New idea to visualize the distribution of PSA sampled parameters

[krijkamp]

When the DARTH framework is uses, it would be nice to show the sampled distributions in an illustrative way.

Example to get started with. But we can improve this

Melt the dataset

df_melt <- melt(df_psa_random, variable.name = "Parameter")

Create a new column to classify parameters by prefix

df_melt <- df_melt %>%
mutate(Group = case_when(
str_starts(Parameter, "p_") ~ "Probabilities",
str_starts(Parameter, "u_") ~ "Utilities",
str_starts(Parameter, "c_") ~ "Costs",
TRUE ~ "Other"
))

Optionally check the groups

table(df_melt$Group)

df_probs <- df_melt %>% filter(Group == "Probabilities")

Probabilities:
ggplot(df_probs, aes(x = value, y = Parameter)) +
geom_density_ridges(scale = 1.5, rel_min_height = 0.01) +
theme_bw(base_size = 14) +
ggtitle("Probabilities")

df_utils <- df_melt %>% filter(Group == "Utilities")

ggplot(df_utils, aes(x = value, y = Parameter)) +
geom_density_ridges(scale = 1.5, rel_min_height = 0.01) +
theme_bw(base_size = 14) +
ggtitle("Utilities")

df_costs <- df_melt %>% filter(Group == "Costs")

ggplot(df_costs, aes(x = value, y = Parameter)) +
geom_density_ridges(scale = 1.5, rel_min_height = 0.01) +
theme_bw(base_size = 14) +
ggtitle("Costs")

df_other <- df_melt %>% filter(Group == "Other")

ggplot(df_other, aes(x = value, y = Parameter)) +
geom_density_ridges(scale = 1.5, rel_min_height = 0.01) +
theme_bw(base_size = 14) +
ggtitle("Other")