免费文献传递   相关文献

SIC algorithm based model updating for near infrared analysis of Salvia miltiorrhiza alcohol extraction process

基于SIC算法的丹参醇提过程近红外定量模型更新研究



全 文 :2016
y

z ©
41
ª©

}
Vol41,No.5 March,2016
·
@efg@
·
£¤
SIC
϶Ž¸¹ÐÑÒÓ
ÔÕÖ× šØÙژ™
ƒ„…
1,

1,2,

1,
Cˆ‰
1,
ŠH‹
1,2,
Œ5
1,2
(1
hiçXZ?¨ çZCø›£qrç=

hi
100029;
2
çZéùú£ãän¡â¢£hiûabcde

hi
100029)

op
] 
2¸¹ÐÑÒÓ¯˜™—ö

ÉÊ=䕛ÉÚÒÓ(ŽÔÕÖ+ç

!2
HPLC
(×¹¯¹Ì¹

ÉÊ°Ç:
<¼¶Êõ¸¹9ⅡAŽ× .•šØ,±¸¹Ñ-ÒÓ¼nŒ(。,¤/,ˆöʎŸ KÿÂ,+ŒË/BšØ!Ü
Â

£¤ú

µ^Qö#“#Jʎ϶
(simpleintervalcalculation,SIC)
²Ê¤$,Ñ-ÒÓÔÕÖ× šØÙÚ


ran
domselection(RS)
Ùڕ¶¼n§ë

£¤
SIC
϶ŽÙڒy
:RMSECV,RMSEP,RPD
‡×¯
00068g·L-1,00054g·
L-1,314;
£¤
RS
•¶ŽÙڒy
:RMSECV,RMSEP,RPD
‡×¯
00064g·L-1,00068g·L-1,250。
’yóC
SIC
§
RS
ÙÜ
,¯
ÌH$,Ñ-ÒӎŸ àôªŒ(Ñ6¿˜™£»


qrs
] 
¸¹ÐÑÒÓ

ÔÕÖ

šØÙÚ
;SIC
϶

°Ç:<¼¶

tu)(
] 20150918

vwxT
] 
hiû‹Œü¨g*xy
(7154217);
çZéùú£ãän¡â¢£hiûabcde
2015¯
üü«°Hg^Éfnðýé
x
(Z151100001615065);
hiçXZ?¨‹Æþÿ
(2015JYBXS063)

yzA{
] 
Œ5

—°

žŸéÏj
,Tel:(010)84738621,Email:yjqiao@263net;


žŸ

qr‡ˆˆçZ¡ânú£ãä
,Tel:
(010)84738621,Email:xubing@bucmeducn

A{„…
] 
ƒ„…

Οqré
,Tel:(010)84738621,Email:jiashuaiyun@163com
SICalgorithmbasedmodelupdatingfornearinfraredanalysisof
Salviamiltiorrhizaalcoholextractionprocess
JIAShuaiyun1,XUBing1,2,YANGChan1,CUIXianglong1,SHIXinyuan1,2,QIAOYanjiang1,2
(1.ResearchCenterofTraditionalChineseMedicineInformationEngineering,BeijingUniversityof
ChineseMedicine,Beijing100029,China;
2.BeijingKeyLaboratoryforTraditionalChineseMedicineProductionProcesControland
QualityEvaluation,Beijing100029,China)
[Abstract] Thenearinfrared(NIR)spectroscopyforoflinemonitoringofalcoholextractionprocessofSalviamiltiorhizawasinves
tigated,withhighperformanceliquidchromatography(HPLC)determinationofvalueforreferenceThepartialleastsquaresmethod
wasadoptedtoestablishthetanshinoneⅡAquantitativecalibrationmodel,soastodetectextractionprocessofSalviamiltiorhizaBe
causethediferencesbetweenbatchesofrawmaterialsmayendangertherobustnessoftheoriginalmodel,thesimpleintervalcalcula
tion(SIC)wasappliedinupdatingthenearinfraredquantitativemodelfortraditionalChinesemedicineextractionprocessforthefirst
time,andcomparedwithRandomSelection(RS)methodSIC′sfinalupdatingresultsshowedthatrootmeansquarewithcrossvalida
tion(RMSECV),rootmeansquareerorofprediction(RMSEP)andresidualpredictivedeviation(RPD)oftanshinoneⅡA were
00068g·L-1,00054g·L-1and314,respectively;butRS′finalupdatingresultsshowedthatRMSECV,RMSEPandRPD
were00064g·L-1,00068g·L-1and250,respectivelyThisstudysuggestedthatSICissuperiortoRS,andprovidedare
searchfoundationforqualitycontrolandmonitoringofSmiltiorhizaextractionprocessinthefuture
[Keywords] alcoholextractionprocessofSalviamiltiorhiza;nearinfrared;modelupdating;simpleintervalcalculation;partial
·328·
2016
y

