Tải bản đầy đủ (.doc) (59 trang)

Tiểu luận khai phá dữ liệu GOM CỤM DỮ LIỆU VÀ CÔNG CỤ WEKA

Bạn đang xem bản rút gọn của tài liệu. Xem và tải ngay bản đầy đủ của tài liệu tại đây (862.41 KB, 59 trang )

 
LỜI CẢM ƠN
  !"
#$%&'()*+, !(/0,'(
).%&1)&23$44
56478+9:10/%;#-<
%;9:3:"%&;$(3=&44$44
>47,/?#&.09/&;@&/;,AB4
,%'30"C0+7&D$4
>47E
 ! "#$%& "'' ()$*'$
"+,&- 
.)"/
FGFH%I;4J9FKKLKLML
#/
 
MỤC LỤC
01234
505
678'
94':;8<<2=>?02@8AB=CDE
''F "G )#/E
'$F "G HE
'E;&IJ %
'%K "L "MN
'NO&PQ+R "&S "L "T
'N'>- T
'N''U VT
'N'$G#W
'N$S&XW
'N$'U L)W


'N$$0Y&Z&[\
'NEVU#G#Z&]^ \
'NE'H _`a\
'NE$.U#b
'NEEcd &  b
'NE%eU#b
'T&Q "&S "& b
'WfX "G#Z&]^ '*
'\Z&Y +R "'''
94$=94=g0h'$
$'U V'$
$'' )HU V'$
$'$i+V ]j"XG#Z&i& U V'E
$'EL]U V'N
$'%O _&I "U V'N
$'N&ZY 'T
$$+R "&F "G&_U#- "U#G#Z&]^ 'T
$$'k&IU#G#Z&]^ 'b
$$$lU#- "Y&&mU#G#Z&]^ 'b
$$En &I#$*
$$%+[ "&S "& o2 pI)&  q$*
$EY&& %N$'
FGFH%I;4J9FKKLKLML
#/
 
$E'V&$'
$E$O]P`L "]j_]^ ]j)&Q& k&$'
$EEO&PQk ]HV&O& $$
$EE'O& c &L$$
$%E$O&  H"&I^$%

$%EEO& &Z"&I^$N
$E%Y&& %N$T
$%+R "U Vrsf#sP EE
$%'V&EE
$%$^ af#sPEE
$%EU Vrsf#sPofqE%
$N+R "] "&Y&& EN
$TG#]iZ ET
$T'PQ&+R "G ET
$T$SJ G#&#Z & ]iZ EW
$TE &+[ "]O "&#Z Eb
$WZ&Y +R "$%*
94Et5>?02@8AB8ut=vn<%'
EES "Lvs%E
EE'V&S "Lvs%E
EE$c iX vs%%
EEE`PF vs%%
E$% J n_IsIIsIsPP%N
E$N J n_IsIPPp#%T
EETk&I&Y)w%W
EET'k&I&YA%W
EET$k&I&Y<Ipp%W
EEW- ")Vx ( "n_IsI%\
y2@8ut=<2=>?02@8%b
%']z&&Q "%b
%''Y&& L "%b
%'$O&PQ" +R "&IJ %b
  M+R "&IJ N*
x ( "M" {N*
| "+,} "/ x "L "]j&` ")&&~+R "

&IJ N*
 &` ")&Y&& )s PN*
x ( "M+R "&IJ {N*
•ts €s{•&Y)wN'
•s€s{lk&psZ&GXN'
FGFH%I;4J9FKKLKLML
#/
 
•l!C{Y&.]Ol!C&s_&i_N'
•<{c)]j)]&^N'
•Is&s‚ )) &P{.iX "]j) "w c ]&^ N'
•‚ {.#&Y&& )s ]j")L)N'
•sI{l) J ƒ.]&^N'
„08uAB9h=‚2…N$
'Z&Y N$
B202@8<3tNE
FGFH%I;4J9FKKLKLML
#/B
 
FGFH%I;4J9FKKLKLML
#/B
@/0,NO3B44(&NO3 GFPH !
MỞ ĐẦU
• L do chn đề tài
I "  " ()"† ]U#!]‡&I•& )O&&I "  "
+V " "c xV  k&Mˆ -/)#& S " "&I&x
]‡] "x "L "& S " Hˆ - "U  "!&+R "
).!& !&^&I+, "x " !#/!&c ( !)S&I+, "!"L!‰
&S "P /Q+[ "&S "& ]‡]+[_`a&I "&k&Xˆ -]‡
] "&( "c ] "j! "]+[+&I&I "RP•&Y&I "#

