relational Data Model
• SQL
• Association Rule Algorithms
• Data Warehousing
• Scalability Techniques
• Neural Networks
• Decision Tree Algorithms
• Similarity Measures
• Hierarchical Clustering
• IR Systems
• Imprecise Queries
• Textual Data
• Web Search Engines
• Algorithm Design Techniques
• Algorithm Analysis
• Data Structures

Data Mining Lecture 1 45

7 4,(,)

%.- 3-.-

%((.*-((,(

• (-++-
• (#*,-+-.-
• (A,H-
• (,--

Data Mining Lecture 1 46

,+%

• F,,,+D&-+D&,,

.,&8,&- ,)777!

• 9%

H-,( ,--.-,,)! ,, -(-,,,.,

• : =

%+&--*C523 ,.,H,*#!

• F,,+#(-,)%/2

##,)7+7+&,,)!

Data Mining Lecture 1 47

• ,)-
• ((#-(#*,--+

.( #*,-,,,.,( + (+#-#*,- %A),--&- =8-.,H(-,

• ,-,-

3.,+,

3(-,+,)

=(()-,+,)(-, 3,,,-+--,-

?012

Data Mining Lecture 1 48

,-.)(-)

• ,,,-,8)(-

((-+-,+, ()777

,-,-,

2,(-.-#*,-

%&(-, ,,A)* 3,--#

(-.-#*,-*8 #*,-2#*,-(+, 3(-))-.)

?032

9

Data Mining Lecture 1 49

$ ,,(%

3.) 3(, 0H-

%

0(, F.FC 2) ,8) 7*,,,)- $

Data Mining Lecture 1 50

!

• %((,)-.(

,(-

• 2,.,(-+,)--*

*-(,

• 2%%,--,--

,(-.,- #*,-

• +(-.)((
• %(,H-

,(,,--,) !

• ,((-)
• D-