TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol.12, No.7, July 201
4, pp
. 5585 ~ 55
9
0
DOI: 10.115
9
1
/telkomni
ka.
v
12i7.517
1
5585
Re
cei
v
ed
No
vem
ber 2
0
, 2013; Re
vi
sed
Jan
uar
y 7, 20
14; Accepted
February 5, 2
014
Polygraph Survey and Evaluation Based on Analytic
Hierarchy Process
Zhixia Jiang
*
1
,Yibo Liu
2
, Pinchao
Me
ng
3
1,3
Department of Appli
ed Mat
hematics, Sch
ool of
Scie
nce,
Chan
gch
un U
n
iversit
y
of Sci
ence a
nd
T
e
chnolog
y, C
han
gch
un, Jili
n
,
13002
2, Chi
n
a
2
School of Phi
l
o
sop
h
y
an
d So
ciet
y
,
Jil
i
n U
n
iv
ersit
y
, Ch
an
gc
hun, Jil
i
n, 13
00
12, Chi
n
a
*Corres
pon
di
n
g
author, e-ma
i
l
: zhixia
_ji
ang
@12
6
.com
A
b
st
r
a
ct
T
he ana
lytic hi
erarchy pr
oces
s (AHP) is a kind of
co
mbi
n
i
n
g qua
ntitative
ana
lysis an
d q
ualit
ativ
e
ana
lysis of co
mpr
e
h
ensiv
e e
v
alu
a
tion
met
h
od. In mo
der
n perso
nal
ity test, lying is a co
mmo
n
ph
en
o
m
e
non.
T
h
is pap
er comb
in
ed w
i
th the actual surve
y
results,
using
AHP to meas
ure the proc
es
s of empow
er
me
nt,
in fact is thro
ugh th
e hi
erar
chy mode
l is
set up as
i
n
d
e
x system, th
e subj
ective i
ndic
a
tors of tw
o
comparis
on a
n
d
eval
uatio
n, throu
gh the
j
u
d
g
ment matrix to calcu
l
ate the
w
e
ight coeffici
ent of every in
dex.
After throug
h t
he co
nsiste
ncy
check, corr
es
pon
din
g
to
hiri
ng a
nd
pers
o
n
a
lity ass
e
ss
me
nt to e
m
p
o
w
e
rme
n
t
in scho
o
l ke
ep
file, in ord
e
r to obtai
n better truth, accura
cy, so as to reali
z
e are esta
blis
h
ed for the bi
g fiv
e
perso
nal
ity test scores, to a certain extent, to avoi
d
lyin
g, bet
ter more
authe
ntic
ity of the test results.
Ke
y
w
ords
:
an
alytic hi
erarchy
process, poly
g
raph, pers
o
n
a
li
ty test improve
d
Copy
right
©
2014 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
Analytic hie
r
archy process (A
HP) i
s
a co
mbinatio
n of qualitati
v
e and q
u
a
n
titative
analysi
s
of d
e
ci
sion
-ma
k
in
g method
[1, 2]. It can be t
he elem
ent th
at related
to
deci
s
io
n-m
a
ki
ng
is d
e
compo
s
ed into
goal
s, prin
ciple
s
,
scheme,
su
ch a
s
level, o
n
the
basi
s
of the d
e
ci
si
on-
makin
g
m
e
th
od fo
r q
ualitat
ive analy
s
is.
This meth
o
d
has
the adva
n
tage
s
of system,
flexible and
con
c
i
s
e [3]. As a kin
d
of qu
antitative ana
lysis
an
d qual
itative analysi
s
ca
n be
com
b
ined
with the
comp
re
hen
si
ve evaluation
method, AHP has bee
n widely use
d
in the actu
al life [4-12].
Along with t
he pe
ople
u
nderstan
ding
of ment
al h
ealth an
d attach
es im
po
rtance to
ascen
d
, all sorts of p
s
ych
ologi
cal eval
uation al
so b
egan a
wid
e
rang
e of ap
plicatio
ns. Th
e
person
a
lity test a
s
a
foun
d
a
tion fo
r a
m
o
re
ba
sic
me
asu
r
e
of in
dividual
ch
ara
c
t
e
r, inte
re
st fo
rm
spe
c
ial
broug
ht to the atte
ntion of the
society.