U & Ak ]H)&Z ]jŠ]+[&S "& &mRP•
F "] "Ik&]+[G &U)&&Ij !S "]Š]+[ "+,&]z&&c
"/o>&  "qZ&GXM]‡]s).
H& S "&I "&-&ZH]Š]‡)&( " Y &xM  "+,&I "
U &RP•V ]jŠ&jiZ   "‹&L]O "Œ& 
Š") H+V "&ZY  •&Y& 
]+[L "&I "ˆ - #† V ]+[Z&m&mˆ -RP•!
)#/o) ssI  "q!&I& U &.oI&p &s"s sq!a&#Z&&S "&
o pI)& &sI#q!_Pk&&Q "coIii&#ŽP&&P&Pq!& &  ( "
o"sIpI) s)& "q!+R "& & )H)oPp&)& "
)s&"sPqi& M#Z&I "( iX o&s_&
)  "q!•sio•si)  "q!Y&Z&[oPP& IsP)  "q!
a&#Z&&Y&SoI"Ps&&sI#q!U L)oP&sI "q!U VoPPp& q•
I "  " ()&I+V]U#!+R "U V]‡]+[]H_k&! + "
S "Š+R "&ZY U . R  _R ‘   "
+R "# c V)+R "Š)O&[&Zik&[Ic "
P`L "U .f#sP "X]^ I’ "&O& ]OYI,I.AJ
Y# ŠIk&jj‰P`L " + "Z&GX&J+]+[&X] "
lU#- "U#G#Z&]^ i’ "&Y&& %N-&Ic "+[ "&S "&
o2 pI)&  q"X)&jPQ† † PP “&+• "RiX M&Y&& 
&O& Š"&I^"+[ "&S "& V  k&Pƒ]+[/ ]jU    +
)O&"X‹sIP&Œ&I "/ -&O& U V# c !)O&k
]HM&Y&& &Ic )O&U# Š&jz. H† &I "U#G#Z&]^ 
fc . ]Š!)O&&O& Š&j]+[} " H† &Ic )O&]+, "]L&jM
FGFH%I;4J9FKKLKLML 
#/K
@/0,NO3B44(&NO3 GFPH !
U#H]Š)"X)GXG&IJ U k>]Š-/ &O& ]jU
 k ]HIk&G &I/ "]+[ H / "c x!ŠIk& H
S "&IJ ]+[S "iQ&I "  " ()"† ]U#f&.”y>5

94= g0h! t  5‚t<2 =>?02@8AB
•5<2=>?02@8vn<– ’) "c x+R "U
V")L)L "&I "S "L "MS "L•s&I "

• Mục đích nghiên cứu
"c x)O&PQ+R "&Y&& L "&I "i& U
V")L)  #
• Đối tượng và phạm vi nghiên cứu
"c x+R "_U#- "U#G#Z&]^ -&Ic &Y&& %N!
+R "U Vrsf#sP+R "")L))s P
• Phương pháp nghiên cứu
"c xa&#Z&!U &!&F "[!]z&!G&I&I  "k ]H
† &Z&]H&
FGFH%I;4J9FKKLKLML 
#/Q
@/0,NO3B44(&NO3 GFPH !
CHƯƠNG 1
TỔNG QUAN KHAI PHÁ DỮ LIỆU VÀ MÁY HỌC
1.1. Tổng quan máy hc
#/o s0sI  "q)O& " / "c x&Y&&
|)#& Š&j/]+[ )o s&q!P-&-]O "MG#
&IJ /_U#- "  "Y&-G P&&I. "&&Ic RP•
P-#j M "U#ˆ -IO "V S "ei")/&m
Lz)w!)— /&( "+, "!/ŠP-+V "w !•#/j)
&I  "L&I+V]Šj)&IS X  "Z&GXM "]+I
/) ]j&P`L ".  "Z&GX #&. c   "P-&F "G&
  "&I+, "[)V
Š "!)#/S "P`L "  "G P&]R ˜H&YM
Š)& iO)O&&YY #&+, "]+["/&Yk #
z&Y/Y #x  ")w]+[ij‰ &s)O&

]Š]j)#Š&j]/j]+[Yk # i",™ ". !]Š
S "X& iOPƒ]+[/)O& _
O&&Z &IJ )#/")$"]. {
Giai đoạn hc{&Q "U &RP• Y IP-&+R "&-"
]Q&+[ "Z&GXMU &Š&j Š)]Q&+[ "&I "
V!&.IY&!&c ] V]Q&+[ ")V
Giai đoạn thử nghiệm{Y&]+[&.IX]+[j) ").i’ "
)O&PQ)& & &-&&Ic )O&† M&Yk # z&Ic )O&
&YV 
1.2. Tổng quan về khai phá dữ liệu
Š " "Q&m+R "Ic "i&!)O&&I "  "
. "G &I/ " k&&Q "c)#/Q "cŠ " "Q&m& /
]Š k ). ]Z ]O _& /!) ")Q &Z&Y)Š&j Y
I&Ic  H & /&I+Vj)&` Š&I "&-&Z"+[.!)#/Š "
"QIk& j&I "&-&‰ & & H #w ]Z &-&‰ !P‘ P "j)&`
]jiZ& Š&- &Q&&Z )S "† ,)O&x ")  &x
FGFH%I;4J9FKKLKLML 
#/R
@/0,NO3B44(&NO3 GFPH !
G&IJ & )SJ !&F "Z& 
"&I^]+[k#&m&Y&I+VP-&()—U
&+[ "V ]j)&mI)w[!ŠŠ&jj
]+[[)w]X)iX& &F "G&+]+[iZ&&I+V]Š!Š
Š&j-)w]Š]+I ]O "}[!j]+[Š&jic ^
j&k])w
• ( "U &M  "+,S "]†#]M&+VH
M!&Q]O&( "&I+• "MIk&V c)]Š  "]x "
). )ƒM•&Y&HX ( "&&Y!+&I! ( "-& & !†
)H)!P-& &.H#c )S "I— Š)S&I+, ". &I H^
L!xS "e. &I H"o]QV"U  "!S "# .!

