４．追加のプログラムの使い方
# Test2.1 for latin Table
# Cite from
# Washio, Yasutoshi,実験計画法入門 改訂版,
# "Introduction of Eexperimentation Design", 2nd. Ed.,1997,
# ISBN:4-542-50330-5, P193 ex. 10.3.
#
library("xtable")
y<-c(64.2,64.6,65.1,64.5,65.0,63.9,65.1,64.1,65.0,64.7,64.3,64.6,64.3,64.5,
64.3,63.9,64.9,64.1,65.0,62.9,64.8,64.2,65.3,64.0,64.8,64.8,64.0)
res1<-latin.table.L27(y)
res2<-summary(aov(data=res1,formula=y ~ f01+f02+f05+f09+f13+f01:f02+f02:f05))
res1<-xtable(res1)
print(res1,type="html")
res2<-xtable(res2)
print(res2,type="html")
# Test1.1 for Latin Table
# Cite from
# Washio, Yasutoshi,実験計画法入門 改訂版,
# "Introduction of Eexperimentation Design", 2nd. Ed.,1997,
# ISBN:4-542-50330-5, P166 ex. 9.3.
tmppar<-par(no.readonly=TRUE)
par(ask=TRUE)
y<-c(35,48,21,38,50,43,31,22)
result<-latin.table.L8(y)
colnames(result) <-c("B","C","B:C","D","B:D","A","nF","y")
result
summary(aov(data=result,formula=y ~ A+B+C+D+B:C+B:D))
summary(aov(data=result,formula=y ~ A+B+C+D+B:D))
attach(result)
interaction.plot(x.factor=B,trace.factor=D,response=y,col=c(1,2))
detach(result)
factable11<-latin.factor.plot(ltable=result,ncol=6,fun="mean",
mfrow=c(2,3),ylim=c(25,45),title="Latin Table L8 Test")
factable11
par(tmppar)
# Test 3.1 for latin Table
# Cite from
# Taguchi, Gen-ichi et. al., コンピュータによる情報設計の技術開発,
# "Engeenering R&D of Information Design with Computer Simulation", 2004,
# ISBN:4-452-51115-4, P173 Table 14.3
tmppar<-par(no.readonly=TRUE)
par(ask=TRUE)
y<-c(15.51,15.73,19.87,10.09,15.25,17.94,13.40,14.64,17.52,13.26,16.56,
17.16,15.72,15.04,18.27,14.83,15.80,16.85)
result<-latin.table.L18(y)
colnames(result) <-c("A","B","C","D","E","F","G","H","SN")
result
summary(aov(data=result,formula=SN ~ A+B+C+D+E+F+G+H))
factable31<-latin.factor.plot(ltable=result,ncol=8,fun="mean",
mfrow=c(2,4),ylim=c(13,19),title="Latin Table L18 Test")
factable31
par(tmppar)
# Test for Taguchi Method Standard SN 1
# Cite from
# Taguchi, Gen-ichi et. al., コンピュータによる情報設計の技術開発,
# "Engeenering R&D of Information Design with Computer Simulation", 2004,
# ISBN:4-452-51115-4, P. 41 Table 3.8
fno<- c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,1)
fsig<- c(1,1,1,2,2,2,3,3,3,4,4,4,5,5,5)
fnoise<- c(0,1,2,0,1,2,0,1,2,0,1,2,0,1,2)
noisedata<- c(3.255,3.013,3.497,
6.245,5.833,6.657,
9.089,8.590,9.588,
11.926,11.400,12.460,
14.825,14.320,15.330)
noise<- data.frame(fno=as.factor(fno),fsig=as.factor(fsig),fnoise=as.factor(fnoise),noisedata=noisedata)
res<-taguchi.param.sn.std(noise)
res
# Test for Taguchi Method Standard SN 2
# Cite from
# Taguchi, Gen-ichi et. al., コンピュータによる情報設計の技術開発,
# "Engeenering R&D of Information Design with Computer Simulation", 2004,
# ISBN:4-452-51115-4, P. 32 Table 3.2
fno<- c(1,1,1,2,2,2,3,3,3,4,4,4,5,5,5,6,6,6,7,7,7,8,8,8,9,9,9,10,10,10,11,11,11,12,12,12,13,13,13,14,14,14,15,15,15,16,16,16,17,17,17,18,18,18)
fsig<- c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1)
