R/kdclusterpatt.R
In jpison/KDSeries: Knowdledge Discovery in Time Series package
##############################################################################################
# Written by: Dr. Francisco Javier Martínez de Pisón Ascacíbar (fjmartin@dim.unirioja.es) (2005)
# EDMANS (Engineering Data Mining And Numerical Simulations) Research GROUP.
# License: GPL version 2 or newer
# --------------------------------------------------------------------------------------------
# Project Engineering Group
# Department of Mechanical Engineering. C/ Luis de Ulloa, 20.
# 26004. Universidad de La Rioja. Spain
# --------------------------------------------------------------------------------------------
# EDMANS Members=Francisco Javier Martínez de Pisón Ascacibar (fjmartin@dim.unirioja.es), Manuel Castejón Limas(manuel.castejon@unileon.es),
# ,Alpha V. Pernía Espinoza (alpha.veronica@dim.unirioja.es), Joaquín B. Ordieres Meré (joaquin.ordieres@dim.unirioja.es),
# ,Eliseo P. Vergara González (eliseo.vergara@dim.unirioja.es), Fernando Alba Elías (fernando.alba@dim.unirioja.es)
# ,Ana González Marcos (ana.gonzalez@unileon.es)
##############################################################################################
##############################################################################################
# (kdclusterpatt.R) Cluster Similar Segment from Segment Matrix (MATPatt=OBTAINED FROM kdextract.R)
##############################################################################################
# INPUT PARAMETERS:
# MATPatt=Different type of segment found (FROM kdextract() or kdmergepatt())
# INCParam= 1.- Number of Segments. Divide one Segment into frames in order to obtain shape distance
# 2.- Cluster dendrogram from shape distance by groups (g) or by height (h)
# 3.- Value of threshold dendrogram from shape distance
# 4.- Cluster dendrogram from length distance by groups (g) or by height (h)
# 5.- Value of threshold dendrogram from length distance
# DECParam= 1.- Number of Segments. Divide one Segment into frames in order to obtain shape distance
# 2.- Cluster dendrogram from shape distance by groups (g) or by height (h)
# 3.- Value of threshold dendrogram from shape distance
# 4.- Cluster dendrogram from length distance by groups (g) or by height (h)
# 5.- Value of threshold dendrogram from length distance
# HORParam= 1.- Cluster dendrogram from length distance by groups (g) or by height (h)
# 2.- Value of threshold dendrogram from length distance
# --------------------------------------------------------------------------------------------
# OUTPUT PARAMETERS:
# 1. $MATSort:Patterns Sorted
# [col1=Position, col2=Code Pattern (FamilyCode (see below))
# 2. $FamilyINC: Patterns of family INC as they appear
# 3. $FamilyDEC: Patterns of family DEC as they appear
# 4. $FamilyHOR: Patterns of family HOR as they appear
# 5. $PointsINC: Points used to cluster INC patterns
# 6. $GroupsINCShape[[X]]: INC Patterns Family X by Shape
# 7. $GroupsINCLength[[X]]: INC Patterns Family X by Length
# 8. $LengthPattINC[[X]]: INC Pattern's Length of Family X
# 9. $PointsDEC: Points used to cluster DEC patterns
# 10. $GroupsDECShape[[X]]: DEC Patterns Family X by Shape
# 11. $GroupsDECLength[[X]]: DEC Patterns Family X by Length
# 12. $LengthPattDEC[[X]]: DEC Pattern's Length of Family X
# 13. $GroupsHORLength[[X]]: HOR Patterns Family X by Length
# 14. $LengthPattHOR[[X]]: HOR Pattern's Length of Family X
# --------------------------------------------------------------------------------------------
kdclusterpatt <- function(MATPatt, INCParam=c(10,"g",4,"g",4), DECParam=c(10,"g","4","g",4), HORParam=c("g","4"))