. !•qk&X  "]H]Š  "#c  U &]d#&
&Ij 
1.3. Quá trình khai phá dữ liệu
I+V&c !U i&"&Y& "‹)SJ o)sqŒ‹)wo&&sI qŒ
} "&I "SJ )O&k&I‹G#)SV Œ!Š&j&F "Z&
G H&I+, "[oPsq!&I ")w)O&k&ILiO!&~)‡
i•)O&PQ&&I+, "[z&I ")O&)H  ~MS "" ;&IJ 
&I&x&Z  &si+VP{
šf+V'{Y[G P&&m "  "& &Y
f+V #Z&PxG &I/ "J Z_]^ Pk ]H&J& iO
G&IJ PX ! Š&I• c S!]H #™ ""Q " + " S "
"† )H)&k&i.•"]. _]^ #c† "+,} ">Y#
"].  #Šƒ"]. &Q  H&,"  k&
šf+V${/ /_]^ &Y]&I+, "c G 
šf+VE{0)P._i~ ‰!&H _`a>U &&Y[
]jP`L "&Q&R x "L "
šf+V%{X)iV&#j ]F&J)I  "]z&I+ "L "!
"X)iV&HziZ !ij‰ .].+[ "ik&iZ 
šf+VN{0-/ x ( "&F "Z&!U V!G#!Z&[
U  Š)
FGFH%I;4J9FKKLKLML 
#/S
@/0,NO3B44(&NO3 GFPH !
šf+VT{&I[IG#Z&]^ Y&]+[/ /]+[ Š).]jP`L "
&Y & R ‰jR 
1.4. Ứng dụng của khai phá dữ liệu
)O&i+V&I "G&IJ )&I&x")Š
&Y&& &#c } "+V)O&PQG]^ HGX& &
k Y ]+[]j&J)I)wz)SJ &I "Š)O&!
)L] &J)I)w)SJ ] "& &.&I "

RP• + "w — i^sk&i• " "# #
Š Hx "L "&I "&-&Z!)O&PQx "L "]j J M
Š&j]+[L "]j]+IG#Z&]^ &I "   +{
Marketing{x "L "i")U & M "-
Š]R ) "!_]^ Z.Is& "i")GX "!^
&I "!&+ "›!U ].  "!` "!#PX d)&Z&Z
 )L!iQ&I` "!Z ^GX "
Tài chính và thị trường chứng khoán{=L "U &&˜&
L "&cijM "!U ]. &X  Y ]+[!U &]†&+
&  +x " !"k#x " Y !G•&J &+R "!] "&
 !& ˜" !>-i"M.FZ&I "&^&I+, "
x " !
Bảo hiểm{=L "U &)x]OIMI_X#I]QV&m ".
 "!^L#Z +[&J)Z) ")iXj)!
Quá trình sản xuất{x "L ""XG#Z&P-&Q+M " &
"#c  +)#)Š! U P-! "#c Y&œ&Z&Z&Q+&I "G&IJ 
PX _k&!iQ&IU _+• "&Z&ZPX d)!‘ ".  +G&IJ &-]O "-
#c† "
Y hc và chăm sóc sức khỏe{x "L "i")&J)I)w&I "
)lš; "!U && ")]j&J)Ii &Y&!U &)Qc 
"&Ix "i !d ] U &X +• "M&Q!GXM
]H&I^œ&Q+MG&IJ _`a&I ")O&i  !)QG "
&J &I. "i  U V&J &I. "#c )S MiPˆ!
"IŠ Hx "L "Š&j]+[L "&I "
ˆ - +/!)S&I+R "!_‡O/!•
FGFH%I;4J9FKKLKLML 
#/T
@/0,NO3B44(&NO3 GFPH !
1.5. Một số phương pháp khai phá dữ liệu thông dụng
J ''ij‰ )O&&Y"XH)O&]j)&Ic J ].

 )O& "+,#&H  "U  "&.)O&PQ&,]j)&I "Gx>
]+[U .V{  " "+,S "ŠX ( "&IX [  " "+,
&J &I. "# [] "•&I. "&&Q&o&x&.&,]j)]ŠŠX ( "&IX [ "U
 "q  L  ]     M           -  ]  oIs& q    )S &X
osPI& q
1.5.1. Dự Đoán
>-] -"&I^)O&iZ z&I+, "&I "&Y]j-]
"&I^&I "&+R "z+iZ&MiZ † G &U)
KTKK.0
U V)O&&I " O "G &I/ "M&]U#
)O&ˆ - "c xŠ H&Ij / "V HX ( "x "L "&-&Z
L]MU VU &]†ij‰ )O&)SJ 
 _)V ,P`L "]z]j) Š&I "SJ  #
&+, "]+[P`L "]jU V&Y&I "&+R ")&Y #+
]+[_]^ &OV 
# Š!U V/)O&) _.)O&)L)O&
&I "PQV&I+VJ '$&k#P-U VM# [
&I "V"U  "Š&jP`L ")H U .]j&-]O "G#Z&]^ 
 " "+,# [&I "&+R "Š c ##S "
FGFH%I;4J9FKKLKLML 
#/U
@/0,NO3B44(&NO3 GFPH !
KTKQF'*
U &G#)O&+R "U &&Q "c]j-]Š "&I^
M)O&z)O&PQiZ L&OoiZ ]x "q&s)O&&Y[iZ ]OY
oiZ } "]j-iqŠ™ "]+[Y L "]j] "GX&]O "M
iZ ]OY]QViZ L&O
AL +] "_Pk&)O&i  U PƒZ&-&Ic &YZ&GX_|&
")d ] !-i †M "+,&c} "]QV)O&PX d )V-
&Ic .&]O "GX "&c} "

