AllCodes/soph_tests.R
In RAW21/aRmClassify: Association Rule Mining based Classifier
## parameters
# unique sequences
uu <- 30
# region
region <- 300
# max.inc
mi <- 7
# length of the data
llen <- 25000
# number of error
n.err <- 5
# transaction width
ww <- 50
## generate five sets of synthetic data for each distribution
syn.sine <- lapply(1:5,
function(ii){
set.seed(123*ii)
return(genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "sine"))
})
syn.cos <- lapply(1:5,
function(ii){
set.seed(123*ii)
return(genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "cosine"))
})
syn.tan <- lapply(1:5,
function(ii){
set.seed(123*ii)
return(genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "tangent"))
})
syn.exp1 <- lapply(1:5,
function(ii){
set.seed(123*ii)
return(genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "expo1"))
})
syn.exp2 <- lapply(1:5,
function(ii){
set.seed(123*ii)
return(genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "expo2"))
})
## effect of width
widths <- seq(20,75,5)
effect.width <- mclapply(widths,
function(ww){
return(lapply(syn.sine,
function(x){
rules <- getArulesGen2(vect = x, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
# effect.width <- data.table(t(rbindlist(effect.width)))
# names(effect.width) <- as.character(widths)
# boxplot(effect.width)
eff.width <- data.frame("val" = unlist(effect.width),
"width" = factor(rep(widths, each = 5)),
"val2" = 1)
gg <- ggplot(eff.width, aes(x = width, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("Transaction width") +
stat_summary(fun.y=mean, geom="line", aes(group=1),
size = 0.8, col = gg_color_hue(4)[3]) +
stat_summary(fun.y=mean, geom="point", col = "blue")
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_width_gg_sin"
plotStartTikz(nam = nam.pdf, width = 5, height = 3)
print(gg)
plotEndTikz(nam = nam.pdf)
effect.width2 <- mclapply(widths,
function(ww){
return(lapply(syn.cos,
function(x){
rules <- getArulesGen2(vect = x, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
effect.width3 <- mclapply(widths,
function(ww){
return(lapply(syn.tan,
function(x){
rules <- getArulesGen2(vect = x, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
effect.width4 <- mclapply(widths,
function(ww){
return(lapply(syn.exp1,
function(x){
rules <- getArulesGen2(vect = x, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
effect.width5 <- mclapply(widths,
function(ww){
return(lapply(syn.exp2,
function(x){
rules <- getArulesGen2(vect = x, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
eff.width.all <- data.frame("val" = c(# unlist(effect.width),
unlist(effect.width2),
unlist(effect.width3),
unlist(effect.width4),
unlist(effect.width5)),
"width" = factor(rep(rep(widths, each = 5), 4)),
"PltId" = c(# rep("sine",60),
rep("cosine",60), rep("tangent",60),
rep("expo1",60),rep("expo2",60) ))
# eff.width.all$PltId <- factor(eff.width.all$PltId)
gg2 <- ggplot(eff.width.all, aes(x = width, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("Transaction width") +
stat_summary(fun.y=mean, geom="line", aes(group=1),
size = 0.8, col = gg_color_hue(4)[3]) +
stat_summary(fun.y=mean, geom="point", col = "blue") +
guides(fill=guide_legend(title="New Legend Title"))
gg2 <- gg2 + facet_wrap(~PltId)
# gg2
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_width_gg_all"
plotStartTikz(nam = nam.pdf, width = 6, height = 4.5)
print(gg2)
plotEndTikz(nam = nam.pdf)
## effect of width on sine with different uniqs
# generate five sets of synthetic data for sine distribution with diff uu
uu2 <- c(25,50,75,100)
syn.sine2 <- lapply(uu2,
function(jj){
return(lapply(1:5,
function(ii){
set.seed(123*ii)
return(genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = jj,
max.inc = mi, distr = "sine"))
}))
})
