Data management Figure for ITSTGCN

STGCN

Author

SEOYEON CHOI

Published

July 20, 2023

Import

library(ggplot2)

library(dplyr)

df <- read.csv("./df_fig.csv")

head(df)

	X	dataset	method	mrate	mtype	lags	nof_filters	inter_method	epoch	mse	calculation_time	model
	<int>	<chr>	<chr>	<dbl>	<chr>	<int>	<dbl>	<chr>	<dbl>	<dbl>	<dbl>	<chr>
1	0	fivenodes	STGCN	0.0		2	12		50	0.7293743	80.98522	GConvGRU
2	1	fivenodes	STGCN	0.0		2	12		50	0.7290817	80.89179	GConvGRU
3	2	fivenodes	STGCN	0.7	rand	2	12	linear	50	1.8922616	81.97655	GConvGRU
4	3	fivenodes	STGCN	0.7	rand	2	12	nearest	50	2.2112885	87.80387	GConvGRU
5	4	fivenodes	STGCN	0.8	rand	2	12	linear	50	2.0728178	103.64874	GConvGRU
6	5	fivenodes	STGCN	0.8	rand	2	12	nearest	50	2.5664744	98.34010	GConvGRU

A data.frame: 6 × 12

fivenode

fivenodes <- data.frame(rbind(
    df %>% filter(dataset=='fivenodes' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==2 & epoch==50 & model=='GConvGRU'  & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate == 0.7  | mrate ==  0.8)),
    df %>% filter(dataset=='fivenodes' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==2 & epoch==50 & model=='GConvLSTM' & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate == 0.7  | mrate ==  0.8)),
    df %>% filter(dataset=='fivenodes' & mtype=='rand' & inter_method == 'linear' & nof_filters==4 & lags==2 & epoch==50 & model=='GCLSTM' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate == 0.7  | mrate ==  0.8)),
    df %>% filter(dataset=='fivenodes' & mtype=='rand' & inter_method == 'linear' & nof_filters==4 & lags==2 & epoch==50 & model=='LRGCN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate == 0.7  | mrate ==  0.8)),
    df %>% filter(dataset=='fivenodes' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==2 & epoch==50 & model=='DyGrEncoder' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate == 0.7  | mrate ==  0.8)),
    df %>% filter(dataset=='fivenodes' & mtype=='rand' & inter_method == 'linear' & lags==2 & epoch==50 & model=='EvolveGCNH' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate == 0.7  | mrate ==  0.8)),
    df %>% filter(dataset=='fivenodes' & mtype=='rand' & inter_method == 'linear' & lags==2 & epoch==50 & model=='EvolveGCNO' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate == 0.7  | mrate ==  0.8)),
    df %>% filter(dataset=='fivenodes' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==2 & epoch==50 & model=='TGCN' & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate == 0.7  | mrate ==  0.8)),
    df %>% filter(dataset=='fivenodes' & mtype=='rand' & inter_method == 'linear' & nof_filters==2 & lags==2 & epoch==50 & model=='DCRNN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate == 0.7  | mrate ==  0.8))))

fivenodes$model <- factor(fivenodes$model, levels = c("GConvGRU","GConvLSTM","GCLSTM","LRGCN","DyGrEncoder","EvolveGCNH","EvolveGCNO","TGCN","DCRNN"))
fivenodes$method <- factor(fivenodes$method, levels = c('STGCN','IT-STGCN'))

후보 1

ggplot(fivenodes, aes(x=mrate,y= mse,group=mrate)) + facet_wrap(model~method) + 
geom_boxplot(fill='grey',color='black',width=0.7,outlier.color = 'darkblue',outlier.shape = 2) + theme_classic()
# ggsave("random_list_fivenodes.png")

후보 2

ggplot(fivenodes, aes(x=mrate,y= log10(mse),group=mrate)) + facet_wrap(model~method,,ncol=4) + 
geom_boxplot(fill='grey',color='black',width=0.7,outlier.color = 'darkblue',outlier.shape = 2) + 
theme_classic() 
# ggsave("random_list_fivenodes.png")

후보 3

ggplot(fivenodes, aes(x = mrate, y = log10(mse), group = mrate)) +
  geom_boxplot(fill = 'grey', color = 'black', width = 0.7, outlier.color = 'darkblue', outlier.shape = 2) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank()) +
  facet_wrap(model ~ method, ncol = 4, scales = "free_y",) +
  xlab("Missing Rate") +
  ylab("MSE(log scale)") +
  scale_x_continuous(labels=c('','','','',''))
# ggsave("random_list_fivenodes.pdf")