1.5.2. Mô tả
S&X&Y&I "&J)I)SJ )  "+,Š&jj
]+[ic ^]+[
KTQKV
L&c M+R "U L) Š)]Q&+[ "
&+R "&- &I "&YL)P]Q&+[ "&O} ")O&
L)&+R "] "— ]Q&+[ "&OL) PƒS "&+R "] "
U L))O&LM+R "/S "")P&S ""Q " +
U V!U L)S "]—~X]^  "ˆ&I+V)w
k # AJ&Z!Š&jU L))O&/i’ "G P&!
&I "U V/i’ "LI "+R " #i. PƒS "&j
iZ&Z&GXL)&]+[Pƒ +&Z i•&]†G&IJ AJY#!&S "
&+, "† Š)O&#c "Hˆ -]Š]j] "L)&]+[U L)
]+[P`L " H&I "x "L "HU .&^&I+, "!U .
 "! Y . ")w!U .&I "vsi!•"IU L)—
Š&j]+[P`L " +)O&i+V&H _`&Y&& 
FGFH%I;4J9FKKLKLML 
#/W
@/0,NO3B44(&NO3 GFPH !
J 'E&k#P-U L)&Y#&I "EL){+aI’ "
L) "c  |]j)&OH HR )O&L)
KTQQXD(Y0
L&cMU &Y&Z&[ #& ]+I)Qc 
""&I^&I "RP•w]†IM"X&Y&
&YY&Z&[&J)]+[Y&Z&[]+[&- G$i+V{
f+V'{J)&k&X&Y)LFiZ !)O&&Y)LFiZ ]+[_]^ G
& ]O&I[&~)‡ ]O&I[-&j
f+V${ IY&Z&[). &m&Y)LFiZ !Y&X&~)‡ ]O
&I[-&j]O& Y#-&j
+R " #]+[P`L "Ik&GX&I "ˆ - +Is& "!

U &G#Z&]^ !GX   !•
1.5.3 Khai phá dữ liệu với cây quyết định
U#G#Z&]^ )O&&I "  "J &x)S&X&I-G  k&!‰
j k&]QV "+,} "U#G#Z&]^ )O&k&IU#!ij‰ )O&k
]HG#Z&]^  &&I ""• V)O&&O& ]H !)  &m &&I "
"• V)O&"&I^M&O& ]H &+R "x "!) &"• V)O&"&I^
&O& G#Z&]^ 
KTRK$Z+NO3
O&&Y&Si"))wi ]†#+iZ ]F&m&F "
&j†Z&&Sij‰ )O& _O&Z&[M
)w&Q "cP-]He M#c "
I "&Z"V&-! Š "&S•• Š)x]O ‰H #
Š "#c  U   +!Š].+[ "S " 
_!>]Š! "&&+, "&H _`ao "ˆ!‹)P.Œq]j.i~&k&X
FGFH%I;4J9FKKLKLML 
#/M
@/0,NO3B44(&NO3 GFPH !
&Si^ ‰"]. &H _`a #™ "Š&jiZ ]F&Sj &^
R 
KTRQ%&
U#G#Z&]^ ]+[&.& i’ "† +[&o]G#q)O&&Y
& &Y !)&Y ]+[&.& M#Z&m† &`M} "
)O&V &oS "X &q]j)U   MU#AU
 &. &Š&j-&Ic j)&I)O&# H&O& ]j_]^ 
U 
KTRR[,,
A-/ M#Z&I "&Y&& U V-U#G#Z&]^ 
/ &O&  ]jj)&I&.) &MU# "&) ")Q / &O
& PU V&Y)w&Q& k&+Y# "&† XŠ)O&&c
d ]j] "k ]H #&cd Pƒ]+[&IJ i#&I "&Y&&

_U#- "U#G#Z&]^ •+R "$
KTRS\/
I ""]. &.U# "&Š&j"V. &&Ij MU#i’ "PQ
iX & &Q&j&.) &!]OPU&Q]MU##"&I^&Q&jM+[ "&S "
& &&c)"]. &.U# "&Š&j} "+R "‹O)S&X
"•  k&Œo ))>sPI& 0s "&q#"&I^&Q&jM"+[ "&S "&
o2 pI)&  q]j&eU#
1.6. Hệ thống thông tin
&Q "&S "& S "Lij‰ &I&x+V. ")O&iX "
")O&x "V&O&   "x "V ]Q&+[ "ž)w
Định nghĩa 1.1. &Q "&S "& )O&iO&x
qpA!<!o8!
=
&I "]Š
8&YI "!. ]Q&+[ "œ<&YI "!. &O
& œ


=
<

AA
/
VA

&Y"&I^M&O& 
œ ]

p)&S "& !V)/
]/



^


)p"&I^
/
G/_

q!o

Bảng 1.1:&Q "&S "& H &+[ "X))
FGFH%I;4J9FKKLKLML 
#/`
@/0,NO3B44(&NO3 GFPH !
8 U  & R ]† X))
_
'
 Š Š Š
_
$
 Š S " Š
_
E
fJ &+, " Š Š S "
_
%
‚k& S " S " S "
_
N

