Commit: 3387593ddd7bde7d7c045fd0eab12f8a90329889 - myprojects-hg-reborn (hg)

Commit MetaInfo

• Revisión: 3387593ddd7bde7d7c045fd0eab12f8a90329889 (tree)
• Tiempo: 2024-10-09 00:29:55
• Autor: Lorenzo Isella <lorenzo.isella@gmai...>
• Commiter: Lorenzo Isella

Log Message

A code to generate some synthetic data.

Cambiar Resumen

• add: R-codes/synthesis.R

Diferencia incremental

diff -r 154d30e4eb43 -r 3387593ddd7b R-codes/synthesis.R
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/R-codes/synthesis.R	Tue Oct 08 17:29:55 2024 +0200
@@ -0,0 +1,55 @@
+rm(list=ls())
+library(tidyverse)
+library(janitor)
+library(charlatan)
+
+source("/home/lorenzo/myprojects-hg/R-codes/stat_lib.R")
+
+set.seed(1234)
+
+nn <- ch_name(30) |>
+    sort()
+
+
+ll <- tibble(x= sample(nn, 100e3, replace=T, prob=(1:30)/sum(1:30))) |>
+    group_by(x) |>
+    group_split()
+
+p <- runif(30,0,1) ## because the failure probability is 1-success probability
+
+df_p <- tibble(real_success_rate=p, group=nn) 
+
+res <- map2(ll,p, \(z,y) tibble(x=rbinom(nrow(z),1, y)))
+
+df_ll <- ll |>
+    list_to_df() |>
+    rename("name"="x")
+    
+df_res <- res |>
+    list_to_df() |>
+    rename("success"="x") ## |>
+    ## mutate(failure=round(failure, 0))
+
+
+test <- df_res |>
+    group_by(source) |>
+    summarise(n=n(), mean_success=mean(success)) |>
+    ungroup() |>
+    arrange(mean_success)
+
+
+df_out <- tibble(name=df_ll$name, success=df_res$success) |>
+    left_join(y=df_p, by=c("name"="group"))
+
+test2 <- df_out |>
+    group_by(name) |>
+    summarise(n=n(), mean_success=mean(success),
+              real_rate=real_success_rate[1]) |>
+    ungroup() |>
+    arrange(mean_success)
+
+
+write_csv(df_out, "artificial_data.csv.gz")
+
+
+print("So far so good")