후보 4

p1 = ggplot(fivenodes, aes(x = mrate, y = log10(mse), group = interaction(mrate, method))) +
  geom_boxplot(aes(fill = method), color = 'black', width = 0.7, outlier.color = 'black', outlier.shape = 2) +
  scale_fill_manual(values = c("steelblue", "pink"), labels=c('Classic', 'Proposed')) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank(),
       legend.position = "") +
  labs(x = " ", y = "", fill = "") +
  scale_x_continuous(labels = c('', '', '', '', '')) +
  facet_wrap(~model, ncol = 3, scales = "free_y")
# ggsave("random_list_fivenodes.pdf")

ggplot(fivenodes, aes(x = mrate, y = log10(mse), group = interaction(mrate, method))) +
  geom_boxplot(aes(fill = method), color = 'black', width = 0.7, outlier.color = 'black', outlier.shape = 2) +
  scale_fill_manual(values = c("steelblue", "pink"), labels=c('Classic', 'Proposed')) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank(),
       legend.position = 'bottom') +
  labs(x = "Missing Rate", y = "MSE(log scale)", fill = "") +
  scale_x_continuous(labels = c('', '', '', '', '')) +
  facet_wrap(~model, ncol = 3, scales = "free_y")
# ggsave("random_list_fivenodes.pdf")

ChickenpoxDatasetLoader(lags=4)

chickenpox <- rbind(
    df %>% filter(dataset=='chickenpox' & mtype=='rand' & inter_method == 'linear' & nof_filters==16 & lags==4 & epoch==50 & model=='GConvGRU'  & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='chickenpox' & mtype=='rand' & inter_method == 'linear' & nof_filters==32 & lags==4 & epoch==50 & model=='GConvLSTM' & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='chickenpox' & mtype=='rand' & inter_method == 'linear' & nof_filters==16 & lags==4 & epoch==50 & model=='GCLSTM'  & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='chickenpox' & mtype=='rand' & inter_method == 'linear' & nof_filters==8 & lags==4 & epoch==50 & model=='LRGCN' &  
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='chickenpox' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==4 & epoch==50 & model=='DyGrEncoder' & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='chickenpox' & mtype=='rand' & inter_method == 'linear' & lags==4 & epoch==50 & model=='EvolveGCNH' & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='chickenpox' & mtype=='rand' & inter_method == 'linear' & lags==4 & epoch==50 & model=='EvolveGCNO' & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>%filter(dataset=='chickenpox' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==4 & epoch==50 & model=='TGCN'  &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>%filter(dataset=='chickenpox' & mtype=='rand' & inter_method == 'linear' & nof_filters==16 & lags==4 & epoch==50 & model=='DCRNN'  &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)))

chickenpox$model <- factor(chickenpox$model, levels = c("GConvGRU","GConvLSTM","GCLSTM","LRGCN","DyGrEncoder","EvolveGCNH","EvolveGCNO","TGCN","DCRNN"))
chickenpox$method <- factor(chickenpox$method, levels = c('STGCN','IT-STGCN'))

ggplot(chickenpox, aes(x=mrate,y= mse,group=mrate)) + facet_wrap(model~method,ncol=4) + 
geom_boxplot(fill='grey',color='black',width=0.7,outlier.color = 'darkblue',outlier.shape = 2) + theme_classic()

ggplot(chickenpox, aes(x=mrate,y= log10(mse),group=mrate)) + facet_wrap(model~method,ncol=4) + 
geom_boxplot(fill='grey',color='black',width=0.7,outlier.color = 'darkblue',outlier.shape = 2) + theme_classic()

ggplot(chickenpox, aes(x=mrate,y= log10(mse),group=mrate)) + 
geom_boxplot(fill='grey',color='black',width=0.7,outlier.color = 'darkblue',outlier.shape = 2) + 
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank()) +
  facet_wrap(model ~ method, ncol = 4, scales = "free_y") +
  xlab("Missing Rate") +
  ylab("MSE(log scale)") +
  scale_x_continuous(labels=c('','','','',''))
# ggsave("random_list_chickenpox.pdf")