‚k& Š Š Š
_
T
fJ &+, " S " Š S "
_
W
‚k& Š Š Š
_
\
 Š S " S "
l|&fX "''!&Q "&S "& ")\]Q&+[ "
Ÿ_!!_ 8
\'
…=
%&O
& <¡ U  &!UR!U]†!X))Ÿ
1.7. Bảng quyết định
Định nghĩa 1.2.fX "G#Z&]^ )O&. "]zi&M&Q "&S "& !
&I "]Š&Y&O& <i")&Y I, {&Y&O& ]H
 &Y&O& G#Z&]^ >+ Y#!iX "G#Z&]^  )O& 
&Q "&S "& >¡o8!∪>!A!pq!&I "]Š∩>¡∅
S ")k&& &F "G&Š&j"X&Z&>e"))O&&O& G#Z&]^ 
# k&!o&I+, "[Š H&O& &Ji’ ")O&|)‡S Š&jG#
H)O&&O& q+Y#!)]Q&+[ "_&I "iX "G#Z&]^ ]+[)S&Xi’ "
)O&|&Ro
'
o_q!
$
o_q!!


o_q!o_qq
l|&L!iX "G#Z&]^ >¡ 8!∪ Ÿ!A!pŸ&I "]Š{
•
Ÿ!!!!!!! 8
\WTN%E$'
=
S "" ]Q&+[ "
•¡ !!!Ÿ&Y&O& ]H o{&Šœ{Hœ{U
z "œ{>} "&Qq
•¡ Ÿ&O& G#Z&]^ o{Z&GXq
Bảng 1.2: fX "G#Z&]^ H &+[ "I) • "
 &Š H U  z " >} "&Q Z&GX
    

'
s †)&+V ¢ S " f^I)

$
s  AmX Š S "

E
‚U) k AmX Š S "

%
s k AmX S " f^I)

N
f. †)&+V z " S " f^I)

T

‚U)  z " S " S "

W
‚U) †)&+V z " S " S "
FGFH%I;4J9FKKLKLML 
#/KL
@/0,NO3B44(&NO3 GFPH !

\
s k ¢ Š S "
Định nghĩa 1.3. iX "G#Z&]^ >¡o8! ∪ Ÿ!A!pq"/&Y
a9Nb
N^a

ž

qo

=
)H 
9
‘ "]^ M
l|&L!iX "G#Z&]^ &I "iX "'$
 )H 
9
‘ "]^ {
Ÿl! l8ž
$'
=
V

Ÿ!! l
N%''
=

Ÿ!!!! l
\WTE$$
=

ŸŸ Ÿ! Ÿ! Ÿ! Ÿ! Ÿ! Ÿ! Ÿ! 8ž
\WTN%E$'
=
AY#)H 
9
‘ "]^ ¡
$'
>9>9 ∪

Ÿ!!!! Ÿ!! 
\WTE$N%'

∪=
Ÿ!!!!!!! 
\WTN%E$'
=
^
=
>-L&Ic !‰&k#
qo

Nb

&Y]Q&+[ "]+[U V] "
o +q&I "8 ZP`L "&Y&O& ]H 
O&  &O  &    ∈   ]+[  "/    S "  †  &Z&  &I "  >  Z
qoqo
Ÿ 
NbNb
9

=
"+[.† &Z&&I ">
 ŠiX "G#Z&]^ >¡o8!∪ Ÿ!A!pq]OY Z)/&O& 
∈]H† &Z&&I ">
Định nghĩa 1.4. l|&iX "G#Z&]^ >¡o8!∪ Ÿ!A!pq]Q&+[ "
_!#∈8 Š_#)U&w  &I "> Zo_q¡o#q + "o_q£o#q
Q&+[ "_]+["/ k&G &I "> ZS "& &.)O&]Q&+[ "
#)U&w V_>]+["/ k&G  Z)/]Q&+[ "&I "_∈8
]H k&G 
Mệnh đề 1.1. l|&iX "G#Z&]^ >¡o8!∪ Ÿ!A!pqŠt

oq¡
_∈8¤_]Q&+[ " k&G Ÿ
R  ! Z> k&G &Jt

oq¡8
1.8. Kết luận chương 1
•+R " #]‡&IJ i#&F "G H!x "L "M
!"V&)O&PQ+R "&S "L "
•IJ i#&F "G HU#G#Z&]^  +)S&Xk&IU#!&H _`a
!&.U#!&cd &  +R "&eU#
FGFH%I;4J9FKKLKLML 