fnoise<- c(0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2,0,1,2)
noisedata<- c(14.32,13.86,14.77,
20.73,20.19,21.26,
27.28,26.70,27.85,
14.57,14.18,14.95,
20.89,20.43,21.34,
22.60,21.86,23.32,
14.86,14.53,15.19,
16.90,16.34,17.46,
23.02,22.39,23.64,
15.47,15.07,15.87,
21.93,21.46,22.38,
23.86,23.09,24.61,
13.36,12.92,13.79,
19.84,19.31,20.35,
25.94,25.37,26.50,
14.87,14.53,15.20,
16.79,16.22,17.34,
23.18,22.55,23.80)
noise<- data.frame(fno=as.factor(fno),fsig=as.factor(fsig),fnoise=as.factor(fnoise),noisedata=noisedata)
tres<- taguchi.param.sn.std(noise)
tres
tresult<-latin.table.L18(tres$SN)
colnames(tresult)<- c("A","B","C","D","E","F","G","H","SN")
tresult
tmppar<-par(no.readonly=TRUE)
factabletg<-latin.factor.plot(ltable=tresult,ncol=8,fun="mean",
mfrow=c(2,4),ylim=c(29.5,33.5),title="Standard SN")
factabletg
summary(aov(data=tresult,formula=SN ~ A+B+C+D+E+F+G+H))
par(tmppar)
# Test for Taguchi Method tune
# Cite from
# Taguchi, Gen-ichi et. al., コンピュータによる情報設計の技術開発,
# "Engeenering R&D of Information Design with Computer Simulation", 2004,
# ISBN:4-452-51115-4, P. 44 Table 3.11
m<- c(2.685,5.385,8.097,10.821,13.555)
y<- c(2.890,5.722,8.600,11.633,14.948)
taguchi.param.tune(y,m)
# Test for Taguchi Method Smaller Better SN
# Cite from
# Taguchi, Gen-ichi et. al., 開発・設計段階の品質工学,
# "Quality Engeenering in R&D and Design", 1988,
# ISBN:4-542-51101-4, P. 75 Table 4.1
R1<- c(12, 6, 9, 8,16,18,14,16)
R2<- c(12,10,10, 8,14,26,22,13)
R3<- c(10, 3, 5, 5, 8, 4, 7, 5)
R4<- c(13, 5, 4, 4, 8, 2, 5, 4)
R5<- c( 3, 3, 2, 3, 3, 3, 3,11)
R6<- c( 4, 4, 1, 4, 2, 3, 4, 4)
R7<- c(16,20, 3, 9,20, 7,19,14)
R8<- c(20,18, 2, 9,33,10,21,30)
data<- data.frame(R1=R1,R2=R2,R3=R3,R4=R4,R5=R5,R6=R6,R7=R7,R8=R8)
res1<- taguchi.param.sn.smaller(data)
res2<- latin.table.L8(res1)
colnames(res2)<- c("A","B","A:B","C","A:C","D","E","SN")
res2
latin.factor.plot(ltable=res2,ncol=7,fun="sum",
mfrow=c(2,4),ylim=c(-93,-72),title="Smaller Better SN")
# Test for Taguchi Method Larger Better SN
# Cite from
# Taguchi, Gen-ichi et. al., 開発・設計段階の品質工学,
# "Quality Engeenering in R&D and Design", 1988,
# ISBN:4-542-51101-4, P. 92 Table 5.1
K1<- c( 3, 8, 9, 9,10,14, 7,10,14, 8, 4, 9, 7,11,13,10,12,15)
K2<- c(12,10, 9,10,13,15,10,12,15, 8,14,10,10,12,14,11,14,17)
data<- data.frame(K1=K1,K2=K2)
res1<- taguchi.param.sn.larger(data)
res2<- latin.table.L18(res1-20) # 教科書と合わせるため，仮平均20dbを引く．
colnames(res2)<- c("A","B","C","D","E","F","G","e","SN")
res2
latin.factor.plot(ltable=res2,ncol=8,fun="sum",
mfrow=c(2,4),ylim=c(-18,12),title="Larger Better SN")
summary(aov(SN ~ B+C+D+E+F,res2))
# Test for Taguchi Method Nominal Best SN
# Cite from
# Taguchi, Gen-ichi et. al., 開発・設計段階の品質工学,
# "Quality Engeenering in R&D and Design", 1988,
# ISBN:4-542-51101-4, P. 109 Table 6.2
R<- c(1,1,1,2,2,2,3,3,3)
L<- c(1,2,3,1,2,3,1,2,3)