{
SimiPatt <- NULL
###################### Obtains groups using hierarchical clustering (INC) ###########################
# mar=margins (Bot,Left,Top,Right), cex.axis=axis's text
# lab=axis y size title
# mgp=Text Axis's Distance (title, labels, line)
par(bg="white", mar=c(2,2,2,2), cex.axis=0.7, cex.lab=0.8, lab=c(10,5,6), mgp=c(1,0.5,0), xaxs="i")
NF <- layout(matrix(c(1,2,3,4),2,2,byrow=TRUE))
BorderPlot=2
# FamilyINC=0001ssssssllllll (0001=INC, s=shape type, l=length type)
FamilyINC <- rep(0,dim(MATPatt$INC)[1])
# Cluster INCREMENTAL Segments
MATPos <- MATPatt$INC[,1]
MATLen <- MATPatt$INC[,2]
PointsINC <- NULL
for (i in 1:as.numeric(INCParam[1]))
{
PointsINC <- cbind(PointsINC, MATPatt$TSerie[MATPos+MATLen*(i/as.numeric(INCParam[1]))]-MATPatt$TSerie[MATPos])
}
# Cluster by Shape
Hc <- hclust(dist(PointsINC))
plot(Hc,hang=-1,main="Cluster Dendrogram INC Patterns by SHAPE",xlab="INCPatterns",labels=FALSE)
if (INCParam[2]=="h")
{
GroupsINCShape <- rect.hclust(Hc, h=as.numeric(INCParam[3]), border=1:30)
}
else
{
GroupsINCShape <- rect.hclust(Hc, k=as.numeric(INCParam[3]), border=1:as.numeric(INCParam[3]))
}
# Cluster by Length
Hc <- hclust(dist(MATLen))
plot(Hc,hang=-1,main="Cluster Dendrogram INC Patterns by LENGHT",xlab="INCPatterns",labels=FALSE)
if (INCParam[4]=="h")
{
GroupsINCLength <- rect.hclust(Hc, h=as.numeric(INCParam[5]), border=1:30)
}
else
{
GroupsINCLength <- rect.hclust(Hc, k=as.numeric(INCParam[5]), border=1:as.numeric(INCParam[5]))
}
# Plots Cluster INCSegments by SHAPE
ColorsPat <- rep(0,dim(PointsINC)[1])
for (i in 1:length(GroupsINCShape))
{
ColorsPat[GroupsINCShape[[i]]] <- i
# FamilyINC=0001ssssssllllll (0001=INC, s=shape type, l=length type)
FamilyINC[GroupsINCShape[[i]]] <- 4096+i*64
}
plot(PointsINC[1,],type="l",ylim=range(PointsINC),col=ColorsPat[1],main="INC Clusters by Shape")
for (i in 2:(dim(PointsINC)[1]))
{
lines(PointsINC[i,],col=ColorsPat[i])
}
# Plots Cluster INCSegmentsby LENGTH
TablePat <- rep(0, length(GroupsINCLength))
LengthMax <- TablePat
LengthPattINC <- as.list(NULL)
for (i in 1:length(GroupsINCLength))
{
# FamilyINC=0001ssssssllllll (0001=INC, s=shape type, l=length type)
FamilyINC[GroupsINCLength[[i]]] <- FamilyINC[GroupsINCLength[[i]]]+i
TablePat[i] <- length(GroupsINCLength[[i]])
LengthMax[i] <- max(MATLen[GroupsINCLength[[i]]])
LengthPattINC <- c(LengthPattINC, list(MATLen[GroupsINCLength[[i]]]))
}
boxplot(LengthPattINC,col=1:length(TablePat),ylim=c(0,max(LengthMax)+5),main="Boxplot INC Length Cluster",xlab="Cluster",ylab="Length")
text(1:length(TablePat),LengthMax+2,TablePat)
########################################## END INC SEGMENTS ##########################################
###################### Obtains groups using hierarchical clustering (DEC) ###########################