#/KK
@/0,NO3B44(&NO3 GFPH !
•IJ i#&F "G Ha&#Z&&Q "&S "& iX "G#Z&]^ 
CHƯƠNG 2
CÁC PHƯƠNG PHÁP PHÂN LỚP
2.1. Phân lớp dữ liệu
2.1.1. Khái niệm về phân lớp
U V¥'¦•&Y&-&Ic &Y)w  ""&I^#
‡ MV&I ")O&&O& U VP`L " Š&I "U V
)V "&Š&jU &I+, "[U V&S "L "{
Lớp mô tả{& c MV S "&O.) "c x
U VAL!]Y  "&O & V&i& !
 " & ]+[_]^ &OV i•&O& ]zi&AJY#!
 "& ŠU V•]U#P-U V& c &I "} ")O&M ".
)" "S "Šik&§)O&]z]j)z&O& ]zi& ]jU
i&
Lớp quyết định{V•]U#Z&GXM-] -&S "& M
&O& ! "+,&&+, "_s) "&Oi& H-] O&L)
 "&‰&k#]Š‡Pk&o‘ ".  +‡Pk&)i  ".&!‡Pk&&H
"`z‡Pk&&H #!•qS "&+, "U~]+[]z&I‡Pk&Pƒ&( "
o&OV'q#S "&( "o&OV*q0V]+[]^  "ˆ&I+Vi•)O&†
S "" )w!&xi•&O&  "&Š&j&k#I’ "V)O&)
M&O& +Y#)O&&j Š&ji^U PV ZŠ)O&&O
& +[& ""V. ]‡]^ &I+V0Y&]+[} "]jU V-
FGFH%I;4J9FKKLKLML 
#/KQ
@/0,NO3B44(&NO3 GFPH !
&O& ! +Y#k ]H•]U#&.IY& ""Q "VY&&-&Z
 "&Q&
2.1.2. Các bước chính để giải quyết bài toán phân lớp

f+V'{/oI  "qL]Mi+V #_U#- ")O&)SJ _]^ 
)O&&YVSJ  #]+[_U#- "i’ "U &
iOM)O&RP•!)iO]+[_]^ i•"&I^M
&O& XP`)iO&OH)O&&I "V]‡]+[
]^  "ˆ&I+V!]H #]+[_]^ i•)O&&I "&O& !"/
&O& U VI " "X Mi& U V!)iO
]+[_s))O&)w!)O&L!#)O&]Q&+[ " "iO]+[
U &]j_U#- ")SJ U Vk#&m&I "&Y/#
k # o&I  "&Ps&q "iOIc "˜&.& &Y
k # — "/  ")wk # o&I  "P)sPq]+[
/  "w c &m)w
SJ ]+[]+IP]‡U &_ "&Yk # &+, "
Š. "  "Y&U V!U#G#Z&]^ #S "&x& /AL!
J $š'Š)O&iX "H&S "&  "!)O&)SJ U VoY&
FGFH%I;4J9FKKLKLML 
#/KR
@/0,NO3B44(&NO3 GFPH !
U Vq]+[_U#- "G&IJ /•f+V'Š&j_]^   " "&
Y#  " "iJ &+, "M)O&` "0Y&U V #Š&j
]+[P`L "]jU .)w&I "&+R "!™ " + Š "k)O&
&I&xx&I "RP•
f+V${>j)&I]+[} "]j+V+[ "]O _MiOU VZ
]O _k Y ]+[&JŠ&j} "iOU V]jU V
)w)V
†&c !] "]O _M)SJ #iOU V #i’ "P`
L ")O&&Y)w]‡]+[U V]jj)&Io&sP&q"/iOj)&Io&sP&Ps&q
 ")w #]+[/  "w c ]OYV)w]‡/•f+V'"/
)wj)&`o&sP&P)sq
O _M)O&)SJ U V-&Ic iOj)&`&¨  "
)w&`]+[U V] "i’ ")SJ U V]Š0+a! Z]O _M

)SJ  #-&Ic &Yk # !&J)SJ  #]+[] "&Q+!
ŠU V] " & &Ic )w]‡/!&I "&I+, "[ #!)SJ 
+V "&VP-G&otsIp&& "qMAJY#XP`L ")O&iO
&`Z]O _M)O&)SJ ]+[_s)_|&Š&jk Y ]+[&J)S
J ]Š]+[} "]jU ViOz]Q&+[ "&I "&+R "
AL!)SJ U V]+[_U#- "&I "f+V'i’ "U &M
FGFH%I;4J9FKKLKLML 
#/KS
@/0,NO3B44(&NO3 GFPH !
 "]‡iZ&!]+[} "]j-]  ")V&I "&+R "•
J $š$
2.1.3. Mục đích phân lớp dữ liệu
 R {‘ ". &I "]/)‡i+] i’ ")#]jŠ&jP•_Z
U .&+&m!&I Š(  U c i+ 
l]^ ]OIMI&I "]R _ #&H -&S "&  "
#&H ! "&Š&j_Z "_ #&H  Š) "Š]OIM
I!&kziJ &+, "]j&m]ŠŠ&j&)X]j]+IG#Z&]^ ] "
]• 
I "ˆ -#! "&#/ "&I ]+[w&Y&d
] P†)!Y#)O&&Q "d ] ]OY!] "& Y#-&I
x "ic  "Ik&† &Z&
>-] &I+V]j]+IG#Z&]^ ]†&+& L "#)i x "
 
k ]H† G &U)"XG#Z&]QViOU V{
o Độ chính xác{]O& Y#M)O&Y&-&¨ _U V
)z}Š)O&PQ]+[ "c)&I/ "R ! + "k ]H
G &I/ "]He ] "]QVVG &I/ "
o Dễ hiểu{O&iOU V‰jPƒ&. "+,P`L "& &+• "R 
&Q "!] "&,™ "" "+,P`L "&I ]+[j†)Z&
GXM)O&Y&]+[]+Ii•&Q "