z ©
41
ª©

}
Vol41,No.5 March,2016
leastsquares
doi:10.4268/cjcmm20160511
  
ÔÕÖ
(nearinfraredspectroscopy,NIR)
+ç‡
§()®mB–É“¤

‹¬ž

(4úÈ w„
Üq
[14],
Ôy:ðE>²Ê¤$,©ÖÒÓoäŸ
 Œà
,NIR
‡§º¤Ê؇§éê

Ì× ™çŽ
TŠÕ?A¥“ ^ŽšØ

ü°Ç:<¼µs
(PLSR)
šØ„
[57]。
0kBŽ$,ô,ÒÓÔÕÖ
× ‡§šØŽÊõ

íÉÊBCˆöŽÒÓ(4
¼nʚ

]øtj

ÍÖF

Ɂ}

tôùg„
®ì%&

$,/qæo[ן uŠ

©ÖÒÓ$
ك

Mu

›õ½À„“_

+%&ÒÓZƟ 
¹=¾¿Ž¬ðÂ
[8]。
ù‚u&²ŽuŠ

(ËI
JŽ“_„®ì

w+vQ
NIR
‡§šØS(Âç
ŽÀ_
[911]。
®ú

ÉÊBCˆö(4ÊõŽ× 
šØ

‹¶[,øBS}ŽÒÓv²

rÚ(4Š
O

+TTvQëwŽS(°K

®ú
,¯
Ñ°ÔÕ
Ö× ‡§o$,ô,ÒÓ$²ÊŽîH‘!Ü
Â

Õ±0ÊõŽšØ¼nÙڑ P

8_Ž Ž°±

šØÙÚÕ­×ÂÚ
Ž,B“ÿv²Ž(4

#Ìù‚/šØŽ.•Ú
$

¬Ú.•!ÊõڎšØ
[12],
2Ñ°šØŽv
²°±é1‘3²Â

šØÙڎZÆo¤üP°
WF6[,“ÿv²(4

‰^•kB
,Ken
nardStone(KS)
϶
[13]、Duplex
϶

> =¶
(Mahalanobisdistance,MD)、
;‡‡§
(PCA)[14]
‘“#Jʎ϶
(simpleintervalcalculation,
SIC)[15]
„íîTŠ“ÿv²×

4^2$,¸
¹“ÐÑ-ÒÓ¯˜™±C

ÉÚ¥ˆöÒÓÔÕ
Ö+ç

Êõ¸¹9ⅡAŽÔÕÖ× ‡§šØ,œ
:#
SIC
϶²Ê¤•ïg`–ÉIJL$,Ñ-
ÒÓ,“ÿv²(4Ž×

!TŠÑ-ÒÓ× 
‡§šØŽÙÚ
,¯
ÔÕÖ+ç()o$,©ÖÒ
Ó$Ž²ÊÑ6¹Ì

1 
q
Agilent1100
+Y>ç‚

Ç%5ÁtDE
),
Antaris
8²6“ÂÔÕÖ+ç‚

Ç%
Thermo
Nicolet
DE
),
$,TU…ØŠ?Ñ-f{

23ñ
ÃO)‚u›õBCDE
)。
¸¹ÝÞ

ˆ

íËã234T•>,)BC
DE

ˆ<‡×¯
20140308,20131226,20130808,
Ö
FD5
),
±Ä¸¹9ⅡA($%W,Åט
™}

ˆ<
110766200619),
“”