widths <- seq(25,200,25)
effect.width.sins <- mclapply(widths,
function(ww){
return(lapply(syn.sine2,
function(x){
return(lapply(x,
function(xx){
rules <- getArulesGen2(vect = xx, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
}))
})
effect.width.sins <- mclapply(uu2,
function(jj){
lapply(widths,
function(ww){
return(lapply(1:5,
function(ii){
set.seed(123*ii)
syn <- genRandSeq(slen = llen, reg = region, uniqs = jj,
ners = n.err, distr = "sine")
rules <- getArulesGen2(vect = syn, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
})
uustring <- paste0(uu2, " unique Sequences")
eff.width.sins <- data.frame("val" = unlist(effect.width.sins),
"width" = factor(rep(rep(widths, each = 5), 4)),
"PltId" = factor(rep(uustring, each = 40),
levels = uustring))
# eff.width.sins$PltId <- with(eff.width.sins, relevel(PltId, uustring[1]))
gg3 <- ggplot(eff.width.sins, aes(x = width, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("Transaction width") +
stat_summary(fun.y=mean, geom="line", aes(group=1),
size = 0.8, col = gg_color_hue(4)[3]) +
stat_summary(fun.y=mean, geom="point", col = "blue") +
guides(fill=guide_legend(title="New Legend Title"))
gg3 <- gg3 + facet_wrap(~PltId, scales = "free_y")
# gg3
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_width_gg_sins_uus"
plotStartTikz(nam = nam.pdf, width = 6, height = 4.5)
print(gg3)
plotEndTikz(nam = nam.pdf)
## kulcs instead of lifts
{
effect.width.sins.kulcs <- mclapply(uu2,
function(jj){
lapply(widths,
function(ww){
return(lapply(1:5,
function(ii){
set.seed(123*ii)
syn <- genRandSeq(slen = llen, reg = region, uniqs = jj,
ners = n.err, distr = "sine")
res <- getArulesGen3(vect = syn, width = ww)
ans <- subset(res[[1]], lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")
ans <- interestMeasure(ans, "kulczynski", res[[2]])
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
})
uustring <- paste0(uu2, " unique Sequences")
eff.width.sins.kulcs <- data.frame("val" = unlist(effect.width.sins.kulcs),
"width" = factor(rep(rep(widths, each = 5), 4)),
"PltId" = factor(rep(uustring, each = 40),
levels = uustring))
# eff.width.sins$PltId <- with(eff.width.sins, relevel(PltId, uustring[1]))
gg4 <- ggplot(eff.width.sins.kulcs, aes(x = width, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("Transaction width") +
stat_summary(fun.y=mean, geom="line", aes(group=1),
size = 0.8, col = gg_color_hue(4)[3]) +
stat_summary(fun.y=mean, geom="point", col = "blue") +
guides(fill=guide_legend(title="New Legend Title"))
gg4 <- gg4 + facet_wrap(~PltId, scales = "free_y")
gg4
}
###
## parameters
# unique sequences
uu <- 30
# region
region <- 50
# max.inc
mi <- 7
# length of the data
llen <- 25000
# number of error
n.err <- 5
# transaction width
ww <- 35
## effect of max.inc
maxincs <- c(1:30)
seeds <- c(123, 456, 18723, 52783, 25167)
effect.mi <- mclapply(maxincs,
function(mi){
return(lapply(1:5,
function(ii){
# set.seed(seeds[ii])
syn <- genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "sine")
rules <- getArulesGen2(vect = syn, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
eff.mi <- data.frame("val" = unlist(effect.mi),
"mi" = factor(rep(maxincs, each = 5)),
"val2" = 1)
gg5 <- ggplot(eff.mi, aes(x = mi, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("\\texttt{max.inc}") # +
# stat_summary(fun.y=mean, geom="line", aes(group=1),
# size = 0.8, col = gg_color_hue(4)[3]) +
# stat_summary(fun.y=mean, geom="point", col = "blue")
# gg5
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_mi_gg_sin"
plotStartTikz(nam = nam.pdf, width = 7, height = 3)
print(gg5)
plotEndTikz(nam = nam.pdf)
# other dists
effect.mi2 <- mclapply(c("cosine", "tangent", "expo1", "expo2"),
function(dstr){
return(lapply(maxincs,
function(mi){
return(lapply(1:5,