p2= ggplot(chickenpox, aes(x = mrate, y = log10(mse), group = interaction(mrate, method))) +
  geom_boxplot(aes(fill = method), color = 'black', width = 0.7, outlier.color = 'black', outlier.shape = 2) +
  scale_fill_manual(values = c("steelblue", "pink"), labels=c('Classic', 'Proposed')) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank(),
       legend.position = '') +
  labs(x = " ", y = "", fill = "") +
  scale_x_continuous(labels = c('', '', '', '', '')) +
  facet_wrap(~model, ncol = 3, scales = "free_y")
# ggsave("random_list_chickenpox.pdf")

ggplot(chickenpox, aes(x = mrate, y = log10(mse), group = interaction(mrate, method))) +
  geom_boxplot(aes(fill = method), color = 'black', width = 0.7, outlier.color = 'black', outlier.shape = 2) +
  scale_fill_manual(values = c("steelblue", "pink"), labels=c('Classic', 'Proposed')) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank(),
       legend.position = 'bottom') +
  labs(x = "Missing Rate", y = "MSE(log scale)", fill = "") +
  scale_x_continuous(labels = c('', '', '', '', '')) +
  facet_wrap(~model, ncol = 3, scales = "free_y")
# ggsave("random_list_chickenpox.pdf")

PedalMeDatasetLoader (lags=4)

pedalme <- rbind(
    df %>% filter(dataset=='pedalme' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==4 & epoch==50 & model=='GConvGRU' & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='pedalme' & mtype=='rand' & inter_method == 'linear' & nof_filters==2 & lags==4 & epoch==50 & model=='GConvLSTM' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='pedalme' & mtype=='rand' & inter_method == 'linear' & nof_filters==4 & lags==4 & epoch==50 & model=='GCLSTM' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='pedalme' & mtype=='rand' & inter_method == 'linear' & nof_filters==8 & lags==4 & epoch==50 & model=='LRGCN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='pedalme' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==4 & epoch==50 & model=='DyGrEncoder' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='pedalme' & mtype=='rand' & inter_method == 'linear' & lags==4 & epoch==50 & model=='EvolveGCNH' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='pedalme' & mtype=='rand' & inter_method == 'linear' & lags==4 & epoch==50 & model=='EvolveGCNO' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='pedalme' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==4 & epoch==50 & model=='TGCN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='pedalme' & mtype=='rand' & inter_method == 'linear' & nof_filters==8 & lags==4 & epoch==50 & model=='DCRNN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)))

pedalme$model <- factor(pedalme$model, levels = c("GConvGRU","GConvLSTM","GCLSTM","LRGCN","DyGrEncoder","EvolveGCNH","EvolveGCNO","TGCN","DCRNN"))
pedalme$method <- factor(pedalme$method, levels = c('STGCN','IT-STGCN'))

ggplot(pedalme, aes(x=mrate,y= mse,group=mrate)) + facet_wrap(model~method,ncol=4) + 
geom_boxplot(fill='grey',color='black',width=0.7,outlier.color = 'darkblue',outlier.shape = 2) + theme_classic()

ggplot(pedalme, aes(x=mrate,y= log10(mse),group=mrate)) + facet_wrap(model~method,ncol=4) + 
geom_boxplot(fill='grey',color='black',width=0.7,outlier.color = 'darkblue',outlier.shape = 2) + theme_classic()

ggplot(pedalme, aes(x=mrate,y= log10(mse),group=mrate)) +
geom_boxplot(fill='grey',color='black',width=0.7,outlier.color = 'darkblue',outlier.shape = 2) + 
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank()) +
  facet_wrap(model ~ method, ncol = 4, scales = "free_y") +
  xlab("Missing Rate") +
  ylab("MSE(log scale)") +
  scale_x_continuous(labels=c('','','','',''))
# ggsave("random_list_pedalme.pdf")

p3 = ggplot(pedalme, aes(x = mrate, y = log10(mse), group = interaction(mrate, method))) +
  geom_boxplot(aes(fill = method), color = 'black', width = 0.7, outlier.color = 'black', outlier.shape = 2) +
  scale_fill_manual(values = c("steelblue", "pink"), labels=c('Classic', 'Proposed')) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank(),
       legend.position = '') +
  labs(x = "", y = "", fill = "") +
  scale_x_continuous(labels = c('', '', '', '', '')) +
  facet_wrap(~model, ncol = 3, scales = "free_y")
# ggsave("random_list_pedalme.pdf")

ggplot(pedalme, aes(x = mrate, y = log10(mse), group = interaction(mrate, method))) +
  geom_boxplot(aes(fill = method), color = 'black', width = 0.7, outlier.color = 'black', outlier.shape = 2) +
  scale_fill_manual(values = c("steelblue", "pink"), labels=c('Classic', 'Proposed')) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank(),
       legend.position = 'bottom') +
  labs(x = "Missing Rate", y = "MSE(log scale)", fill = "") +
  scale_x_continuous(labels = c('', '', '', '', '')) +
  facet_wrap(~model, ncol = 3, scales = "free_y")
# ggsave("random_list_pedalme.pdf")