ëÐ

>çÆ

Ç
%
Fisher
DE
),
ÄÅÅÆ]*

ö·ÄÅÅÚÑB
CDE
),
Ñ-Ê*¯ãô°Æ*

ÌØË|í¯‡
§Æ

2 
•¶
21 
¸¹Ñ-+ôõ
¹Ì

$%,Ò
》2015
y渹ÞÑ-g`I
J

³-¸¹ÝÞ3 C¤
2L
L’!ÃƟ$

ù

ý 
90%
“Ð
,80℃
来-
15h,
ÇȞ±
100r·min-1。
ùå°ßÃÄ

ÊÉ
2min
-(

É
ʋßÃÄ

ÊÉ
6min
-(

|ö-(
15mL,
(

9000r·min-1
=S
10min,
-¼w+‡×( 
NIR
‘
HPLC。
³-¸¹ÝÞ

û


150g,
¼n

ˆÑ-
–É

Ì$©
1,4,5
‘

ˆÉʈ<¯
20140308
Ž
¸¹ÝÞ

©

ˆÉʈ<¯
20131226
Ž¸¹Ý
Þ

©

ˆÉʈ<¯
20130808
Ž¸¹ÝÞ

¯w
ùÒӓÿÂ

©

ˆùå°ßÃÄ

ÊÉ
30s
-
(

ÉʋßÃÄ

ÊÉ
3min。
,¤xˆÑ-ÒÓ
$êÎÉʎOÊ]

©
1~6
ˆÚΎ(4=
‡×¯
23,23,23,22,62,23。
ˆö
1~3
Ž(4É
¯.•Ú

ʤÊõ-!
PLS
šØ

©

ˆ(4ɯ
81U€Ú

©
5,6
ˆÉ¯Ö1S(Ú

o
PLS
š
ØÙژ™$
,SIC
϶ʤ©
4,5
ˆ$,“ÿv
²(4Ž6

©

ˆÊ:U€ÙÚHŽšØÂç

22 NIR
+çÉÚ
NIR
(×IJ¯…ßMuL

Éʄ˕›É
Ú+ç


8mm,
Ëôé1
10000~4000cm-1,
‡6}¯
4cm-1,
Ëôö=
32
ö

|ñ(Ñn


ö+ç

$Ñí+ç

23 
Ñ-+$¸¹9ⅡA, ŽHPLC(×
231 
>çIJ
 
¹Ä

$%,Ò
》2015
yæ>ç
IJ

ÉÊ
AgilentSBC18>ç–(46mm×250mm,
5μm),–ß30℃,¥ž08mL·min-1,Å(uý
270nm,
¼( 
10μL,¥ŠY“”(A)002%ö
»*Í+
(B),
±kè
(0~6min,61% ~61%A;
6~20min,61% ~90%A;20~205min,90% ~
61%A;205~30min,61%~61%A)。
·428·
èe4„

£¤
SIC
϶Ž¸¹ÐÑÒÓÔÕÖ× šØÙژ™
232 
±ÄÍ+ŽFô
 
ã.³ ¸¹9ⅡA±
ā

!ÊëÐÍâב

Š¥HFô
8088
mg·L-1
±ÄÍ+

233 
6ˁÍ+ŽFô
 
-(Ò
NIR
Ž6ˁ
/+6ØÊ
045μmòÕ¥í¥Ò,-Â¥+,áá。
24 
=‰úȑqJ
‡×ÉÊ
SG
Ñm
(savitzkygolaysmoothing,9
q
)、
[Ãv=
(firstderivative,1std)、
¥~¢Ë.•
(multiplicativescatercorection,MSC)、
–—•ù“
Â
(standardnormalvariate,SNV)
‘:uLÌ
(wave
letdenoising,WDS)
¼nSúÈ