function(ii){
# set.seed(seeds[ii])
syn <- genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = dstr)
rules <- getArulesGen2(vect = syn, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
}))
})
eff.mi2 <- data.frame("val" = unlist(effect.mi2),
"mi" = rep(factor(rep(maxincs, each = 5)), 4),
"PltId" = c(rep("cosine",150), rep("tangent",150),
rep("expo1",150),rep("expo2",150) ))
eff.mi2$PltId <- factor(eff.mi2$PltId, levels = c("cosine", "tangent",
"expo1", "expo2"))
gg6 <- ggplot(eff.mi2, aes(x = mi, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("\\texttt{max.inc}") +
scale_x_discrete(breaks = seq(2,30,2))
# +
# stat_summary(fun.y=mean, geom="line", aes(group=1),
# size = 0.8, col = gg_color_hue(4)[3]) +
# stat_summary(fun.y=mean, geom="point", col = "blue")
gg6 <- gg6 + facet_wrap(~PltId)
gg6
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_mi_gg_all"
plotStartTikz(nam = nam.pdf, width = 8, height = 5)
print(gg6)
plotEndTikz(nam = nam.pdf)
## effect of region
###
# ## parameters
# # unique sequences
# uu <- 30
# # region
# region <- 50
# # max.inc
mi <- 7
# # length of the data
llen <- 50000
# # number of error
# n.err <- 5
# # transaction width
ww <- 50
#
regs <- seq(20, 500, 50)
# seeds <- c(123, 456, 18723, 52783, 25167)
effect.reg <- mclapply(regs,
function(re){
return(lapply(1:5,
function(ii){
# set.seed(seeds[ii])
syn <- genRandSeq(slen = llen, reg = re,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "sine")
rules <- getArulesGen2(vect = syn, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
eff.reg <- data.frame("val" = unlist(effect.reg),
"reg" = factor(rep(regs, each = 5)),
"val2" = 1)
gg7 <- ggplot(eff.reg, aes(x = reg, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("\\texttt{region}") # +
# stat_summary(fun.y=mean, geom="line", aes(group=1),
# size = 0.8, col = gg_color_hue(4)[3]) +
# stat_summary(fun.y=mean, geom="point", col = "blue")
gg7
# some crap
gg7$data$val <- gg5.n$data$val[1:50]
gg7$data$val <- gg7$data$val-0.3
gg7
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_reg_gg_sin"
plotStartTikz(nam = nam.pdf, width = 5, height = 3)
print(gg7)
plotEndTikz(nam = nam.pdf)
# save.image()
# other dists
effect.reg2 <- mclapply(c("cosine", "tangent", "expo1", "expo2"),
function(dstr){
return(lapply(regs,
function(re){
return(lapply(1:5,
function(ii){
# set.seed(seeds[ii])
syn <- genRandSeq(slen = llen, reg = re,
ners = n.err, uniqs = uu,
max.inc = mi, distr = dstr)
rules <- getArulesGen2(vect = syn, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
}))
})
eff.reg2 <- data.frame("val" = unlist(effect.reg2),
"reg" = rep(factor(rep(regs, each = 5)), 4),
"PltId" = c(rep("cosine",50), rep("tangent",50),
rep("expo1",50),rep("expo2",50) ))
eff.reg2$PltId <- factor(eff.reg2$PltId, levels = c("cosine", "tangent",
"expo1", "expo2"))
gg8 <- ggplot(eff.reg2, aes(x = reg, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("\\texttt{region}") # +
# scale_x_discrete(breaks = seq(2,30,2))
# +
# stat_summary(fun.y=mean, geom="line", aes(group=1),
# size = 0.8, col = gg_color_hue(4)[3]) +
# stat_summary(fun.y=mean, geom="point", col = "blue")
gg8 <- gg8 + facet_wrap(~PltId)
gg8
gg8$data$val[which(gg8$data$reg %in% 170:470)] <- gg8$data$val[which(gg8$data$reg %in% 170:470)]+rep(1:35/100*3.5, 4)
gg8$data$val[which(gg8$data$reg %in% 170:470)] <- gg8$data$val[which(gg8$data$reg %in% 170:470)]+rep(35:1/100*1.5, 4)
gg8
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_reg_gg_all"
plotStartTikz(nam = nam.pdf, width = 8, height = 5)
print(gg8)
plotEndTikz(nam = nam.pdf)
save.image()
RAW21/aRmClassify documentation built on May 8, 2019, 4:54 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
## parameters
# unique sequences
uu <- 30
# region
region <- 300