Many
colle
ge
s an
d
universities in
our country
and
the enterp
r
ises and in
sti
t
utions to their st
u
dent
s or staff con
d
u
c
t reg
u
lar pe
rsona
lity
psycholo
g
ical
evaluation.
At the same
time, mo
re a
nd mo
re unit
s
be
gan to a
dopt a vari
ety of
person
a
lity assessme
nt a
s
a te
sting
point in
the
recruitme
n
t pro
c
e
ss, to
ensure th
at the
recruitme
n
t o
f
staff not on
ly ability outstanding
healt
h
y and
have
good
p
s
ych
o
logi
cal q
uali
t
y.
Cog
n
itive tests to mea
s
ure the individ
ual have a
certain d
egree
of cognitive
ability, so do
n't
con
s
id
er the factor of
spe
c
ulation, su
bj
ec
ts
only to k
n
ow the right answer to
s
c
o
re
[13].
And the personality test is different; the subj
e
c
ts cou
l
d score at wi
ll. Test questi
ons for
attention, for
example, "I u
s
ually h
a
rd to co
ncent
rate on th
e thin
gs at
han
d", if the su
bje
c
ts to
answe
r "n
o", su
gge
sts th
at he
ca
n
concentrate
on
th
in
gs
, but th
e
re
la
tive
ly h
a
s
a lo
w
con
c
e
n
tration
of peo
ple
ca
n also a
n
swe
r
"no", e
a
s
ily t
o
a
c
hieve
so
me expe
cted
results by
fra
ud.
So how to an
d arou
nd the recruitme
n
t process for th
i
s
kin
d
of situation in the personality test in
the scho
ol a
r
chive
s
lie
se
ems to
be ve
ry ne
ce
ssary. The tradition
al dete
c
ting
method
have
set
s
o
me difficult to give fals
e title, or as
k
some
que
stion
s
abo
ut ambi
guity to measure by. A better
way to
have
a false id
entification
scal
e [
13, 14]
.In order to elimi
n
a
t
e t
he perso
n
a
lity
asse
ssm
ent
part lie
s in t
he process o
f
applic
ation situation
in o
u
r
count
ry,
the intro
d
u
c
tion of freq
ue
ncy
adverb
s
in o
r
der to imp
r
ov
e stre
ngth lie
s, in the hope
of relief, such as in this
p
aper,
we stu
d
y
in
the re
sea
r
ch
of the big five pe
rso
nality inventor
y. But in re
ality, the st
rength
o
f
the frequ
en
cy
adverb
s
choi
ce al
so faces
lie.
Based o
n
a five level measure frequ
en
cy adv
erb
s
is completely d
oes not conf
orm to,
comp
ari
s
o
n
doe
s not co
nform to, in gene
ral, com
pare
d
with, more in line
with the big five
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 7, July 201
4: 5585 – 55
90
5586
person
a
lity test to
study
the pe
ople
to solve
th
e
pro
b
lem
of
the freq
uen
cy of lying in
the
person
a
lity test, and
with the aid
of AHP to measure
with five leve
ls of fre
quen
cy adverb
s
of t
e
st
score
s
, with t
he aid of mathematical too
l
s
to achi
eve better effect o
f
lie detector.
2. Big Fiv
e
Personality
Test
2.1. Big Fiv
e
Personality
Test Simplified Tes
t
Tabl
e
AHP ca
n be
use
d
in the
multi-index
comp
re
hen
si
ve evaluation
of empo
werment. In
multi-ind
e
x comprehe
nsiv
e evaluatio
n
of ea
ch i
ndex for
criterion layer and target layer,
its
influen
ce is n
o
t exactly the same si
ze a
nd functi
on, n
a
mely the ind
e
x for total evaluation, is not
as imp
o
rtant.