£¤wç°Ç:<
¼
(PLS)
šØ

2.•ÚYZŸ=
rcal、.•VKí
•”
(rootmeansquareerorofcalibration,RMSEC)
‘«,U€í•K
(rootmeansquareerorofcross
validation,RMSECV)¯
÷ø¦–

MÂÇÜSúÈ
•¶

äHÉÊ{ñÊÉ°Ç:<¼¶
(synergyin
tervalPLS,siPLS)
±ÊšuļnLM

Êõ
PLS
š
Ø

÷ø¹=¯«,U€í•K
(RMSECV)、
S(V
K핔
(rootmeansquareerorofprediction,RM
SEP)
ËY²ŽYZŸ=
r。
ÔÕÖ+çÉÚÍÊ
ThermoScientificRESULT
qJ

+çSúÈÉÊ
SIMCAP+115(Umetrics,
Sweden)
qJ
,PLS
϶^Ê
PLSToolbox21(Eigen
vectorResearchInc,USA),siPLS
϶ÍÊ
iToolbox
gm[

ío
Matalb
qJÑJ¼TŠ
。SIC
϶$


‘ βŽÿŽÍÊ SICToolboxo MatalbÑJ
¼TŠ
[1617]。
3 
’yªÀ
31 
Ñ-+$¸¹9ⅡAŽ, (גy
Êõ¸¹9ⅡA–—ãä,•Ó¯Y=61167X+
51729(R2=09995,n=8),
äÂé1o
00809~
16176μg,ɯ.•Úˆö1~3Ž¸¹9ⅡAŸ G
±
00233~00867g·L-1,
ɯ(Ëڈö
4~6
¸¹9ⅡAŸ G±‡×¯ 00246~00849,
00169~00873,00269~00813g·L-1。
’yó
Cúˆö

Ö

Ì؈ö¸¹9ⅡAŸ G±íº%o
.•Úé18

32 
+çSúȑ“ LM


ˆö¯—

/!+çqxq
1。1
“

7100cm-1
º¤ÔÕÖ]uàŽ<…ýÎJ

7100~5000cm-1
¯[…ýÎJ
,5000~4000
cm-1
¯{ñÎJ

#Éڎ/!+ç¼nʚ

Ê
š’y
:rcal¯08981,RMSEC¯00078g·L
-1,
RMSECV
¯
00110g·L-1。
,¤<…ýÎJŽà
&±ëì
,5150cm-1
úŽ&à«*‡<α
³Ï½“Š‘Å㓊Ž{ñÎç‡

Šo
4550
cm-1
áԎàþ‘—;­«“ÐŽ
OH
Ͻ“Š
‘
OH
Å㓊Ž{ñÎé–
,¯
Ñ°šØÂç

M
•u=
7000~4000cm-1
/!+ç

!Lúþ‘—
u=
5116~5290,4499~4590cm-1
H¼nʚ

xq
2(
Ú>ß%1‡
),
ʚ’y
:rcal¯09153,
RMSEC
¯
00071g·L-1,RMSECV
¯
00086g·
L-1,
ªwçŽ/!+çY§ë

ʚWyBøÑ°

q
1 
/!
NIR
+çq

ˆö
1)
Fig1 RawNIRspectraofbatch1
q
2 SG
ÑmSúȎ{ñÊÉ°Ç:<¼Êš
(siPLS)
Ç
ÜuĒy

ˆö
1)
Fig2 SynergyintervalPLSofsavitzkygolaysmoothingfor
batch1
®ú

±]+çSúȕ¶¼n§ë

’yx
ó
1,
îx
SG
ÑmSúȶÊõ¸¹9ⅡAPLSš
Ø
,rcal¯09551,RMSEC¯00052g·L
-1,RM
SECV
¯
00081g·L-1,
šØÂçBëwÑ°