# mar=margins (Bot,Left,Top,Right), cex.axis=axis's text
# lab=axis y size title
# mgp=Text Axis's Distance (title, labels, line)
X11()
par(bg="white", mar=c(2,2,2,2), cex.axis=0.7, cex.lab=0.8, lab=c(10,5,6), mgp=c(1,0.5,0), xaxs="i")
NF <- layout(matrix(c(1,2,3,4),2,2,byrow=TRUE))
BorderPlot=2
# FamilyDEC=0002ssssssllllll (0002=DEC, s=shape type, l=length type)
FamilyDEC <- rep(0,dim(MATPatt$DEC)[1])
# Cluster DECREMENTAL Patterns
MATPos <- MATPatt$DEC[,1]
MATLen <- MATPatt$DEC[,2]
PointsDEC <- NULL
for (i in 1:as.numeric(DECParam[1]))
{
PointsDEC <- cbind(PointsDEC, MATPatt$TSerie[MATPos+MATLen*(i/as.numeric(DECParam[1]))]-MATPatt$TSerie[MATPos])
}
# Cluster by Shape
Hc <- hclust(dist(PointsDEC))
plot(Hc,hang=-1,main="Cluster Dendrogram DEC Patterns by SHAPE",xlab="DECPatterns",labels=FALSE)
if (DECParam[2]=="h")
{
GroupsDECShape <- rect.hclust(Hc, h=as.numeric(DECParam[3]), border=1:30)
}
else
{
GroupsDECShape <- rect.hclust(Hc, k=as.numeric(DECParam[3]), border=1:as.numeric(DECParam[3]))
}
# Cluster by Length
Hc <- hclust(dist(MATLen))
plot(Hc,hang=-1,main="Cluster Dendrogram DEC Patterns by LENGHT",xlab="DECPatterns",labels=FALSE)
if (DECParam[4]=="h")
{
GroupsDECLength <- rect.hclust(Hc, h=as.numeric(DECParam[5]), border=1:30)
}
else
{
GroupsDECLength <- rect.hclust(Hc, k=as.numeric(DECParam[5]), border=1:as.numeric(DECParam[5]))
}
# Plots Cluster DECPatterns by SHAPE
ColorsPat <- rep(0,dim(PointsDEC)[1])
for (i in 1:length(GroupsDECShape))
{
# FamilyDEC=0002ssssssllllll (0002=DEC, s=shape type, l=length type)
FamilyDEC[GroupsDECShape[[i]]] <- 8192+i*64
ColorsPat[GroupsDECShape[[i]]] <- i
}
plot(PointsDEC[1,],type="l",ylim=range(PointsDEC),col=ColorsPat[1],main="DEC Clusters by Shape")
for (i in 2:(dim(PointsDEC)[1]))
{
lines(PointsDEC[i,],col=ColorsPat[i])
}
# Plots Cluster DECPatterns by LENGTH
TablePat <- rep(0, length(GroupsDECLength))
LengthMax <- TablePat
LengthPattDEC <- as.list(NULL)
for (i in 1:length(GroupsDECLength))
{
# FamilyDEC=0002ssssssllllll (0002=DEC, s=shape type, l=length type)
FamilyDEC[GroupsDECLength[[i]]] <- FamilyDEC[GroupsDECLength[[i]]]+i
TablePat[i] <- length(GroupsDECLength[[i]])
LengthMax[i] <- max(MATLen[GroupsDECLength[[i]]])
LengthPattDEC <- c(LengthPattDEC, list(MATLen[GroupsDECLength[[i]]]))
}
boxplot(LengthPattDEC,col=1:length(TablePat),ylim=c(0,max(LengthMax)+5),main="Boxplot DEC Length Cluster",xlab="Cluster",ylab="Length")
text(1:length(TablePat),LengthMax+2,TablePat)
########################################## END DEC PATTERNS ##########################################
###################### Obtains groups using hierarchical clustering (HOR) ###########################
# mar=margins (Bot,Left,Top,Right), cex.axis=axis's text
# lab=axis y size title
# mgp=Text Axis's Distance (title, labels, line)
X11()
par(bg="white", mar=c(2,2,2,2), cex.axis=0.7, cex.lab=0.8, lab=c(10,5,6), mgp=c(1,0.5,0), xaxs="i")
NF <- layout(matrix(c(1,2),2,1,byrow=TRUE))
BorderPlot=2
# FamilyHOR=0003000000llllll (0003=HOR, s=shape type, l=length type)
FamilyHOR <- rep(0,dim(MATPatt$HOR)[1])
# Cluster HORIZONTAL Patterns
MATLen <- MATPatt$HOR[,2]
# Cluster by Length
Hc <- hclust(dist(MATLen))
plot(Hc,hang=-1,main="Cluster Dendrogram HOR Patterns by LENGHT",xlab="HORPatterns",labels=FALSE)