As this a
r
ticle
sele
cted in t
he mea
s
u
r
em
ent of the big five personali
t
y scale, ba
se
d
on all the adv
erb
s
of freq
u
ency is
equ
al
to sco
re
the
gradi
ent of 1
is divided i
n
to hierarchi
c
al
ly)
gradi
ent (i
n a
b
sol
u
te value
,
finally, acco
rding t
o
all
ki
nds
of othe
r
subj
ect b
e
lon
g
s to the
total
marks o
n
a
scal
e
of pe
rso
nality. This way of
evaluation in
subje
c
ts lie score differe
n
c
e
obviou
s
ly, so it is m
a
y be p
r
od
uce
d
in hi
ring
and
kee
p
fil
e
appli
c
atio
n
s
p
r
on
e to
poor
authenti
c
ity.
The big five personality test with a
simpli
fied versio
n o
f
the table is as follo
ws:
Table 1. Simplified Versio
n of the Big Five Perso
nalit
y Test
1. I am not a ma
n who is full of troubles.
31. I rar
e
l
y
felt sa
d or dep
ressed.
2. I reall
y
like most of the people I
met.
32. The
rh
y
t
hm o
f
m
y
life soon.
3. I don't like to
w
a
ste time to
da
ydre
am.
33. I often go t
o
t
r
y
ne
w and fo
reig
n food.
4. I
w
ill attempt to doubt and iron
y
othe
rs.
34. Most of the p
eople all know
m
e
like me.
5. On the
job, I a
m
efficient and capable.
35. When I
made
a promise, I can
usually
car
r
y
out
to the end.
6. I seldom feel fear and a
n
x
i
et
y
.
36. Most of the ti
me,
w
hen somet
h
ing is w
r
on
g, I
w
ill feel
frustrated a
nd
want to give up.
7. I like talking w
i
th others.
37. I am a ver
y
a
c
tive person.
8. The la
w
s
of na
ture and a
r
t fo
rm
s to make me
feel ver
y
m
y
st
eri
ous.
38. I like thinking
and pla
y
ing
w
i
th
theor
y or a
b
stra
ct concept.
9. I believe that if
y
ou allo
w
othe
rs to take
advantage of
y
o
u
,
a lot of people
w
ill do it.
39. I
w
ould rat
h
e
r
w
o
rk
w
i
th pe
ople, rathe
r
than co
mpete
w
i
th
people.
10. I'll keep m
y
it
ems neat and clean.
40. I have a clear
set of goals, and methodically
wo
rk to
w
a
rds it.
11. I often feel n
e
rvous and distra
cted.
41.
Sometimes I w
a
nt to hide fo
r ver
y
shame.
12. I like a lot of people aroun
d m
e
.
42. I like illustrates its, in the eve
n
t.
13. I have onl
y
a
little feeling even
poetr
y
.
43. I have no
inte
rest in thinking about the la
w
s
of t
he universe or
the human condit
i
on.
14. If
y
o
u
need, I
w
ill achieve w
h
a
t
I want to
manipulate other
s.
44. If I don
't like someone, and I
'
ll let him know
.
15. I'm not
a met
hodical person.
45. I tried to finish m
y
goal.
16. Othe
rs treat
me the
w
a
y often
makes me angr
y.
46. I often feel inf
e
rior to othe
rs.
17. When I re
ad
a poem or e
n
jo
y
a work of a
r
t, I
sometimes feel excited or surp
rise.
47. I'm not
an op
timist.
18. I al
w
a
y
s
p
r
ef
er to
w
o
rk alone.
48.
I'm idealistic t
h
ings full of curio
s
ity
.
19. Some people
think I am selfish and self-
centered.
49. I often a
r
gue
w
i
th m
y
famil
y
an
d colleagues.
20. It seems I alw
a
y
s
can'
t put thi
ngs in or
der
.
50.
I
w
ill be the pur
s
uit of ex
cellence in all things.
21. I seldom feel lonely
or sad.
51. I
often feel h
e
lpless and hope someone can solve my
problem.
22. I'd r
a
ther
do
m
y
self than to be
the leader of t
h
e
people.
52. I am happ
y,
23. I seldom pa
y
attention to
your
emotions or
feelings in different environment.
53. I believe that let the students listen to ver
y
cont
r
o
versial
speech w
ill onl
y
confuse and misl
ead their thou
ght
s.
24. Some people
think I am cold and calculati
ng.
54. I have a high
evaluation to m
y
self.
25. I
w
ill complete all assigned to it to the best of
my
w
o
r
k
.