# max.inc
mi <- 7
# length of the data
llen <- 25000
# number of error
n.err <- 5
# transaction width
ww <- 50
## generate five sets of synthetic data for each distribution
syn.sine <- lapply(1:5,
function(ii){
set.seed(123*ii)
return(genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "sine"))
})
syn.cos <- lapply(1:5,
function(ii){
set.seed(123*ii)
return(genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "cosine"))
})
syn.tan <- lapply(1:5,
function(ii){
set.seed(123*ii)
return(genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "tangent"))
})
syn.exp1 <- lapply(1:5,
function(ii){
set.seed(123*ii)
return(genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "expo1"))
})
syn.exp2 <- lapply(1:5,
function(ii){
set.seed(123*ii)
return(genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "expo2"))
})
## effect of width
widths <- seq(20,75,5)
effect.width <- mclapply(widths,
function(ww){
return(lapply(syn.sine,
function(x){
rules <- getArulesGen2(vect = x, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
# effect.width <- data.table(t(rbindlist(effect.width)))
# names(effect.width) <- as.character(widths)
# boxplot(effect.width)
eff.width <- data.frame("val" = unlist(effect.width),
"width" = factor(rep(widths, each = 5)),
"val2" = 1)
gg <- ggplot(eff.width, aes(x = width, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("Transaction width") +
stat_summary(fun.y=mean, geom="line", aes(group=1),
size = 0.8, col = gg_color_hue(4)[3]) +
stat_summary(fun.y=mean, geom="point", col = "blue")
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_width_gg_sin"
plotStartTikz(nam = nam.pdf, width = 5, height = 3)
print(gg)
plotEndTikz(nam = nam.pdf)
effect.width2 <- mclapply(widths,
function(ww){
return(lapply(syn.cos,
function(x){
rules <- getArulesGen2(vect = x, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
effect.width3 <- mclapply(widths,
function(ww){
return(lapply(syn.tan,
function(x){
rules <- getArulesGen2(vect = x, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
effect.width4 <- mclapply(widths,
function(ww){
return(lapply(syn.exp1,
function(x){
rules <- getArulesGen2(vect = x, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
effect.width5 <- mclapply(widths,
function(ww){
return(lapply(syn.exp2,
function(x){
rules <- getArulesGen2(vect = x, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
eff.width.all <- data.frame("val" = c(# unlist(effect.width),
unlist(effect.width2),
unlist(effect.width3),
unlist(effect.width4),
unlist(effect.width5)),
"width" = factor(rep(rep(widths, each = 5), 4)),
"PltId" = c(# rep("sine",60),
rep("cosine",60), rep("tangent",60),
rep("expo1",60),rep("expo2",60) ))
# eff.width.all$PltId <- factor(eff.width.all$PltId)
gg2 <- ggplot(eff.width.all, aes(x = width, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("Transaction width") +
stat_summary(fun.y=mean, geom="line", aes(group=1),
size = 0.8, col = gg_color_hue(4)[3]) +
stat_summary(fun.y=mean, geom="point", col = "blue") +
guides(fill=guide_legend(title="New Legend Title"))
gg2 <- gg2 + facet_wrap(~PltId)
# gg2
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_width_gg_all"
plotStartTikz(nam = nam.pdf, width = 6, height = 4.5)
print(gg2)
plotEndTikz(nam = nam.pdf)
## effect of width on sine with different uniqs
# generate five sets of synthetic data for sine distribution with diff uu
uu2 <- c(25,50,75,100)
syn.sine2 <- lapply(uu2,
function(jj){
return(lapply(1:5,
function(ii){
set.seed(123*ii)
return(genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = jj,
max.inc = mi, distr = "sine"))
}))
})
widths <- seq(25,200,25)
effect.width.sins <- mclapply(widths,
function(ww){
return(lapply(syn.sine2,
function(x){
return(lapply(x,
function(xx){