if (HORParam[1]=="h")
{
GroupsHORLength <- rect.hclust(Hc, h=as.numeric(HORParam[2]), border=1:30)
}
else
{
GroupsHORLength <- rect.hclust(Hc, k=as.numeric(HORParam[2]), border=1:as.numeric(HORParam[2]))
}
# Plots Cluster HORPatterns by LENGTH
TablePat <- rep(0, length(GroupsHORLength))
LengthMax <- TablePat
LengthPattHOR <- as.list(NULL)
for (i in 1:length(GroupsHORLength))
{
# FamilyHOR=0003000000llllll (0003=HOR, s=shape type, l=length type)
FamilyHOR[GroupsHORLength[[i]]] <- 12288+i
TablePat[i] <- length(GroupsHORLength[[i]])
LengthMax[i] <- max(MATLen[GroupsHORLength[[i]]])
LengthPattHOR <- c(LengthPattHOR, list(MATLen[GroupsHORLength[[i]]]))
}
boxplot(LengthPattHOR,col=1:length(TablePat),ylim=c(0,max(LengthMax)+5),main="Boxplot HOR Length Cluster",xlab="Cluster",ylab="Length")
text(1:length(TablePat),LengthMax+2,TablePat)
########################################## END DEC PATTERNS ##########################################
# Creates serie using nominal patterns
MATSort <- rbind(cbind(Pos=MATPatt$INC[,1],Code=FamilyINC),cbind(Pos=MATPatt$DEC[,1],Code=FamilyDEC),cbind(Pos=MATPatt$HOR[,1],Code=FamilyHOR))
MATSort <- MATSort[order(MATSort[,1]),]
SimiPatt <- list(MATSort=MATSort, FamilyINC=FamilyINC, FamilyHOR=FamilyHOR, FamilyDEC=FamilyDEC, PointsINC=PointsINC, GroupsINCShape=GroupsINCShape, GroupsINCLength=GroupsINCLength, LengthPattINC=LengthPattINC, PointsDEC=PointsDEC, GroupsDECShape=GroupsDECShape, GroupsDECLength=GroupsDECLength, LengthPattDEC=LengthPattDEC, GroupsHORLength=GroupsHORLength, LengthPattHOR=LengthPattHOR)
return(SimiPatt)
}
jpison/KDSeries documentation built on May 10, 2019, 12:09 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
##############################################################################################
# Written by: Dr. Francisco Javier Martínez de Pisón Ascacíbar (fjmartin@dim.unirioja.es) (2005)
# EDMANS (Engineering Data Mining And Numerical Simulations) Research GROUP.
# License: GPL version 2 or newer
# --------------------------------------------------------------------------------------------
# Project Engineering Group
# Department of Mechanical Engineering. C/ Luis de Ulloa, 20.
# 26004. Universidad de La Rioja. Spain
# --------------------------------------------------------------------------------------------
# EDMANS Members=Francisco Javier Martínez de Pisón Ascacibar (fjmartin@dim.unirioja.es), Manuel Castejón Limas(manuel.castejon@unileon.es),
# ,Alpha V. Pernía Espinoza (alpha.veronica@dim.unirioja.es), Joaquín B. Ordieres Meré (joaquin.ordieres@dim.unirioja.es),
# ,Eliseo P. Vergara González (eliseo.vergara@dim.unirioja.es), Fernando Alba Elías (fernando.alba@dim.unirioja.es)
# ,Ana González Marcos (ana.gonzalez@unileon.es)
##############################################################################################
##############################################################################################
# (kdclusterpatt.R) Cluster Similar Segment from Segment Matrix (MATPatt=OBTAINED FROM kdextract.R)
##############################################################################################