WikiMathsDatasetLoader (lags=8)

wikimath <- rbind(
    df %>% filter(dataset=='wikimath' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==8 & epoch==50 & model=='GConvGRU' & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='wikimath' & mtype=='rand' & inter_method == 'linear' & nof_filters==64 & lags==8 & epoch==50 & model=='GConvLSTM' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='wikimath' & mtype=='rand' & inter_method == 'linear' & nof_filters==64 & lags==8 & epoch==50 & model=='GCLSTM' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='wikimath' & mtype=='rand' & inter_method == 'linear' & nof_filters==32 & lags==8 & epoch==50 & model=='LRGCN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='wikimath' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==8 & epoch==50 & model=='DyGrEncoder' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='wikimath' & mtype=='rand' & inter_method == 'linear' & lags==8 & epoch==50 & model=='EvolveGCNH' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='wikimath' & mtype=='rand' & inter_method == 'linear' & lags==8 & epoch==50 & model=='EvolveGCNO' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='wikimath' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==8 & epoch==50 & model=='TGCN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)),
    df %>% filter(dataset=='wikimath' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==8 & epoch==50 & model=='DCRNN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.8)))

wikimath$model <- factor(wikimath$model, levels = c("GConvGRU","GConvLSTM","GCLSTM","LRGCN","DyGrEncoder","EvolveGCNH","EvolveGCNO","TGCN","DCRNN"))
wikimath$method <- factor(wikimath$method, levels = c('STGCN','IT-STGCN'))

ggplot(wikimath, aes(x=mrate,y= log10(mse),group=mrate)) +
geom_boxplot(fill='grey',color='black',width=0.7,outlier.color = 'darkblue',outlier.shape = 2) + 
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank()) +
  facet_wrap(model ~ method, ncol = 4, scales = "free_y") +
  xlab("Missing Rate") +
  ylab("MSE(log scale)") +
  scale_x_continuous(labels=c('','','','',''))
# ggsave("random_list_wikimath.pdf")

p4 = ggplot(wikimath, aes(x = mrate, y = log10(mse), group = interaction(mrate, method))) +
  geom_boxplot(aes(fill = method), color = 'black', width = 0.7, outlier.color = 'black', outlier.shape = 2) +
  scale_fill_manual(values = c("steelblue", "pink"), labels=c('Classic', 'Proposed')) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank(),
       legend.position = '') +
  labs(x = "", y = "", fill = "") +
  scale_x_continuous(labels = c('', '', '', '', '')) +
  facet_wrap(~model, ncol = 3, scales = "free_y")
# ggsave("random_list_wikimath.pdf")

ggplot(wikimath, aes(x = mrate, y = log10(mse), group = interaction(mrate, method))) +
  geom_boxplot(aes(fill = method), color = 'black', width = 0.7, outlier.color = 'black', outlier.shape = 2) +
  scale_fill_manual(values = c("steelblue", "pink"), labels=c('Classic', 'Proposed')) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank(),
       legend.position = 'bottom') +
  labs(x = "Missing Rate", y = "MSE(log scale)", fill = "") +
  scale_x_continuous(labels = c('', '', '', '', '')) +
  facet_wrap(~model, ncol = 3, scales = "free_y")
# ggsave("random_list_wikimath.pdf")