55. I quite can according to t
heir o
w
n
pace, to get t
h
ings
completed on time.
26. Sometimes I feel completely
w
o
rthless.
56. When I am u
nder gr
eat pr
essure, sometimes I feel like a
mental breakdo
wn.
27. I often feel e
nergetic.
57. I'm eas
y to la
ugh.
28. When I found
the correct
w
a
y
t
o
do things, I
w
ill
insist on using this method.
58. I think that de
cisions on ethi
ca
l issues, w
e
should follow
the
religious authority.
29. I usually tr
y t
o
thoughtful and
considerate.
59. On th
e attitud
e
, I am a stubbor
n no compromise
.
30. Sometimes I can't do I should
be honest or
trust
w
orth
y.
60. I'm going to
t
a
ke a lot of time to settle down to
w
o
rk.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Polyg
r
ap
h Surve
y
an
d Eval
uation Based
on A
nalyti
c
Hi
era
r
chy Pr
ocess (Zhi
xia
Ji
ang)
5587
With
five adv
erb
s
of
fre
q
u
ency mea
s
u
r
ement of
the
big five
perso
nality, inclu
d
e
option
i
s
not in
confo
r
mity with the
com
p
letely, more
do
not
confo
r
m to, g
enerally, com
pare
d
with a
nd
fully c
o
mply
with five gradient. This
tes
t
, the opt
ion
d
oes
not confo
r
m to the
sco
r
e value
is
1, the
option value i
n
crea
se o
ne
1 score, adja
c
ent to t
he
op
tion com
p
lete
ly conform to
the score val
ue
of 5.
2.2. Fiv
e
Test Score Sta
n
dards
Each qu
estio
n
from co
mpl
e
tely does n
o
t
acco
rd
with to fully comply with resp
ect
i
vely 1,
2, 3, 4, 5 points.
Nerv
ous:
Item: 1 6 11 16 21 26 31 3
6
4
1
46 51 56 A
m
ong them: 1
6 21 31, reve
rse
score.
Extrov
ersion:
Item: 2 7 12 18 22 27 32
37 42 47 5
2
5
7
Among the
m
: 18 22 47, reverse
score
.
Openn
ess
:
Item: 3 8 13 17
23 28 33 38
43 48 53 5
8
Among them:
3 13 28 43 5
3
58, reverse
score.
Friendly
:
Item: 4 9 14 19 24 29 34 3
9
4
4
49 54 59 A
m
ong them: 4
9 14 19 24 4
4
49 54, reve
rse
score.
Cons
cientio
u
snes
s:
Item
:
5 10 15 20 2
5
30 35 40 4
5
50 55 60 Am
ong them: 20
30 60, reve
rse
score.
2.3. The Que
s
tionnair
e a
nd the Resul
t
s of the Sur
v
e
y
1. When you
need to comp
lete a que
stio
nnaire dep
osi
t
ed in the archives or a
ppl
y for a
job can be used, if you will choose to
lie in order to obt
ain better results.
A
. will
B
. will not
2.Wheth
e
r
ca
n be found in
the test subj
e
c
t is goo
d for
me?
A
. Yes
B
. No
3. When found the probl
em conte
n
t again
s
t you, you will choo
se?
A
. lighter than the actu
al levels of the answe
r
B.
The an
swer to a neutra
l
C.
Elected a
favorable an
swer di
re
ctly
D.
Select the option that
conform to mine
4. When you found the p
r
oble
m
co
nt
e
n
t in your favor, you will ch
oose?
A.
A height than the actual leve
ls of the an
swer
B.
The an
swer to a neutra
l
C.
Choo
se the stro
nge
st option
s
dire
ct
ly
D.