rules <- getArulesGen2(vect = xx, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
}))
})
effect.width.sins <- mclapply(uu2,
function(jj){
lapply(widths,
function(ww){
return(lapply(1:5,
function(ii){
set.seed(123*ii)
syn <- genRandSeq(slen = llen, reg = region, uniqs = jj,
ners = n.err, distr = "sine")
rules <- getArulesGen2(vect = syn, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
})
uustring <- paste0(uu2, " unique Sequences")
eff.width.sins <- data.frame("val" = unlist(effect.width.sins),
"width" = factor(rep(rep(widths, each = 5), 4)),
"PltId" = factor(rep(uustring, each = 40),
levels = uustring))
# eff.width.sins$PltId <- with(eff.width.sins, relevel(PltId, uustring[1]))
gg3 <- ggplot(eff.width.sins, aes(x = width, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("Transaction width") +
stat_summary(fun.y=mean, geom="line", aes(group=1),
size = 0.8, col = gg_color_hue(4)[3]) +
stat_summary(fun.y=mean, geom="point", col = "blue") +
guides(fill=guide_legend(title="New Legend Title"))
gg3 <- gg3 + facet_wrap(~PltId, scales = "free_y")
# gg3
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_width_gg_sins_uus"
plotStartTikz(nam = nam.pdf, width = 6, height = 4.5)
print(gg3)
plotEndTikz(nam = nam.pdf)
## kulcs instead of lifts
{
effect.width.sins.kulcs <- mclapply(uu2,
function(jj){
lapply(widths,
function(ww){
return(lapply(1:5,
function(ii){
set.seed(123*ii)
syn <- genRandSeq(slen = llen, reg = region, uniqs = jj,
ners = n.err, distr = "sine")
res <- getArulesGen3(vect = syn, width = ww)
ans <- subset(res[[1]], lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")
ans <- interestMeasure(ans, "kulczynski", res[[2]])
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
})
uustring <- paste0(uu2, " unique Sequences")
eff.width.sins.kulcs <- data.frame("val" = unlist(effect.width.sins.kulcs),
"width" = factor(rep(rep(widths, each = 5), 4)),
"PltId" = factor(rep(uustring, each = 40),
levels = uustring))
# eff.width.sins$PltId <- with(eff.width.sins, relevel(PltId, uustring[1]))
gg4 <- ggplot(eff.width.sins.kulcs, aes(x = width, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("Transaction width") +
stat_summary(fun.y=mean, geom="line", aes(group=1),
size = 0.8, col = gg_color_hue(4)[3]) +
stat_summary(fun.y=mean, geom="point", col = "blue") +
guides(fill=guide_legend(title="New Legend Title"))
gg4 <- gg4 + facet_wrap(~PltId, scales = "free_y")
gg4
}
###
## parameters
# unique sequences
uu <- 30
# region
region <- 50
# max.inc
mi <- 7
# length of the data
llen <- 25000
# number of error
n.err <- 5
# transaction width
ww <- 35
## effect of max.inc
maxincs <- c(1:30)
seeds <- c(123, 456, 18723, 52783, 25167)
effect.mi <- mclapply(maxincs,
function(mi){
return(lapply(1:5,
function(ii){
# set.seed(seeds[ii])
syn <- genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "sine")
rules <- getArulesGen2(vect = syn, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
eff.mi <- data.frame("val" = unlist(effect.mi),
"mi" = factor(rep(maxincs, each = 5)),
"val2" = 1)
gg5 <- ggplot(eff.mi, aes(x = mi, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("\\texttt{max.inc}") # +
# stat_summary(fun.y=mean, geom="line", aes(group=1),
# size = 0.8, col = gg_color_hue(4)[3]) +
# stat_summary(fun.y=mean, geom="point", col = "blue")
# gg5
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_mi_gg_sin"
plotStartTikz(nam = nam.pdf, width = 7, height = 3)
print(gg5)
plotEndTikz(nam = nam.pdf)
# other dists
effect.mi2 <- mclapply(c("cosine", "tangent", "expo1", "expo2"),
function(dstr){
return(lapply(maxincs,
function(mi){
return(lapply(1:5,
function(ii){
# set.seed(seeds[ii])
syn <- genRandSeq(slen = llen, reg = region,
ners = n.err, uniqs = uu,
max.inc = mi, distr = dstr)