Windmillsmall(lags=8)

windmillsmall <- rbind(
    df %>% filter(dataset=='windmillsmall' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==8 & epoch==50 & model=='GConvGRU' & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.7)),
    df %>% filter(dataset=='windmillsmall' & mtype=='rand' & inter_method == 'linear' & nof_filters==16 & lags==8 & epoch==50 & model=='GConvLSTM' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.7)),
    df %>% filter(dataset=='windmillsmall' & mtype=='rand' & inter_method == 'linear' & nof_filters==16 & lags==8 & epoch==50 & model=='GCLSTM' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.7)),
    df %>% filter(dataset=='windmillsmall' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==8 & epoch==50 & model=='LRGCN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.7)),
    df %>% filter(dataset=='windmillsmall' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==8 & epoch==50 & model=='DyGrEncoder' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.7)),
    df %>% filter(dataset=='windmillsmall' & mtype=='rand' & inter_method == 'linear' & lags==8 & epoch==50 & model=='EvolveGCNH' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.7)),
    df %>% filter(dataset=='windmillsmall' & mtype=='rand' & inter_method == 'linear' & lags==8 & epoch==50 & model=='EvolveGCNO' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.7)),
    df %>% filter(dataset=='windmillsmall' & mtype=='rand' & inter_method == 'linear' & nof_filters==12 & lags==8 & epoch==50 & model=='TGCN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.7)),
    df %>% filter(dataset=='windmillsmall' & mtype=='rand' & inter_method == 'linear' & nof_filters==4 & lags==8 & epoch==50 & model=='DCRNN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.6 | mrate ==  0.7)))

windmillsmall$model <- factor(windmillsmall$model, levels = c("GConvGRU","GConvLSTM","GCLSTM","LRGCN","DyGrEncoder","EvolveGCNH","EvolveGCNO","TGCN","DCRNN"))
windmillsmall$method <- factor(windmillsmall$method, levels = c('STGCN','IT-STGCN'))

ggplot(windmillsmall, aes(x=mrate,y= log10(mse),group=mrate)) +
geom_boxplot(fill='grey',color='black',width=0.7,outlier.color = 'darkblue',outlier.shape = 2) + 
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank()) +
  facet_wrap(model ~ method, ncol = 2, scales = "free_y") +
  xlab("Missing Rate") +
  ylab("MSE(log scale)") +
  scale_x_continuous(labels=c('','','','',''))
# ggsave("random_list_windmill.pdf")

p5 = ggplot(windmillsmall, aes(x = mrate, y = log10(mse), group = interaction(mrate, method))) +
  geom_boxplot(aes(fill = method), color = 'black', width = 0.7, outlier.color = 'black', outlier.shape = 2) +
  scale_fill_manual(values = c("steelblue", "pink"), labels=c('Classic', 'Proposed')) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank(),
       legend.position = '') +
  labs(x = "", y = "", fill = "") +
  scale_x_continuous(labels = c('', '', '', '', '')) +
  facet_wrap(~model, ncol = 3, scales = "free_y")
# ggsave("random_list_windmill.pdf")

ggplot(windmillsmall, aes(x = mrate, y = log10(mse), group = interaction(mrate, method))) +
  geom_boxplot(aes(fill = method), color = 'black', width = 0.7, outlier.color = 'black', outlier.shape = 2) +
  scale_fill_manual(values = c("steelblue", "pink"), labels=c('Classic', 'Proposed')) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank(),
       legend.position = 'bottom') +
  labs(x = "Missing Rate", y = "MSE(log scale)", fill = "") +
  scale_x_continuous(labels = c('', '', '', '', '')) +
  facet_wrap(~model, ncol = 3, scales = "free_y")
# ggsave("random_list_windmill.pdf")

Montevideobus (lags=4)

monte <- rbind(
    df %>% filter(dataset=='monte' & mtype=='rand' & inter_method == 'nearest' & nof_filters==12 & lags==4 & epoch==50 & model=='GConvGRU' & 
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.7 | mrate ==  0.8)),
    df %>% filter(dataset=='monte' & mtype=='rand' & inter_method == 'nearest' & nof_filters==12 & lags==4 & epoch==50 & model=='GConvLSTM' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.7 | mrate ==  0.8)),
    df %>% filter(dataset=='monte' & mtype=='rand' & inter_method == 'nearest' & nof_filters==12 & lags==4 & epoch==50 & model=='GCLSTM' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.7 | mrate ==  0.8)),
    df %>% filter(dataset=='monte' & mtype=='rand' & inter_method == 'nearest' & nof_filters==2 & lags==4 & epoch==50 & model=='LRGCN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.7 | mrate ==  0.8)),
    df %>% filter(dataset=='monte' & mtype=='rand' & inter_method == 'nearest' & nof_filters==12 & lags==4 & epoch==50 & model=='DyGrEncoder' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.7 | mrate ==  0.8)),
    df %>% filter(dataset=='monte' & mtype=='rand' & inter_method == 'nearest' & lags==4 & epoch==50 & model=='EvolveGCNH' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.7 | mrate ==  0.8)),
    df %>% filter(dataset=='monte' & mtype=='rand' & inter_method == 'nearest' & lags==4 & epoch==50 & model=='EvolveGCNO' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.7 | mrate ==  0.8)),
    df %>% filter(dataset=='monte' & mtype=='rand' & inter_method == 'nearest' & nof_filters==8 & lags==4 & epoch==50 & model=='TGCN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.7 | mrate ==  0.8)),
    df %>% filter(dataset=='monte' & mtype=='rand' & inter_method == 'nearest' & nof_filters==12 & lags==4 & epoch==50 & model=='DCRNN' &
                  ( mrate == 0.3  | mrate == 0.5  |mrate ==  0.7 | mrate ==  0.8)))