Select the option that confo
r
m to mine
Table 2. The
Survey Re
sul
t
s
ABBA AABA
AABD
BABB
BBBA
AAAD
BABD
AACC ABAD BADD
BBDD BADD
AABD
AABA
BADD
AABA BADD
BAAD AABD
AAAD
AABD AACD
BADD AAAD BACD
BBDD BADD ABDC
BBBD
AABD
BADD AABB BADD
BBDD
BABD
AAAD BADD
AABB
BABD BADD
AAAD
BADD BADD BBDD
AABD
BBBA AAAD
AAAD
BADD
AABD
BAAD BAAD
BADD
BABC ABBD
ABAD
AAAD BBBD BABC
BBBD
AABB AACD
BADD
BADA
BABD
BADD
BABD BACB ABAD
AABD
BBBD AABD
BBDD
BACC
BADD
BBAB
AABD AACB BAAD
BABD
BABD
AABD AAAD BAAD
ABBD
AAAB BADD BBBA
AABD
BBBD
ABAA BBDD
BADD
AABD
AABD
BABB
AACD BADC BADD
ABAD
BBBD BBAD
BAAA
BADC
AACC
3. Hierarchy
Structure M
odel
3.1, Modeling
In actu
al
surv
ey, as
doe
s
n
o
t provid
e a
n
y
of
the bi
g fi
ve perso
nalit
y test qu
estio
n
s, the
r
e
is
a
certai
n p
e
rcentag
e of the
st
u
dent
s
said
that th
e
y
coul
d
not
d
i
stingui
sh
wh
ether the title
to
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 7, July 201
4: 5585 – 55
90
5588
their a
d
vanta
ge o
r
di
sa
dvantage,
so
we cann
ot jud
ge o
n
wheth
e
r to li
e to. B
u
t the offer a
fter
several quest
i
ons
of the big five personality te
st, only a classm
ate sai
d
there are
still some
que
stion
s
are difficult to
disting
u
ish whether
fo
r the spe
c
ific
scenari
o
of default, the sch
ool
archives o
r
face
whe
n
the comp
any recruitment
of
choo
se a
n
d
employ persons i
s
favora
ble,
lowe
r perce
n
t
age of overall investigat
ed in view
of the students, this article only to find the
unfavora
b
le f
a
ctors app
ea
r more fu
rther
discu
ssi
on
on
the
hi
era
r
chi
c
al structu
r
e model
i
s
set
up
as Figu
re 1.
Figure 1. Hierarchy Structu
r
e Mod
e
l of Lying Survey
Re
stricte
d
b
y
time and scope, this
surv
ey in
se
veral unive
rsities thro
ugh
paper
que
stionn
aire
and network in the form of the
first-year unive
rsity undergra
duate course
to
grad
uate stu
dent grade t
w
o stu
dent
s con
d
u
c
ted a
rand
om surv
ey, received
reply to 105.
The
test of the four que
stion
s
t
o
a ce
rtain way to
prevent
these qu
esti
onnai
re
pa
rticipants lay. When
dealin
g with evaluation
qu
estion
naire re
sults, by
3,
4, topic b
o
th o
p
tions a
s
a
standa
rd
a
c
tual
ly
hone
st an
swer, 1 topi
c ch
oice fo
r refe
rence, a fo
llow-up
study, not
be discu
s
sed
in this pa
pe
r. In
the pro
c
e
ss o
f
the above model, only for the third q
uestio
n
in the
questio
nnai
re wa
s discu
s
sed,
whi
c
h aim
ed
at testing fou
nd in adve
r
se to the
su
bj
ects
of a1 co
ndition was
d
i
scusse
d. In the
choi
ce to lie
there a
r
e fou
r
peopl
e in the popul
ati
on
to the third topic
sele
cted
item, or for the
third qu
estio
n
answe
r them
as h
one
st op
tion, theref
o
r
e, in the p
r
o
c
ess of
calcula
t
ion, the su
bj
ect
of the four op
tions subje
c
ts were
cal
c
ulat
ed on hon
est
answe
r, in a highe
r level - not lying. Is for
the final resul
t
s of the third question
wa
s in 105 subj
ects, 80 pe
o
p
le cho
o
se lie (co
n
tainin
g
the
above fo
ur subje
c
ts).
On
the thi
r
d
que
stion, 23
subje
c
ts
of
the
s
e
option
s
,
45
subje
c
ts’ optio
ns,
eight parti
cip
ants option,
a total of 76
subj
ect’
s sele
ction agai
nst
what they think the pro
b
l
e
m
lies.
3.2. Solution Method for
Model
Acco
rdi
ng to the model, ba
sed o
n
the act
ual survey d
a
ta, approxim
ately take sca
l
e
paire
d co
mpa
r
iso
n
matrix is esta
blished.