·528·
2016
y

z ©
41
ª©

}
Vol41,No.5 March,2016
ó
1 
]+çSúȕ¶Ž§ë
Table1 ComparisonofdiferentpreprocessingmethodsforNIRspectra
•¶
n“ 
®<=
.•Ú U€Ú
rcal RMSEC/g·L-1 RMSECV/g·L-1 BIAScal rpre RMSEP/g·L-1 RPD BIASpre
raw 7 09153 00071 00086 00530 04487 00212 084 00120
1std 7 08855 00082 00135 00062 01434 00327 054 00196
MSC 9 09660 00045 00087 00036 05574 00178 100 00107
SNV 9 09661 00045 00087 00036 05565 00177 100 00107
WDS 10 09441 00058 00072 00048 05367 00284 063 00235
SG 10 09551 00052 00081 00040 08097 00136 131 00090
  
¼[çÉÊÊÉ°Ç:<¼¶
(siPLS)
϶¼
n¡¢uÄLM
,siPLS
¹=›C¯

ÊÉ=
20,
Ç
wn“ ®<=
10,
{ñ=
3,
LMŽÇÜuį
5365~5500,5502~5637,6187~6322cm-1,
x
q
2(
²>ß%1‡
)。
:ÊÇIuÄ{ñ¼nÊ
š
:rcal¯09811,RMSEC¯00034g·L
-1,RM
SECV
¯
00053g·L-1,rpre¯09431,RMSEP¯
00058g·L-1,RPD
¯
306。RMSEC,RMSECV,
RMSEP
ºBø_þ

YZŸ=

‘
RPD
BøÑ°

ËÌ«
RPD
Ñ°§ëà{

33 
.•šØŽÊõªS(
4^±

ˆ¸¹Ñ-+(4¼n
PLS
ʚ

RMSEC,RMSECV,PRESS
o

ñn“ ®<=O

*+¤Ñ!

ù
RMSEP
o

OÇþ

H*+ww

xq
3。
Ëñx¦–“_±

M•n“ ®<=
¯

O

Êõ.•šØ

úIJL.•šØ.•Ú
‘S(Ú$Ž
HPLC
¹Ì¹ª
NIR
S(¹ŽYZZ
Ÿxq
4,
óCšØ.•‘S(ÂçHA

ˆö
4~

ŽU€’yxó
2,
îxˆö

Ž
RPD
íw¤
5,6
ˆŽ
,RMSEP
‘
BIASpreí:¤ˆö5,6,óC¸¹
Ñ-ÒÓ$æoˆÊKÿ

/BŽšØ‹¶°±Ñ
-ÒÓ$ڎ“ÿv²

Õ­±/šØ¼nÙ
Ú P

q
3 
ÇÜn“ ®<ŽM•
Fig3 Calibrationcharacteristicsvsnumberoflatentfactors
q
4 
.•ÚªS(Ú$¹Ì¹ªS(¹ŽYZZŸq
Fig4 CorelationdiagramfortanshinoneⅡA
ó
2 
ˆö
4~6
ŽU€’y
Table2 Valdationstatisticsofbatches46
ˆö
.•Ú S(Ú
rcal RMSEC/g·L-1 RMSECV/g·L-1 BIAScal rpre RMSEP/g·L-1 RPD BIASpre
4 09811 00034 00053 00027 09431 00058 306 00044
5 09811 00034 00053 00027 08886 00105 194 00088
6 09811 00034 00053 00027 09304 00089 190 00074
  