res2<- data.frame(R=as.factor(R),L=as.factor(L),y=rep(0,9))
N1<- c(21.5,10.8, 7.2,13.1, 9.0, 6.6, 8.0, 6.8, 5.5)
N2<- c(38.5,19.4,13.0,20.7,15.2,11.5,12.2,10.7, 9.1)
data<- data.frame(N1=N1,N2=N2)
res1<- taguchi.param.sn.nominal(data)
res1
tmppar<-par(no.readonly=TRUE)
par(ask=TRUE)
res2$y<- res1[ ,"SN"]
latin.factor.plot(ltable=res2,ncol=2,fun="mean",
mfrow=c(1,2),ylim=c(7,10),title="Nominal Best SN")
res2$y<-res1[ ,"Sig"]
latin.factor.plot(ltable=res2,ncol=2,fun="mean",
mfrow=c(1,2),ylim=c(18,25),title="Nominal Best Signal")
par(tmppar)
#
#Wheatstone Bridge Program Test
#
# References
#
# Taguchi, Gen-ichi et. al., 開発・設計段階の品質工学,
# "Quality Engeenering in R&D and Design", 1988,
# ISBN:4-542-51101-4
#
# Miyagawa, Masami, 品質を獲得する技術 タグチメソッドがもたらしたもの,
# "Technology for Getting Quality What the Taguchi Method Has Brought Us", 2000,
# ISBN:4-8171-0339-6
#
# Funao, Nobuo, The R Tips, 2005, ISBN:4-86167-039-X
#
# RjpWiki,http://www.okada.jp.org/RWiki/index.php?RjpWiki
#
# R Development Core Team, R: A language and environment for statistical computing (Ver. 2.21), 2005,
# ISBN 3-900051-07-0, URL http://www.R-project.org.
#
# nF RA RC RD EE RF RB XX
ilvl1<- c(0, 1, 1, 1, 1.5, 4.99, 0, 1)
ilvl0<- c(0,100, 10, 10, 5, 10, 0, 1)
ilvl2<- c(0,191, 19.1,19.1,24, 15, 0, 1)
ilvl<- matrix(c(ilvl1,ilvl0,ilvl2),nrow=8,ncol=3)
fac.names<- c("nF","RA","RC","RD","EE","RF","RB","XX")
rownames(ilvl)<- fac.names
# nF RA RC RD EE RF RB XX
olvl1<- c(0, 0.95,0.95,0.95,0.9,0.95,0.95,-0.001)
olvl0<- c(0, 1, 1, 1, 1.0,1, 1, 0)
olvl2<- c(0, 1.05,1.05,1.05,1.1,1.05,1.05, 0.001)
olvl<- matrix(c(olvl1,olvl0,olvl2),nrow=8,ncol=3)
rownames(olvl)<- fac.names
ilvl.result<- lvl.make.L18(ilvl)
olvl.result<- lvl.make.L18(olvl)
rownames(ilvl.result)<- fac.names
rownames(olvl.result)<- fac.names
ilvl.result<- wheatstoneBridge.RB(ilvl.result,RR=100)
y.result<- wheatstoneBridge.L18(ilvl.result,olvl.result)
y.sn<- taguchi.param.sn.nominal(y.result)
ltable.sn<- latin.table.L18(y.sn[,"SN"])
colnames(ltable.sn)<- c(fac.names,"SN")
ltable.sig<- latin.table.L18(y.sn[,"Sig"])
colnames(ltable.sig)<- c(fac.names,"Sig")
tmppar<-par(no.readonly=TRUE)
par(ask=TRUE)
latin.factor.plot(ltable=ltable.sn,ncol=8,mfrow=c(2,4),ylim=c(-1,19),title="W Bridge SN")
latin.factor.plot(ltable=ltable.sig,ncol=8,mfrow=c(2,4),ylim=c(38.5,40.5),title="W Bridge Signal")
par(tmppar)
summary(aov(SN ~ RA+RC+RD+EE+RF,ltable.sn))
kilvl.result<- matrix(c(0,1,10,19.1,24,4.99,0,1),ncol=1,nrow=8)
rownames(kilvl.result)<- fac.names
kilvl.result<- wheatstoneBridge.RB(kilvl.result,RR=100)
silvl.result<- matrix(ilvl0,ncol=1,nrow=8)
rownames(silvl.result)<- fac.names
silvl.result<- wheatstoneBridge.RB(silvl.result,RR=100)
ky.result<- wheatstoneBridge.L18(kilvl.result,olvl.result)
sy.result<- wheatstoneBridge.L18(silvl.result,olvl.result)
ky.sn<- taguchi.param.sn.nominal(ky.result)
sy.sn<- taguchi.param.sn.nominal(sy.result)