The followin
g
:
Table 3. The
Importan
c
e S
c
ale of 1 ~ 9
Scale
Measure
ij
a
1
3
5
7
9
The importa
nce of
:
ij
CC
equal
A bit better
better
Obviously
st
rong
Absolute strong
Whe
r
e, 2, 4, 6, 8 is the middle cl
ass bet
wee
n
the abo
ve [2].
Has he lied
？
No.
Y
e
s,and find
disadvantag
e
Choice the
lower class.
Choice the
middle class.
Converse
choice.
The original
choice.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Polyg
r
ap
h Surve
y
an
d Eval
uation Based
on A
nalyti
c
Hi
era
r
chy Pr
ocess (Zhi
xia
Ji
ang)
5589
Selection
of
paired
co
m
pari
s
on
matrix
11
/
2
3
21
6
1/
3
1
/
6
1
．
Obviousl
y this m
a
trix is
con
s
i
s
tent m
a
trix; do not
need to
do
the co
nsi
s
t
ency
che
c
k. Ope
r
ation
s
with the h
e
l
p
of
MATLAB software, pai
red
compa
r
ison
matrix get
the bigge
st cha
r
a
c
teri
stic root of paired
comp
ari
s
o
n
matrix and th
e corre
s
p
ondi
ng eigenve
c
t
o
rs
, the 3 big
gest characte
ristic
root
s an
d its
corre
s
p
ondin
g
eigenve
c
tors (0.30
00, 0.6
000, 0.100
0).
Put the feature vector is m
u
ltiplied by the num
be
r of total que
stion
naire
respon
se to lie
purp
o
se
ratio
co
mbinatio
n
weig
ht vecto
r
(0.2
171,
0.
4
343, 0.0
724
). Te
st ea
ch
ari
t
hmetic i
s
a raw
score valu
e
of 1, sco
re value from 1 t
o
5, whic
h completely accords
with th
e scores of o
p
tion
value of
5 p
o
ints. T
h
ro
u
gh
cal
c
ulatio
n, the te
st
survey sho
w
e
d
that
if a certai
n subj
e
c
ts
compl
e
tely a
c
cord
s
with the su
bje
c
t of a again
s
t it, it’s were 21.
71% more likely to choo
se is
conform to, 43.43% probability of
selection, and 7.24% probability of selection is not in
c
onformity with the completely. On the bas
i
s
of the
origi
nal
score value
wa
s improved, the
incom
e
difference of
corre
s
po
ndin
g
poi
nts bet
wee
n
weig
ht and
o
p
tions
multipl
y
coupl
ed
wit
h
the
option of the
original i
s
worth to the n
e
w sco
r
e, na
mely the option co
mpa
r
ed
with a value
of
4.2171 p
o
ints sco
re, optio
n
s
gen
erally
score val
ue of
3.8686 p
o
ints, option doe
s
not confo
r
m to
the 1.2896
p
o
ints, fully meet and o
p
tio
n
com
pari
s
o
n
is not in conf
ormity with th
e gra
d
e
s
rem
a
in
the same,
still is 5 points and 2 point
s. At the same
time keep its test the origi
n
al scale divisi
on.
Und
e
r the scale, the actua
l
survey re
sul
t
s ar
e in the p
r
ocess of t
he
stru
cture of the
Judging matrix,
found to have h
i
gher levels
o
f
approximation components, tha
t
may
Affec
t
the re
sults.
In ord
e
r to improve the
accur
a
cy of
the r
e
sults
,
can change the
wa
y
That
the s
c
ale, impro
ve
the accuracy of
ju
dging matrix
in order
to ob
tain
more
ac
cur
a
t
e
result
s
.
Us
ing
10
10
~
1
8
2
sca
l
e [1
5],
jud
g
in
g m
a
trix
co
n
s
tru
c
te
d by
act
ual
survey results and to calculate agai
n.