monte$model <- factor(monte$model, levels = c("GConvGRU","GConvLSTM","GCLSTM","LRGCN","DyGrEncoder","EvolveGCNH","EvolveGCNO","TGCN","DCRNN"))
monte$method <- factor(monte$method, levels = c('STGCN','IT-STGCN'))

ggplot(monte, aes(x=mrate,y= log10(mse),group=mrate)) + facet_wrap(model~method,ncol=4) + 
geom_boxplot(fill='grey',color='black',width=0.7,outlier.color = 'darkblue',outlier.shape = 2) + 
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank()) +
  facet_wrap(model ~ method, ncol = 4, scales = "free_y") +
  xlab("Missing Rate") +
  ylab("MSE(log scale)")+
  scale_x_continuous(labels=c('','','','',''))
# ggsave("random_list_monte.pdf")

p6 = ggplot(monte, aes(x = mrate, y = log10(mse), group = interaction(mrate, method))) +
  geom_boxplot(aes(fill = method), color = 'black', width = 0.7, outlier.color = 'black', outlier.shape = 2) +
  scale_fill_manual(values = c("steelblue", "pink"), labels=c('Classic', 'Proposed')) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank(),
       legend.position = 'bottom') +
  labs(x = "", y = "", fill = "") +
  scale_x_continuous(labels = c('', '', '', '', '')) +
  facet_wrap(~model, ncol = 3, scales = "free_y")
# ggsave("random_list_monte.pdf")

ggplot(monte, aes(x = mrate, y = log10(mse), group = interaction(mrate, method))) +
  geom_boxplot(aes(fill = method), color = 'black', width = 0.7, outlier.color = 'black', outlier.shape = 2) +
  scale_fill_manual(values = c("steelblue", "pink"), labels=c('Classic', 'Proposed')) +
  theme(strip.background = element_blank(), strip.placement = "outside", panel.grid = element_blank(),
       legend.position = '') +
  labs(x = "Missing Rate", y = "MSE(log scale)", fill = "") +
  scale_x_continuous(labels = c('', '', '', '', '')) +
  facet_wrap(~model, ncol = 3, scales = "free_y")
# ggsave("random_list_monte.pdf")

ggplot3

source('ggplot3.R')

ERROR: Error in library(tidyverse): there is no package called ‘tidyverse’

figsize(20,20)
(p1|p2)/(p3|p4)/(p5|p6)
# ggsave("random_mrate.pdf",width = 20,height = 20)