o Thời gian để hc{Ak ]H #]zi&G &I/ "&Q "]+[P`L "
&I ")S&I+, "&#]F&+, "_#c !]H]Š#c†&Q "X
/Ik&  )O&Y&U Vz  Š "]He )O&Y&]‡]+[
/}[V&-&Z," /X  ™ "#]+[P`L "
&I "&Q "ŠPQ+[ ")w/&+R "]Q&
2.1.4. Độ chính xác trong phân lớp dữ liệu
O _M&Y&& ]QV&Y/&+, "Ik&V]O
 _M&Y&& ]QV&Yj)&IH #S "XS "
&+, "_#c _X#I!]zi&&I "x "L ")#/!]QV&Y/
&JŠZ&GXIk&&Q& + "]QV&Yj)&I&JIk&&S "&+, "&J]O
FGFH%I;4J9FKKLKLML 
#/KT
@/0,NO3B44(&NO3 GFPH !
 _]QV+iZ&&x+_]^ ]+[V)VG &I/ "
+R " "]j-] ]+[ H "+,k Y  k&P`L "&Y
]‡iZ&&I+V&s +P{
>} ")O&† &I+Vo&Y/q]jk # &Q "
_U#- ")SJ o&YY&q
YY& #Pƒ]+[j)&I&Ic &Y— .!Z&GX]+[]s)P
P VZ&GXU V]‡iZ&&I+V¨U V] "&Ic &Yj)&I
)O&+V] ]O _M&YY&]+[_U#- "&I "G&IJ /
2.1.5. Các cách tiếp cận
U V— ]+["/U VŠ")P&oPsIPsPPp& q!
)O&&I "  "ˆ -FiZ  k&M)#/o) ssI  "q
o&)  "qŠ H]jij‰ )O&)SJ U VŠIk&
H&Y&& "XG#Z& Š&Y&& U V&ciji"){U#G#Z&
]^ !2>E!%N!). "sI!P#Y G# .!rsf#sP!AoI&As&I
 sq!•k&X&ZY  #_U#- "  ")SJ ]HŠX ( "
U V)O&)w)V+iZ&-  ")w&+R "&-]‡]+[/
I "&-&ZL "! "&&k#I’ "S "Š)O&&Y&& U V

&Q& k&&I ")/&I+, "[))&Y&& e&Q&&I ")O&PQ&I+, "[
L&j ]Š)&S
>&," Š. !&I ".)i&. #e "c x]Z )O&
PQ+R "&ZY U VŠ+V "w )L&jU#G#Z&]^ !2>E!
%N!+R "rsf#sPG#]iZ 
2.2. Phương pháp tổng quát xây dựng cây quyết định
f+V'{ Z&. & &,!&k&X]Q&+[ "k # ]H&O)O&
V ]Š&J & #&  &Š&c  ‡ V "M]Q
&+[ "
f+V${ I+, "[ "+[.!P`L ")O&]O]!/ &O& ]H U
&Q& k&&Y)wk # Š&. &
f+VE{ .)O&+[ " & M & &,i’ "PQ"&I^ M
&O& ]+[/  )  &m &]Z  & )O&"&I^
M&O& IU ]Q&+[ "k #  &
&+R "x "
FGFH%I;4J9FKKLKLML 
#/KU
@/0,NO3B44(&NO3 GFPH !
f+V%{ & &]+["/&†  k&!&I•&  &! Z&k&X]Q&+[ "
)w&.]Š]H&O} ")O&V0z.f+V'šE]QV)
&+&†  k&
l|&)O&L_U#- ")O&U#G#Z&]^  +P{
fX "$'ij‰ &S "& H'%]Q&+[ "
'
!•!
'%
fX "$')O&
iX "G#Z&]^ V&Y&O& ]H ¡ "&I,!&]O!Od)!ŠŸ
&O& G#Z&]^ ¡ RŸ
fX "$š'{fX "G#Z&]^  "R

8 "I, &O Od) Š R

'
• " Š "  ¢ S "

$
• " Š "  .  S "

E
HU# Š "  ¢ Š

%
+ ©)  ¢ Š

N
+ & f ¢ S "

T
+ & f .  Š

W
HU# & f ¢ S "

\
• " ©)  ¢ Š

b
• " & f ¢ Š

'*

+ ©) f .  Š

''
• " ©) f .  Š

'$
HU# ©)  .  Š

'E
HU# Š " f ¢ Š

'%
+ ©)  .  S "
J $šE) .)O&U#G#Z&]^ x "V&YfX "$š'oVV
‹ŠŒ "ˆRp!V‹S "Œ "ˆS "RqU#G#Z&]^  #
]+IG#Z&]^ Š#S "ŠRp&§&O&S "PQ]†
  o"  &I,!  &  ]O!  O  d)!  Šq  A  L  )O&  )w I&4#<cI
I3dcICd[c9/&B4Cce% ]+[U VS "o&x "
&I, ’ " Š "!]Od)"Š). &JPƒS "Š "+,Rpq
FGFH%I;4J9FKKLKLML 
#/KW
@/0,NO3B44(&NO3 GFPH !
U#G#Z&]^ ML&Ic Š&j]+["X& +P{ &x
"&I^M&O& G#Z&]^ #&O& U V o&O& ‹R
Œq & &+R "x "V&O & &O & ]H #
&O& U V! &"Q™ "]+[_s) +)O& & ]zi&!•]U# 
&O& ‹"&I,Œ  MU#&m)O& &ik&§&+R "x "V)O&"
&I^M&O& ]H ]+[/ 0+aU#G#Z&]^ &Ic S "ŠP-&)
"M&O& ‹&]OŒ&I "& † U#!&O&  +Y#]+["/
 "&O& +&mi•J&O&  #S "X +• "]Z G