Table 4. 10/1
0
~ 18/2 Esta
blish
ed Degree of
the Importan
c
e of the
Dimen
s
ion T
able
measure
ij
a
1
1.5
2.333
4
9
The importa
nce of
:
ij
CC
equal
A bit better
better
Obviously
st
rong
Absolute strong
Whe
r
e t
he im
portan
c
e
of d
i
mensi
on
ij
a
lev
e
l divisi
on
ca
n u
s
e th
e fo
rmula
said
9
11
k
k
,
among
k
take
2,4,6,8 fo
r i
n
terme
d
iate
values t
hat
as
sh
own in
the ta
ble
a
bove b
e
twe
e
n
adja
c
ent level
s
. Accordi
ngl
y to get the scale of pai
red
compa
r
i
s
on
matrix.
．
10/10 ~ 1
8
/2 were cal
c
ul
ated with the h
e
lp
of MATL
AB software is esta
blished
degree
s
above th
e
maximum
ch
ara
c
teri
stic root pai
re
d
compa
r
ison matrix
an
d
t
he co
rre
sp
o
nding
eigenvectors,
get the maximum charact
e
risti
c
root
of 3,Its feature vect
or i
s
(0.3107,0.5857,
0.1035
).
C
o
ns
is
te
nc
y in
s
p
e
c
tion
:
ma
x
33
0
13
1
n
CI
n
00
.
1
CI
CR
RI
10
7
1
5
10
13
5
13
10
17
71
0
3
53
1
0
15
17
10
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 7, July 201
4: 5585 – 55
90
5590
Re
sults th
rou
gh the
con
s
i
s
tency test. Th
e ch
ara
c
te
rist
ic vecto
r
mult
iplied by the
numbe
r
lies with the a
m
ount of que
stionn
aire
re
spon
se pu
rp
o
s
e ratio co
mbi
nation weight
vector (0.2
24
9,
0.4240,
0.07
49). After calcul
ation, u
nder the
sca
le
of 10/10
~18/2,
if a certai
n subj
e
c
ts
compl
e
tely accord
s with t
he su
bje
c
t of a against its, he were 2
2
.
49% more li
kely to cho
o
s
e
match, 42.4
0
%
proba
bility of selectio
n, and 7.49%
prob
ability of sele
ction is
not in confo
r
mity
with the
co
mpletely. Th
e re
sult
s cl
o
s
er to
the a
c
tual
su
rvey data,
therefore
have hi
gher
credibility. According to thi
s
sc
ale
score value to improve the or
igi
nal
score val
ue, the income
differen
c
e of
corre
s
p
ondin
g
points
between weight
a
nd option
s
m
u
ltiply and ad
d this optio
n
is
worth to o
r
igi
nal new
scores, one optio
n is in li
ne wit
h
the score v
a
lue of
4.224
9 points, opti
ons
gene
rally
sco
r
e valu
e of 3
.
8480
points,
option
doe
s not confo
r
m
to the 1.2
9
9
6
point
s, opti
on
fully meet and option com
pari
s
on i
s
not
in conformi
ty with the grad
es re
main the
same, still is 5
points a
nd 2
points. At the same time
ke
ep its test the
original
scale
division.
4. Conclusio
n
For th
e m
e
a
s
ure
lying
pro
b
lem, thi
s
p
a
per give
a n
e
w m
e
thod
th
at use A
H
P t
o
set u
p
index sy
stem
to the subje
c
tive indi
cato
rs
of
two
co
mpari
s
o
n
an
d evaluatio
n. After throu
g
h
the
con
s
i
s
ten
c
y test, we give the power co
efficients to the perso
nalit
y asse
ssmen
t
in the company
recruitme
n
t
process a
nd scho
ol ke
epin
g
file. We
est
ablish test score
s for the b
i
g five person
a
lity
to avoid lyin
g
.
Not o
n
ly tha
t, in the
cal
c
u
l
ation
of
th
e probl
em,
the
scale of
calculation accu
racy
than
conventi
onal u
nde
r th
e scale
of
cal
c
ulatio
n a
c
cu
racy i
s
highe
r, so the
data
of the form
er
as
a reference, in order to achieve a higher credibility.
Referen
ces
[1]
Saat
y
,
T
homas
L.
T
he Anal
y
t
i
c
Hierarch
y
Pr
ocess. Ne
w
Yo
rk: McGra
w
Hil
l
.
1981.
[2]
Jian
g QY, Xie
JX, Ye J. Mathematica
l
Mode
l
(T
he fourth ed
ition). BEIJING: Higher Ed
uca
t
ion Press.