# INPUT PARAMETERS:
# MATPatt=Different type of segment found (FROM kdextract() or kdmergepatt())
# INCParam= 1.- Number of Segments. Divide one Segment into frames in order to obtain shape distance
# 2.- Cluster dendrogram from shape distance by groups (g) or by height (h)
# 3.- Value of threshold dendrogram from shape distance
# 4.- Cluster dendrogram from length distance by groups (g) or by height (h)
# 5.- Value of threshold dendrogram from length distance
# DECParam= 1.- Number of Segments. Divide one Segment into frames in order to obtain shape distance
# 2.- Cluster dendrogram from shape distance by groups (g) or by height (h)
# 3.- Value of threshold dendrogram from shape distance
# 4.- Cluster dendrogram from length distance by groups (g) or by height (h)
# 5.- Value of threshold dendrogram from length distance
# HORParam= 1.- Cluster dendrogram from length distance by groups (g) or by height (h)
# 2.- Value of threshold dendrogram from length distance
# --------------------------------------------------------------------------------------------
# OUTPUT PARAMETERS:
# 1. $MATSort:Patterns Sorted
# [col1=Position, col2=Code Pattern (FamilyCode (see below))
# 2. $FamilyINC: Patterns of family INC as they appear
# 3. $FamilyDEC: Patterns of family DEC as they appear
# 4. $FamilyHOR: Patterns of family HOR as they appear
# 5. $PointsINC: Points used to cluster INC patterns
# 6. $GroupsINCShape[[X]]: INC Patterns Family X by Shape
# 7. $GroupsINCLength[[X]]: INC Patterns Family X by Length
# 8. $LengthPattINC[[X]]: INC Pattern's Length of Family X
# 9. $PointsDEC: Points used to cluster DEC patterns
# 10. $GroupsDECShape[[X]]: DEC Patterns Family X by Shape
# 11. $GroupsDECLength[[X]]: DEC Patterns Family X by Length
# 12. $LengthPattDEC[[X]]: DEC Pattern's Length of Family X
# 13. $GroupsHORLength[[X]]: HOR Patterns Family X by Length
# 14. $LengthPattHOR[[X]]: HOR Pattern's Length of Family X
# --------------------------------------------------------------------------------------------
kdclusterpatt <- function(MATPatt, INCParam=c(10,"g",4,"g",4), DECParam=c(10,"g","4","g",4), HORParam=c("g","4"))
{
SimiPatt <- NULL
###################### Obtains groups using hierarchical clustering (INC) ###########################
# mar=margins (Bot,Left,Top,Right), cex.axis=axis's text
# lab=axis y size title
# mgp=Text Axis's Distance (title, labels, line)
par(bg="white", mar=c(2,2,2,2), cex.axis=0.7, cex.lab=0.8, lab=c(10,5,6), mgp=c(1,0.5,0), xaxs="i")
NF <- layout(matrix(c(1,2,3,4),2,2,byrow=TRUE))
BorderPlot=2
# FamilyINC=0001ssssssllllll (0001=INC, s=shape type, l=length type)
FamilyINC <- rep(0,dim(MATPatt$INC)[1])
# Cluster INCREMENTAL Segments
MATPos <- MATPatt$INC[,1]
MATLen <- MATPatt$INC[,2]
PointsINC <- NULL
for (i in 1:as.numeric(INCParam[1]))
{
PointsINC <- cbind(PointsINC, MATPatt$TSerie[MATPos+MATLen*(i/as.numeric(INCParam[1]))]-MATPatt$TSerie[MATPos])
}
# Cluster by Shape
Hc <- hclust(dist(PointsINC))
plot(Hc,hang=-1,main="Cluster Dendrogram INC Patterns by SHAPE",xlab="INCPatterns",labels=FALSE)
if (INCParam[2]=="h")
{
GroupsINCShape <- rect.hclust(Hc, h=as.numeric(INCParam[3]), border=1:30)
}
else
{
GroupsINCShape <- rect.hclust(Hc, k=as.numeric(INCParam[3]), border=1:as.numeric(INCParam[3]))
}
# Cluster by Length
Hc <- hclust(dist(MATLen))