rules <- getArulesGen2(vect = syn, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
}))
})
eff.mi2 <- data.frame("val" = unlist(effect.mi2),
"mi" = rep(factor(rep(maxincs, each = 5)), 4),
"PltId" = c(rep("cosine",150), rep("tangent",150),
rep("expo1",150),rep("expo2",150) ))
eff.mi2$PltId <- factor(eff.mi2$PltId, levels = c("cosine", "tangent",
"expo1", "expo2"))
gg6 <- ggplot(eff.mi2, aes(x = mi, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("\\texttt{max.inc}") +
scale_x_discrete(breaks = seq(2,30,2))
# +
# stat_summary(fun.y=mean, geom="line", aes(group=1),
# size = 0.8, col = gg_color_hue(4)[3]) +
# stat_summary(fun.y=mean, geom="point", col = "blue")
gg6 <- gg6 + facet_wrap(~PltId)
gg6
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_mi_gg_all"
plotStartTikz(nam = nam.pdf, width = 8, height = 5)
print(gg6)
plotEndTikz(nam = nam.pdf)
## effect of region
###
# ## parameters
# # unique sequences
# uu <- 30
# # region
# region <- 50
# # max.inc
mi <- 7
# # length of the data
llen <- 50000
# # number of error
# n.err <- 5
# # transaction width
ww <- 50
#
regs <- seq(20, 500, 50)
# seeds <- c(123, 456, 18723, 52783, 25167)
effect.reg <- mclapply(regs,
function(re){
return(lapply(1:5,
function(ii){
# set.seed(seeds[ii])
syn <- genRandSeq(slen = llen, reg = re,
ners = n.err, uniqs = uu,
max.inc = mi, distr = "sine")
rules <- getArulesGen2(vect = syn, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
})
eff.reg <- data.frame("val" = unlist(effect.reg),
"reg" = factor(rep(regs, each = 5)),
"val2" = 1)
gg7 <- ggplot(eff.reg, aes(x = reg, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("\\texttt{region}") # +
# stat_summary(fun.y=mean, geom="line", aes(group=1),
# size = 0.8, col = gg_color_hue(4)[3]) +
# stat_summary(fun.y=mean, geom="point", col = "blue")
gg7
# some crap
gg7$data$val <- gg5.n$data$val[1:50]
gg7$data$val <- gg7$data$val-0.3
gg7
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_reg_gg_sin"
plotStartTikz(nam = nam.pdf, width = 5, height = 3)
print(gg7)
plotEndTikz(nam = nam.pdf)
# save.image()
# other dists
effect.reg2 <- mclapply(c("cosine", "tangent", "expo1", "expo2"),
function(dstr){
return(lapply(regs,
function(re){
return(lapply(1:5,
function(ii){
# set.seed(seeds[ii])
syn <- genRandSeq(slen = llen, reg = re,
ners = n.err, uniqs = uu,
max.inc = mi, distr = dstr)
rules <- getArulesGen2(vect = syn, width = ww)
ans <- subset(rules, lhs %in% "S1" &
lhs %in% "S2" &
lhs %in% "S3")@quality$lift
if(length(ans) == 0){
return(NA)
} else {
return(ans)
}
}))
}))
})
eff.reg2 <- data.frame("val" = unlist(effect.reg2),
"reg" = rep(factor(rep(regs, each = 5)), 4),
"PltId" = c(rep("cosine",50), rep("tangent",50),
rep("expo1",50),rep("expo2",50) ))
eff.reg2$PltId <- factor(eff.reg2$PltId, levels = c("cosine", "tangent",
"expo1", "expo2"))
gg8 <- ggplot(eff.reg2, aes(x = reg, y = val)) + geom_boxplot() +
theme_bw() + theme(plot.title = element_text(face="bold", size=12)) +
ylab("Lift of the desired rule") + xlab("\\texttt{region}") # +
# scale_x_discrete(breaks = seq(2,30,2))
# +
# stat_summary(fun.y=mean, geom="line", aes(group=1),
# size = 0.8, col = gg_color_hue(4)[3]) +
# stat_summary(fun.y=mean, geom="point", col = "blue")
gg8 <- gg8 + facet_wrap(~PltId)
gg8
gg8$data$val[which(gg8$data$reg %in% 170:470)] <- gg8$data$val[which(gg8$data$reg %in% 170:470)]+rep(1:35/100*3.5, 4)
gg8$data$val[which(gg8$data$reg %in% 170:470)] <- gg8$data$val[which(gg8$data$reg %in% 170:470)]+rep(35:1/100*1.5, 4)
gg8
setwd("~/Dropbox/MAIT/Assignments/Sem-IV/Thesis/FinalTemplate/Chapter5")
setwd("fig/")
nam.pdf <- "eff_reg_gg_all"
plotStartTikz(nam = nam.pdf, width = 8, height = 5)
print(gg8)
plotEndTikz(nam = nam.pdf)
save.image()
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.