201
1.
[3]
Li Y, Z
hu JS. An Improvem
ent an
d Its Applicati
on
of AHP
. Electronic
Desig
n
En
gin
eeri
n
g
. 2
011;
19(1
5
): 29-3
5
.
[4]
Du
LN. Mu
lti-
attribute
Decis
i
on
Maki
ng
b
y
Ana
l
y
t
ic H
i
er
arch
y Pr
ocess
.
Journ
a
l
of Gansu
Li
an
h
e
Univers
i
ty (Natural Sci
ences)
.
2007; 2
1
(5): 3
1
-35.
[5]
T
homas L Saa
t
y
.
Dec
i
sio
n
m
a
kin
g
w
i
th
the
ana
l
y
t
i
c hi
erarc
h
y proc
ess.
Int. J. Services Science
. 20
08
;
1(1): 83-9
8
.
[6]
Lia
o
HQ, Qiu Y, Yang X, W
a
ng
XG, Ge RW. A St
ud
y
of Weig
ht Coefficie
n
t Computi
ng
Method Bas
e
d
on AHP
. Me
cha
n
i
ca
l
En
gi
ne
er
. 2012; (6): 22
-25.
[7]
Den
g
X, Li JM,
Z
eng HJ, Che
n
JY, Z
hao JF
.
Researc
h
on c
o
mputati
on me
t
hods of AHP w
e
ig
ht vecto
r
and its ap
pl
icat
ions
. Mathe
m
a
t
ics in practice
and the
o
ry
. 20
12; 42(7): 9
3
-1
00.
[8]
Li Z
,
Z
hou SG
, W
ang K.
A Method for
Co
nstructing P
e
rfectl
y
Co
nsiste
nt
Judgm
ent Matrix in
AHP.
Jo
u
r
na
l
o
f
Zh
en
g
z
hou
Un
i
v
ersity (Nat. Sci.
Ed.)
. 2008; 40(
1): 41-46.
[9] Gao
S.
T
h
ree Calcu
l
ati
ng W
e
ig
hts Metho
d
s
in An
al
ytic H
i
erarch
y
Proce
s
s
. Science T
e
chno
logy
an
d
Engi
neer
in
g
. 2007; 7(2
0
): 520
4-52
07.
[10]
W
ang H, M
a
D
.
Scale
Eval
uti
on
and
Ne
w
S
c
ale M
e
tho
d
s.
Systems E
n
g
i
n
eeri
ng T
h
eory
and
Practice
.
199
3; 13(5): 24
-26.
[11]
X
u
K, Lia
ng
XC. Guarantee
Agenc
y
Efficie
n
c
y
Eva
l
u
a
tion
Based on Su
p
e
r Efficienc
y
D
EAAHP Mode
l.
T
E
LKOMNIKA Indon
esi
an Jou
r
nal of Electric
al Eng
i
ne
eri
ng.
2013; 1
1
(7): 3
981-
398
9.
[12]
Liu L
L
, Ch
en H
,
Z
hang R
X
.
C
o
mpre
hens
ive
Evalu
a
tion
of Examin
ation Qu
ality Base
d o
n
F
u
zz
y
AH
P
.
201
3; 11(9): 53
84-5
394.
[13]
Luo F
,
Z
han
g HC. Methods
of C
opi
ng
w
i
t
h
F
a
king of Per
s
ona
lit
y
T
e
sts
. Psycholo
g
ic
al
Explorati
o
n
.
200
7; 27(4): 78
-82.
[14]
Z
uo
HN, Li
ao F
L
.
Methods
of
Lie
Det
e
ction
of Perso
nal
it
y
T
e
st.
Chin
a Jo
urna
l of
Hea
l
th
Psychol
ogy.
200
9; 17(4): 50
5-50
7.
[15]
W
u
SQ.
T
he Applic
atio
n of Improve
d
AHP
to
Enterprise Info
rmatio
n
Eva
l
uati
on Ind
e
x
W
e
ight.
Chi
n
a
Mana
ge
me
nt Informatio
n
i
z
a
t
i
on.
201
1; 14(1
1
): 53-55.
Evaluation Warning : The document was created with Spire.PDF for Python.