plot(Hc,hang=-1,main="Cluster Dendrogram INC Patterns by LENGHT",xlab="INCPatterns",labels=FALSE)
if (INCParam[4]=="h")
{
GroupsINCLength <- rect.hclust(Hc, h=as.numeric(INCParam[5]), border=1:30)
}
else
{
GroupsINCLength <- rect.hclust(Hc, k=as.numeric(INCParam[5]), border=1:as.numeric(INCParam[5]))
}
# Plots Cluster INCSegments by SHAPE
ColorsPat <- rep(0,dim(PointsINC)[1])
for (i in 1:length(GroupsINCShape))
{
ColorsPat[GroupsINCShape[[i]]] <- i
# FamilyINC=0001ssssssllllll (0001=INC, s=shape type, l=length type)
FamilyINC[GroupsINCShape[[i]]] <- 4096+i*64
}
plot(PointsINC[1,],type="l",ylim=range(PointsINC),col=ColorsPat[1],main="INC Clusters by Shape")
for (i in 2:(dim(PointsINC)[1]))
{
lines(PointsINC[i,],col=ColorsPat[i])
}
# Plots Cluster INCSegmentsby LENGTH
TablePat <- rep(0, length(GroupsINCLength))
LengthMax <- TablePat
LengthPattINC <- as.list(NULL)
for (i in 1:length(GroupsINCLength))
{
# FamilyINC=0001ssssssllllll (0001=INC, s=shape type, l=length type)
FamilyINC[GroupsINCLength[[i]]] <- FamilyINC[GroupsINCLength[[i]]]+i
TablePat[i] <- length(GroupsINCLength[[i]])
LengthMax[i] <- max(MATLen[GroupsINCLength[[i]]])
LengthPattINC <- c(LengthPattINC, list(MATLen[GroupsINCLength[[i]]]))
}
boxplot(LengthPattINC,col=1:length(TablePat),ylim=c(0,max(LengthMax)+5),main="Boxplot INC Length Cluster",xlab="Cluster",ylab="Length")
text(1:length(TablePat),LengthMax+2,TablePat)
########################################## END INC SEGMENTS ##########################################
###################### Obtains groups using hierarchical clustering (DEC) ###########################
# mar=margins (Bot,Left,Top,Right), cex.axis=axis's text
# lab=axis y size title
# mgp=Text Axis's Distance (title, labels, line)
X11()
par(bg="white", mar=c(2,2,2,2), cex.axis=0.7, cex.lab=0.8, lab=c(10,5,6), mgp=c(1,0.5,0), xaxs="i")
NF <- layout(matrix(c(1,2,3,4),2,2,byrow=TRUE))
BorderPlot=2
# FamilyDEC=0002ssssssllllll (0002=DEC, s=shape type, l=length type)
FamilyDEC <- rep(0,dim(MATPatt$DEC)[1])
# Cluster DECREMENTAL Patterns
MATPos <- MATPatt$DEC[,1]
MATLen <- MATPatt$DEC[,2]
PointsDEC <- NULL
for (i in 1:as.numeric(DECParam[1]))
{
PointsDEC <- cbind(PointsDEC, MATPatt$TSerie[MATPos+MATLen*(i/as.numeric(DECParam[1]))]-MATPatt$TSerie[MATPos])
}
# Cluster by Shape
Hc <- hclust(dist(PointsDEC))
plot(Hc,hang=-1,main="Cluster Dendrogram DEC Patterns by SHAPE",xlab="DECPatterns",labels=FALSE)
if (DECParam[2]=="h")
{
GroupsDECShape <- rect.hclust(Hc, h=as.numeric(DECParam[3]), border=1:30)
}
else
{
GroupsDECShape <- rect.hclust(Hc, k=as.numeric(DECParam[3]), border=1:as.numeric(DECParam[3]))
}
# Cluster by Length
Hc <- hclust(dist(MATLen))
plot(Hc,hang=-1,main="Cluster Dendrogram DEC Patterns by LENGHT",xlab="DECPatterns",labels=FALSE)
if (DECParam[4]=="h")
{
GroupsDECLength <- rect.hclust(Hc, h=as.numeric(DECParam[5]), border=1:30)
}
else
{
GroupsDECLength <- rect.hclust(Hc, k=as.numeric(DECParam[5]), border=1:as.numeric(DECParam[5]))
}
# Plots Cluster DECPatterns by SHAPE
ColorsPat <- rep(0,dim(PointsDEC)[1])
for (i in 1:length(GroupsDECShape))
{
# FamilyDEC=0002ssssssllllll (0002=DEC, s=shape type, l=length type)
FamilyDEC[GroupsDECShape[[i]]] <- 8192+i*64
ColorsPat[GroupsDECShape[[i]]] <- i
}