n“ ®<=í¯
9。
·628·
èe4„

£¤
SIC
϶Ž¸¹ÐÑÒÓÔÕÖ× šØÙژ™
34 
PóÂ<ÚLM
,¤ÊšŽ.•ÚŽÔÕÖ+çîçæoÐS


ošØÙÚý

Õ­#,ÐS+çv²Ž(4
VúÑ

ÒLŽÉ¯<Ú¬ÚÊõšØ

Vú•¶
¯

ÃÄD›
(5)
‘
(6)
žŽÏ
SIC
Ó%¹‘µK
¹

äH‡”‰
SIC
϶Ž
OSC


±-!šØ
Ž.•Ú(4¼n‡•

’yxq
5,
q$ÖB
69
ñ(4
,62
ñØS(4
,7
ñ81(4

81(4¯
ÐSv²(4

#n.(4VúL

ÊÒLŽ
62
ñ
(4¬Úʚ

2ˆö

ɯS(Ú

’y«n“ 
®<¯
9,rpre¯09401,RMSEP¯ 00084g·L
-1,
RPD
¯
203,BIASpre¯ 00069。ùÉÊbVú
  
ØS(4ÊõŽšØS(©

ˆ(

’yxó
2,
n“ ®<¯
9,rpre¯09304,RMSEP¯00089
g·L-1,RPD
¯
190,BIASpre¯00074。,úî2
zC

,PóÂ<ÚÊõŽ
PLS
šØŽS(Âܤ
-!Ž
PLS
šØ

q
5 
-!šØ.•ÚŽ
OSC
‡•/ù
Fig5 SICobjectstatusplotroetheinitialcalibrationset
35 
šØÙÚÐK
”‰D›
(5)
‘
(6),
8S(°±Ý