&IJ _U#- ")SJ MU#
I "&Y&& RP•_U#- "U#G#Z&]^ ek Y &O
& &)"G&IJ U VŠ"&I^I,I.!i")X&O& ]+[
} "]j-] &I "G&IJ /™ " +&O& ]+[P`L "]jj)&I
&.) &MU#
A_U#- "U#G#Z&]^ ]+[&Z  )O&]G!† +[&&m &
"Q_Q "&V&Y  &.) &  ] "_|&! Zj)&I&k#&X
]H m "&J&Y&& Pƒ&. && #]+[" )O&"&I^M ‡ V
&}#]H m "]+[&‡"+[.!&Y&& &Z  / ]j)&Q& k&
&s)O&&c&I+V!U   &s]H  #
i+VU &Ic !&Y&& PƒzG&k&X&Y o]‡]+[q
&Z  "/]G +i+V]†&c V &Y  #
FGFH%I;4J9FKKLKLML 
#/KM
"I,
Od)
Š
Š
• " H)U#
+
S " Š
 f
S " Š
.  ¢
J $šE{ALU#G#Z&]^ x "VfX "$š'
@/0,NO3B44(&NO3 GFPH !
I "f+VE!&cd P`L "-/ &O& ]+[j)O&PQ]
]O}[!)O&PQ]] "]O&†  k&!#)O&G#&•U &Y)w
k # 
2.2.1. Cấu trúc cây quyết định

 !"!#
$%&#"!'()$*##!+#")
%,-&+!.! !/#"!''"!0%1
2,3445
š &) ")O&&O& oiZ ]OYq
š  &+R "x "V)O&"&I^M&O& 
š &)O&VoiZ L&Oq
2.2.2. Xây dựng luật từ cây quyết định
)O&]+, "w &m"Q]Z  &&I "U#&.& )O&Y&!Y& #Š
Z&I)O&iO"&I^M&O& ]+[/ ]jU V!ZX)O&&I "
"&I^M&O& Z&GXl|&U#G#Z&]^ J $šE&Š)O&PQY&P
]U#{
 o"&I,¡• "∧Od)¡fq→oR!Šq
 o"&I,¡H)U#q→oR!Šq
 o"&I,¡+∧Š¡¢q→oR!Šq
 o"&I,¡• "∧Od)¡q→oR!S "q
 o"&I,¡+∧Š¡. q→oR!S "q
FGFH%I;4J9FKKLKLML 
#/K`
@/0,NO3B44(&NO3 GFPH !
-6789!':;'#"'#!
&!7<'8"'#'0:+2,
'289''<6!&!"!=2,3
<">!%4 ?##<@,4@
?AB7/'$
':@"!7<'C'!/DE' 'F<G 'F/# ?
@"!'H"'#I$&,
2.2.3. Entropy
n &I#¥'¦)O&].+[ "& /} "]j]+[ "& S "•o#
+[ " "w c qM)O&P- #MU Q "w c &I+V]+[

]^  "ˆ +P{
iX "G#Z&]^ >¡o8! ∪>qS ")k&& &F "G&!&Š&j
_s)>¡ Ÿ!IoA

q¡U .M8]QV&O& ")V{8
'
!8
$
!
•!8

!]z&
!!'!
qo
qo
(
^9/#N
^9/#N



==
 

(


00P5#&0
$
'

"qo

=
−=
o&I "]Š

 &e
]Q&+[ "&I ">) " ‡ Vq
m]^  "ˆn &I#•&Ic ! "&&k#I’ "{
n &I#¡* Z&k&X)w]H&O} ")O&V
n &I#¡' Z¡$Io8
'
q¡Io8
$
q
*ªn &I#ª' Z&O&I+, "[
Ví dụ{&Y8&I "fX "$š'Š'%)w!&I "]ŠŠb)w&OV«Š¬N
)w&OV«S "¬]Š&Š{
n &I#o8q¡n &I#o¥bŠ!NS "¦q¡šobž'%q"
$
obž'%q-oNž'%q"
$
oNž'%q¡*b%*
2.2.4. Gia lượng thông tin (Information Gain)
+[ "&S "& ¥'¦&cd ] "X ( "M)O&&O& 
]+[} "]jU V)w-"&I^n &I# "&-&cd
#]j/ ]+[&O& Š"&I^n &I# ~ k&} "&O&  #]j
U )X &Y)w+[ "&S "& ]+[& &sS "&xP{
qo
¤¤

¤¤
qoq!o
B
GB
B
P5#&0
P
P
P5#&0P/



−=
&I "]ŠA

&Y"&I^M&O& !>

&Y]Q&+[ "&I ">Š
"&I^&O& i’ " o>!q]+[® ‚PP;  ¥W¦P`L ")]O]
FGFH%I;4J9FKKLKLML 
#/QL

×