plot(PointsDEC[1,],type="l",ylim=range(PointsDEC),col=ColorsPat[1],main="DEC Clusters by Shape")
for (i in 2:(dim(PointsDEC)[1]))
{
lines(PointsDEC[i,],col=ColorsPat[i])
}
# Plots Cluster DECPatterns by LENGTH
TablePat <- rep(0, length(GroupsDECLength))
LengthMax <- TablePat
LengthPattDEC <- as.list(NULL)
for (i in 1:length(GroupsDECLength))
{
# FamilyDEC=0002ssssssllllll (0002=DEC, s=shape type, l=length type)
FamilyDEC[GroupsDECLength[[i]]] <- FamilyDEC[GroupsDECLength[[i]]]+i
TablePat[i] <- length(GroupsDECLength[[i]])
LengthMax[i] <- max(MATLen[GroupsDECLength[[i]]])
LengthPattDEC <- c(LengthPattDEC, list(MATLen[GroupsDECLength[[i]]]))
}
boxplot(LengthPattDEC,col=1:length(TablePat),ylim=c(0,max(LengthMax)+5),main="Boxplot DEC Length Cluster",xlab="Cluster",ylab="Length")
text(1:length(TablePat),LengthMax+2,TablePat)
########################################## END DEC PATTERNS ##########################################
###################### Obtains groups using hierarchical clustering (HOR) ###########################
# mar=margins (Bot,Left,Top,Right), cex.axis=axis's text
# lab=axis y size title
# mgp=Text Axis's Distance (title, labels, line)
X11()
par(bg="white", mar=c(2,2,2,2), cex.axis=0.7, cex.lab=0.8, lab=c(10,5,6), mgp=c(1,0.5,0), xaxs="i")
NF <- layout(matrix(c(1,2),2,1,byrow=TRUE))
BorderPlot=2
# FamilyHOR=0003000000llllll (0003=HOR, s=shape type, l=length type)
FamilyHOR <- rep(0,dim(MATPatt$HOR)[1])
# Cluster HORIZONTAL Patterns
MATLen <- MATPatt$HOR[,2]
# Cluster by Length
Hc <- hclust(dist(MATLen))
plot(Hc,hang=-1,main="Cluster Dendrogram HOR Patterns by LENGHT",xlab="HORPatterns",labels=FALSE)
if (HORParam[1]=="h")
{
GroupsHORLength <- rect.hclust(Hc, h=as.numeric(HORParam[2]), border=1:30)
}
else
{
GroupsHORLength <- rect.hclust(Hc, k=as.numeric(HORParam[2]), border=1:as.numeric(HORParam[2]))
}
# Plots Cluster HORPatterns by LENGTH
TablePat <- rep(0, length(GroupsHORLength))
LengthMax <- TablePat
LengthPattHOR <- as.list(NULL)
for (i in 1:length(GroupsHORLength))
{
# FamilyHOR=0003000000llllll (0003=HOR, s=shape type, l=length type)
FamilyHOR[GroupsHORLength[[i]]] <- 12288+i
TablePat[i] <- length(GroupsHORLength[[i]])
LengthMax[i] <- max(MATLen[GroupsHORLength[[i]]])
LengthPattHOR <- c(LengthPattHOR, list(MATLen[GroupsHORLength[[i]]]))
}
boxplot(LengthPattHOR,col=1:length(TablePat),ylim=c(0,max(LengthMax)+5),main="Boxplot HOR Length Cluster",xlab="Cluster",ylab="Length")
text(1:length(TablePat),LengthMax+2,TablePat)
########################################## END DEC PATTERNS ##########################################
# Creates serie using nominal patterns
MATSort <- rbind(cbind(Pos=MATPatt$INC[,1],Code=FamilyINC),cbind(Pos=MATPatt$DEC[,1],Code=FamilyDEC),cbind(Pos=MATPatt$HOR[,1],Code=FamilyHOR))
MATSort <- MATSort[order(MATSort[,1]),]
SimiPatt <- list(MATSort=MATSort, FamilyINC=FamilyINC, FamilyHOR=FamilyHOR, FamilyDEC=FamilyDEC, PointsINC=PointsINC, GroupsINCShape=GroupsINCShape, GroupsINCLength=GroupsINCLength, LengthPattINC=LengthPattINC, PointsDEC=PointsDEC, GroupsDECShape=GroupsDECShape, GroupsDECLength=GroupsDECLength, LengthPattDEC=LengthPattDEC, GroupsHORLength=GroupsHORLength, LengthPattHOR=LengthPattHOR)
return(SimiPatt)
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.