±mBë
w
SIC
Ó%¹‘µK¹(4ŽS(VKëwŽS(
Wy]I

Ú8.•Ž°±Ý

ëwŽ
SIC
Ó%¹
Ž(4ù‚ÎúšØ.•Ú$

BA¤µšØ°±
륎“ÿv²

8ùÑ°šØŽ!ÜÂ

”‰D
›
(5),
r
h=05
O

JÊ·±

w:¯βmax,ùβmax
òYr¤

ýŽ
RMSEC[15],
á
v=[-2RMSEC,
2RMSEC],
®ú
,SIC
Ó%¹£¹
05
î2ò
95%
Ž
7}O€LMÂ(4Ž“ÿÂ

£¤
SIC
϶Ž

ñ^Ž 

áÊ
SIC
Ó%¹
‘µK¹±](4Ž÷ÿøLMŽØS(4!
-#
SIC
Ó%¹w¤
05
Ž(4ù‚ΚØ.•Ú
$¬Úʚ

LS(Ú[ˆ(4ŽšØÙÚ

ŠB

/ÙÚÐK

TpÑ-ÒÓ

ñˆö

ˆö
1~3),
ÖB(4
69
ñ

2n
69
ñ(4ÊõšØ
“Ⅰ”,Ê
SIC
϶LMŽØS(4
62
ñÊõšØ
“A”。
o
ú£»¼

Ö©

/šØÙڕ¶

•¶[

ÉÚÚ

———
ˆö

ŽÒÓ
NIR
+ç‘
HPLC
¹Ì¹

,D›
(5)
‘
(6)
ŽÏ

LMÂ
SIC
Ó%¹w¤
05
Ž(4

!#n.(4ù‚šØ

Ž.•Ú¬ÚÊ
š

úšØù¯šØ
“B”。
È

”‰D›
(5)
‘
(6),
±ˆö

¼nLM


SIC
Ó%¹w¤
05
Ž(4ù‚šØ

Ž.•Ú¬Úʚ

úšØ¯š
Ø
“C”。
•¶<

ošØ
“Ⅰ”Ž£»¼,ÉÊrandom
selection(RS)
϶8ڎˆö

$ÔM
1/3
ŽPó
Â(4

#n.(4ù‚šØ
“Ⅰ”,¬Úʚ,úš
د
“Ⅱ”,È,Éʵ϶8ˆö5$M 1/3
ŽPóÚ(4ù‚šØ
“Ⅱ”,¬Úʚ,µšØù
¯
“Ⅲ”。
n

/Ùڕ¶ŽšØ!Ü‘S(Âç’y

3,
U€šØÂ玈ö^[Éʈö

Ž=
‰

8ó$î2mŠ

£¤
SIC
Ùڕ¶
,RMSEC,
RMSECV,RPD
º«*+ww
,RMSEP
´*+w:
±

ù±¤
RS
Ùڕ¶

ú¿
RMSECV,RM
SEC,RPD
´žwH_Ž±
,RMSEP
£«žwH
w

uö:z
,2
/•¶ÙÚHŽšØS(Âçº
ܤÙÚýŽšØ



öÙÚH
,SIC
§
RS
ÙÜ

¯¿Ù¼[ç.¥
SIC
Ùڒy

Ê9–‹?
‡•q:½ºÙڒy

xq
6。
ˆö
4,5
Žw1
(4Ž
SIC
Ó%¹ºw¤
05,
w1ðù‚ΚØ
Ž.•Ú

Ê2šØÙÚ

zCˆöʎKÿëw

(Ž,

©Öˆö]+vQցŸ ][(

ÝÕ­šØÙÚ

4 
’À
4^ÉÊÔÕև§()±¸¹#öÐÑÒÓ
¼nŒ(

2¸¹9ⅡA¯¦–‡,Êõ¸¹Ð
ÑÒÓ¸¹9ⅡAŽÔÕÖ×  PLSšØ,O±
]ˆöŽÑ-ÒÓ¼nšØÙÚ

4^Ñ6¿

/šØÙÚÐK

Qö#
SIC
϶²ÊÎ$,Ñ-
ÒӎšØÙÚ

OÑÉÊ
RS
¶±šØ¼nÙ
Ú

uŽ:z

ÏÒ

/šØÙÚÐKHŽšØ
·728·
2016
y

z ©
41
ª©

}
Vol41,No.5 March,2016
  ó3 ]šØÙڕ¶ŽšØÂç§ë
Table3 Comparisonoftheperformanceoftheupdatedmodelsvalidatedbybatch6
ÐK šØ (4=
n“ 
®<=
.•Ú U€Ú
rcal RMSEC/g·L-1 RMSECV/g·L-1 BIAScal rpre RMSEP/g·L-1 RPD BIASpre
1 A 62 9 09804 00035 00058 00030 09401 00084 203 00069
B 84 6 09489 00056 00062 00044 09540 00055 307 00043
C 146 7 09583 00056 00068 00045 09773 00054 314 00045
2 Ⅰ 69 9 09811 00034 00053 00027 09304 00089 190 00074
Ⅱ 76 6 09469 00055 00060 00043 09475 00058 294 00046
Ⅲ 153 7 09587 00050 00064 00040 09713 00068 250 00057
A
£¤šØ

Žˆö
4;B
£¤šØ

Žˆö
5。
q
6 SIC
9–‹?‡•q
Fig6 SICobjectstatusplotofbatch6
ÂçCàܤÙÚý

zC`šØ‹¶°±Ú(4
Z‡Žv²

ªBÙڎšØ#çÙAFS(Ú(
4

±§

/šØÙÚÐK

,’yî2ÝÂ

Ï
SIC2
öÙÚHŽšØŽ
RPD
w¤Ï
RS2
öÙÚ

,RMSEP

RS2
öÙÚHŽ

zC
SIC
ÙÚ
•¶§
RS
¶ÙBW

8ù^ÙÚHšØÙÜ

O
ú˜™Ñ¯ÌH¸¹Ñ-ÒӎŸ àôªoäŒ
(Ñ6¿˜™£»