fivenodes

X	dataset	method	mrate	mtype	lags	nof_filters	inter_method	epoch	mse	calculation_time	model
<int>	<chr>	<fct>	<dbl>	<chr>	<int>	<dbl>	<chr>	<dbl>	<dbl>	<dbl>	<fct>
2	fivenodes	STGCN	0.7	rand	2	12	linear	50	1.892262	81.97655	GConvGRU
4	fivenodes	STGCN	0.8	rand	2	12	linear	50	2.072818	103.64874	GConvGRU
8	fivenodes	IT-STGCN	0.7	rand	2	12	linear	50	1.244197	123.85583	GConvGRU
10	fivenodes	IT-STGCN	0.8	rand	2	12	linear	50	1.428064	116.14180	GConvGRU
14	fivenodes	STGCN	0.7	rand	2	12	linear	50	1.699677	99.00136	GConvGRU
16	fivenodes	STGCN	0.8	rand	2	12	linear	50	2.333100	98.85933	GConvGRU
20	fivenodes	IT-STGCN	0.7	rand	2	12	linear	50	1.267609	114.73862	GConvGRU
22	fivenodes	IT-STGCN	0.8	rand	2	12	linear	50	1.424445	113.63355	GConvGRU
26	fivenodes	STGCN	0.7	rand	2	12	linear	50	1.868236	114.07605	GConvGRU
28	fivenodes	STGCN	0.8	rand	2	12	linear	50	2.389049	99.55423	GConvGRU
32	fivenodes	IT-STGCN	0.7	rand	2	12	linear	50	1.103083	122.35968	GConvGRU
34	fivenodes	IT-STGCN	0.8	rand	2	12	linear	50	1.337490	140.39959	GConvGRU
38	fivenodes	STGCN	0.7	rand	2	12	linear	50	1.816421	115.28978	GConvGRU
40	fivenodes	STGCN	0.8	rand	2	12	linear	50	2.235262	122.47946	GConvGRU
44	fivenodes	IT-STGCN	0.7	rand	2	12	linear	50	1.159646	141.34393	GConvGRU
46	fivenodes	IT-STGCN	0.8	rand	2	12	linear	50	1.301588	147.76632	GConvGRU
50	fivenodes	STGCN	0.7	rand	2	12	linear	50	1.931288	119.68102	GConvGRU
52	fivenodes	STGCN	0.8	rand	2	12	linear	50	2.349104	119.17171	GConvGRU
56	fivenodes	IT-STGCN	0.7	rand	2	12	linear	50	1.195797	135.26314	GConvGRU
58	fivenodes	IT-STGCN	0.8	rand	2	12	linear	50	1.403969	137.56146	GConvGRU
62	fivenodes	STGCN	0.7	rand	2	12	linear	50	1.877509	114.22763	GConvGRU
64	fivenodes	STGCN	0.8	rand	2	12	linear	50	2.517601	116.79539	GConvGRU
68	fivenodes	IT-STGCN	0.7	rand	2	12	linear	50	1.196040	137.43869	GConvGRU
70	fivenodes	IT-STGCN	0.8	rand	2	12	linear	50	1.285712	139.89288	GConvGRU
74	fivenodes	STGCN	0.7	rand	2	12	linear	50	1.614468	120.67241	GConvGRU
76	fivenodes	STGCN	0.8	rand	2	12	linear	50	2.149269	119.84526	GConvGRU
80	fivenodes	IT-STGCN	0.7	rand	2	12	linear	50	1.147181	136.44781	GConvGRU
82	fivenodes	IT-STGCN	0.8	rand	2	12	linear	50	1.276231	135.66005	GConvGRU
86	fivenodes	STGCN	0.7	rand	2	12	linear	50	1.765978	115.49126	GConvGRU
88	fivenodes	STGCN	0.8	rand	2	12	linear	50	2.442478	116.26185	GConvGRU
⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮
9760	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.215530	21.37621	DCRNN
9762	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.206242	35.11960	DCRNN
9764	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.234817	21.35544	DCRNN
9766	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.199285	19.50826	DCRNN
9768	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.268049	16.97555	DCRNN
9770	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.199859	18.76043	DCRNN
9772	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.276552	15.83787	DCRNN
9774	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.208091	17.71081	DCRNN
9776	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.273245	19.52254	DCRNN
9778	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.246838	38.88582	DCRNN
9780	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.280940	21.22349	DCRNN
9782	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.223870	31.15116	DCRNN
9784	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.320345	21.75368	DCRNN
9786	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.211209	25.25301	DCRNN
9788	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.315305	21.44829	DCRNN
9790	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.197776	27.25900	DCRNN
9792	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.267516	20.36553	DCRNN
9794	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.204517	35.18677	DCRNN
9796	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.239286	21.63872	DCRNN
9798	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.197067	26.69481	DCRNN
9800	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.201229	14.04772	DCRNN
9802	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.199647	18.36858	DCRNN
9804	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.201632	14.90250	DCRNN
9806	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.203077	18.51829	DCRNN
9808	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.338037	20.13046	DCRNN
9810	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.318265	25.98562	DCRNN
9812	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.247469	22.01481	DCRNN
9814	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.251032	33.56302	DCRNN
9816	fivenodes	STGCN	0.6	rand	2	2	linear	50	1.285776	21.56314	DCRNN
9818	fivenodes	IT-STGCN	0.6	rand	2	2	linear	50	1.196750	29.83325	DCRNN

A data.frame: 2700 × 12