inst/coxprocess/analyse_normvelocity_initial.R

In jeremyhengjm/GibbsFlow: Gibbs Flow for Approximate Transport

library(ggplot2)
library(ggthemes)
library(GibbsFlow)
setmytheme()

# plot norm of Gibbs velocity for GF-SIS for different initialization 
rm(list=ls())
nsteps <- 80
load("inst/coxprocess/results/normvelocity_gfsis_400.RData")
plot.normvelocity.df <- cbind(normvelocity.df, data.frame(initialization = rep("Prior", nsteps)))


load("inst/coxprocess/results/normvelocity_gfsis_vi_400.RData")
plot.normvelocity.df <- rbind(plot.normvelocity.df, 
                              cbind(normvelocity.df, data.frame(initialization = rep("VB", nsteps))))

load("inst/coxprocess/results/normvelocity_gfsis_ep_400.RData")
plot.normvelocity.df <- rbind(plot.normvelocity.df, 
                              cbind(normvelocity.df, data.frame(initialization = rep("EP", nsteps))))

gnormvelocity <- ggplot(plot.normvelocity.df, aes(x = time, y = median, ymin = lower, ymax = upper, colour = initialization))
gnormvelocity <- gnormvelocity + geom_pointrange(alpha = 1) + 
  xlim(0, 1) + ylim(0, 20) + 
  # scale_y_continuous(breaks = c(0, 40, 80, 120)) +
  xlab("time") + ylab("norm of Gibbs velocity") + scale_color_colorblind()
gnormvelocity
ggsave(filename = "~/Dropbox/GibbsFlow/draft_v3/coxprocess_normvelocity_gfsis_initial.eps", plot = gnormvelocity,
       device = "eps", width = 6, height = 6)

# plot norm of Gibbs velocity for GF-AIS for different initialization 
rm(list=ls())
nsteps <- 80
load("inst/coxprocess/results/normvelocity_gfais_400.RData")
plot.normvelocity.df <- cbind(normvelocity.df, data.frame(initialization = rep("Prior", nsteps)))


load("inst/coxprocess/results/normvelocity_gfais_vi_400.RData")
plot.normvelocity.df <- rbind(plot.normvelocity.df, 
                              cbind(normvelocity.df, data.frame(initialization = rep("VB", nsteps))))

load("inst/coxprocess/results/normvelocity_gfais_ep_400.RData")
plot.normvelocity.df <- rbind(plot.normvelocity.df, 
                              cbind(normvelocity.df, data.frame(initialization = rep("EP", nsteps))))

gnormvelocity <- ggplot(plot.normvelocity.df, aes(x = time, y = median, ymin = lower, ymax = upper, colour = initialization))
gnormvelocity <- gnormvelocity + geom_pointrange(alpha = 1) + 
  xlim(0, 1) + ylim(0, 20) + 
  # scale_y_continuous(breaks = c(0, 40, 80, 120)) +
  xlab("time") + ylab("norm of Gibbs velocity") + scale_color_colorblind()
gnormvelocity
ggsave(filename = "~/Dropbox/GibbsFlow/draft_v3/coxprocess_normvelocity_gfais_initial.eps", plot = gnormvelocity,
       device = "eps", width = 6, height = 6)