|}?~

[1] FiserovaM,IIaA,Maholanyiova,MNondestructiveanalysisof
kraftpulpbyNIRspectroscopy[J]CelChemTechnol,2014,48
(3/4):181
[2] LuypaertJ,MassartDL,VanderHeydenYNearinfraredspec
troscopyapplicationsinpharmaceuticalanalysis[J].Talanta,
2007,72(3):865
[3] TianY,SunJ,YanH,etalArapidandconvenientmethodfor
detectingabroadspectrum ofmalignantcelsfrom malignant
pleuroperitonealefusionofpatientsusingamultifunctionalNIR
heptamethinedye[J]Analyst,2015,140(3):750
[4] RoggoY,ChalusP,MaurerL,etalAreviewofnearinfrared
spectroscopyandchemometricsinpharmaceuticaltechnologies
[J]JPharmaceutBiomed,2007,44(3):683
[5] XiaoX,MaJ,GeF,etalApplicationofnearinfraredspec
troscopyfortherapidanalysisofLoniceraeJaponicaeFlossolution
extractedbywater[J]JInnovOptlHealSci,2014,7(4):
1793
[6] WangP,ZhangH,YangH,etalRapiddeterminationofmajor
bioactiveisoflavonoidcompoundsduringtheextractionprocessof
kudzu(Puerarialobata)bynearinfraredtransmissionspectrosco
py[J]SpectrochimActaA,2015,137:1403
[7] WuZS,XuB,DuM,etalValidationofaNIRquantification
methodforthedeterminationofchlorogenicacidinLoniceraja
ponicasolutioninethanolprecipitationprocess[J]JPharmaceut
Biomed,2012,62:1
[8] XuB,LinZZ,WuZS,etalTargetorientedoveralprocess
optimization(TOPO)forreducingvariabilityinthequalityof
herbalmedicineproducts[J]ChemometrIntelLab,2013,128:
144
[9] 
®åÀ

ÕÖ

®×

„

ÔÕÖ+ç()o+_¨vÅ(
$Ž²Ê˜™
[J]
‡§&k
,2011,27(5):655
[10] DeBeerTRM,VercruysseP,BurggraeveA,etalInlineand
realtimeprocessmonitoringofafreezedryingprocessusingra
manandnirspectroscopyascomplementaryprocessanalytical
technology(PAT)tools[J].EurJPharmSci,2009,98(9):
·828·
èe4„

£¤
SIC
϶Ž¸¹ÐÑÒÓÔÕÖ× šØÙژ™
3430
[11] ManleyMNearinfraredspectroscopyandhyperspectralimaging:non
destructiveanalysisofbiologicalmaterials[J]ChemSocRev,2014,
43(24):8200
[12] LiW,XingL,FangL,etalApplicationofnearinfraredspectros
copyforrapidanalysisofintermediatesofTanreqinginjection[J]J
PharmaceutBiomed,2010,53(3):350
[13] TangJY,HuangM,ZhuQBPuritydetectionmodelupdateof
maizeseedsbasedonactivelearning[J]SpectroscSpectAnal,
2015,35(8):2136
[14] BlancoM,CuevaMestanzaR,PegueroANIRanalysisofphar
maceuticalsampleswithoutreferencedata:improvingthecalibra
tion[J]TaIanta,2011,85(4):2218
[15] RodionovaO,EsbensenK,PomerantsevAApplicationofSIC
(simpleintervalcalculation)forobjectstatusclassificationand
outlierdetection—comparisonwithregressionapproach[J]J
Chemom,2004,18:402
[16] XuB,WuZS,LinZZ,etalNIRanalysisforbatchprocessof
ethanolprecipitationcoupledwithanewcalibrationmodelupda
tingstrategy[J]AnalChimActa,2012,720:22
[17] RodionovaO,PomerantsevASimpleviewonsimpleintervalcal
culation(SIC)method[J]ChemomIntelLabSyst,2009,97
(1):64

€-0
 
ŽŽ

·928·