!"#$%
&#'()*+,-.*/01%2)#$3456,)758 +9:;<
=>?>8@))#A
BCDEFG
HGIGF
IJGFKILG
M#NGKOFGIP
1
Q3Q3
2
1. R+S%5&)*+,-)T+#)$)*+)*+,-8 +/U3)T+#
!"!#$%&''!()*+, $%
,/!01(234)15*67)896:2
69;<=)*17;!%9+:$,
;"9;<=)*!(-3>;
&?
& !"/ 67( <#(17*
,7+@012>!)2$A!9(9*7!0)2*2;
)2=4"-BC7!02>92D1!!
E9)$!(F #(:!8*=4;$G
/!0.($%9!%+
H:>)**)2$%9$$<*I
3J92)1!K+1>& !"2A / *
-3!KL7!00)*$%9!%JM1
E<,N!OP2>-Q!4& !"/ 01')*
!%+:>)*)2L242;)2$%')*C
2'*)2)RD19M19L17*
2;%-1=46j9&;9N# ;%-1 S6
$ )?;- &9T!#+LD)N!89 !'
V>1=1_M1&b!8K!N$9>,!(!$6;1>:!'N
# 7$!8) >N!=1+%;VE4/ 9 S6)2!(
$$%&'AL$4$N# !( ^6/ :$,41
*&(W+HVN2>&X!8/ ^6 S)&
$2;- $.!8&K 6J9*.L$4$/ 2& 1!##
(!8L$7>6j9)S6+c;!#)N# 9 !(!'
9N# Z17; S6L$4$ >k$ / %-1
9 S61>:+
%&'8!($)N# I($ %-1-3A6
*($ !'9L+[.(!'V S6
1-f$4H 4H:#VN%!W 2L&
=)LVN;94l4;2;2$79 +
e9V )N# T"8N# =)j$;.$N#
R 92;V6'9 1&X#R *=)+
.$N)2$%"8$J%/ )N# +.$N
($&;9%-11>:*M1 $%.$N*Z$ZhNN
$%Z17;.$N!#+69)2>!34M1 ;E01
&;)N# 34'! "89&/ N# #R
9 1;6'9 1+
m 1!f> &XL$(1$%D &K )2'*;
N# !(Z9!W& $$ )+
4
3. )% W5#VX5YV.#5Z+[Q+8 +/U3.\5]]5#/
3.1 Thut ton lc t kha Naive Bayesian
d%)d >& ;3!%- !W)`d >&!(,;9Z9&17Z>
$%&- 2;V&Z> !#+ea1:E:>
!8&K !(f);3& $+H41$%&'=A6Z17"
Z1>;9& $"*Z17;9$ )*
1:;9&X($ 9N# ;:& $2:$ )89f
,99W)2$%1GpeqH+m 1!#Zf>-x & 6)E
D;9peqH;&')=Z17/ 91GpeqH)2
=&'/ N+[F ,;9Z9&17/ 1GpeqH!53;
& 6)6jP= . ^y)2=&'/ 1GpeqH7!W2)2!%-2
/ & 6)ED+o6(1WZ9&17/ $ )$:!8)2
& $%-1D 1GpeqH+[1'B :!8M1>4!W
96j9,;9;*D\8!#
x x
x
x x z z
_ v bu _ b
_ v b
_ v bu _ b _ v bu _ b
P x c P c
P c x
P x c P c P x c P c
=
+
5
;!#
x
)2)9
z
)2)8)
Z)2$%$ )67L
o_Zv
b{o_S
;=$2> 42$&K EL
=$0$!84*V[•+[ELm $d >& & $
)9- %-19 1:;9d >&+
:-V)117)1>5$$ )6L"2& $$ )+t9&17N
€+•••‚J)!4x#A)2& $ ;+t9&17Ex)26L"+
6
:N# & 1!17)1>+ ($ K$%$Q1*& $
Z9&17)Ex+
7
$%$Q1& $Z9&17&X)2+€••••+
8
9
4. #/#$%)5],); '
A Keyword Filters Method for Spam via Maximum Independent Sets.
Provided by Science & Engineering Research Support soCiety (SERSC),
May 2013.
Phát triển một số phương pháp lọc thông tin cho hệ tư vấn – Nguyễn Duy
Phương – Luận án tiến sĩ ĐH Công Nghệ 2011 – Mã số 62 48 01 01
/>10