abikoushi

tanuki <- function(x,n){
  letter <- "た"
  len <- nchar(x)
  p <- sort(sample(2:len,n))
  sv <- character(n+1)
  sv[1] <-substr(x,1,p[1])
  for(i in 1:(n-1)){
    sv[i+1] <- substr(x,p[i]+1L,p[i+1])
  }
  sv[n+1L] <- substr(x,p[n]+1L,len)

abikoushi

\documentclass[dvipdfmx, border=2.5cm]{standalone}
\usepackage{tikz}
\usetikzlibrary{arrows.meta, positioning}
\begin{document}
\begin{tikzpicture}

\node[rectangle](t1){$t=1$};
\node[rectangle, below = of t1](t2) {$t=2$};
\node[rectangle, below = of t2](t3) {$t=3$};
\node[rectangle, draw=black, right = of t1](j1) {玉森裕太};

abikoushi

module GG

using Random
using Distributions
using SpecialFunctions
import Distributions: ccdf, cdf, logpdf, pdf, quantile, mean, rand, params, shape, scale
import Distributions: @distr_support

struct GeneralizedGamma{Ta, Tb, Tk} <: ContinuousUnivariateDistribution
    a::Ta

abikoushi

library(estatapi)
library(tidyverse)
myappId <-"ここには自分のアプリケーションIDを入れる"

dat1 <-estat_getStatsData(appId = myappId, statsDataId = "0000010209")

#print(unique(dat1$`Ｉ　健康・医療`)) #ほしい項目を目で探す

nyuin <-dplyr::filter(dat1,`Ｉ　健康・医療`=="#I04103_精神科病院年間新入院患者数（人口10万人当たり）" ,
                      地域=="全国") %>% 

abikoushi

library(mvtnorm)
library(dplyr)
library(ggplot2)

combrect <- function(X){
  B <- combn(nrow(X),2)
  #shuffle <- sample.int(nrow(B))
  apply(B, 2, function(i){
    data.frame(x = X[i[1],1], xmax = X[i[2],1],
               y = X[i[1],2], ymax = X[i[2],2])}) 

abikoushi

library(estatapi)
library(ggplot2)
library(dplyr)


myappId <-"ここには自分のアプリケーションIDを入れる"

#dat_meta <- estat_getStatsList(myappId, "人口動態調査")
#View(dat_meta)

abikoushi

library(estatapi)
library(ggplot2)
library(dplyr)
library(tidyr)

myappId <-"ここには自分のアプリケーションIDを入れる"

# データを探す
# dat_meta <- estat_getStatsList(myappId, "労働力調査")
# View(dat2)

abikoushi

library(gganimate)
library(ggplot2)
#library(tidyr)

sim_pois1 <- function(t_n, p_n, w, seed){
  set.seed(seed)
  l <- 5 - 0.5*w
  r <- 5 + 0.5*w
  counts <- integer(t_n)
  p_list <- vector("list", t_n)

abikoushi

library(BayesLogit)
library(ggplot2)
#y: response variable
#n: binomial size parameter
#X: explanatory design matrix
#lambda: prior parameter
gibbs_logistic <- function(y,n,X,lambda,iter){
  N <- length(y)
  d <- ncol(X)
  B <- diag(lambda, d)

abikoushi

library(BayesLogit)
library(ggplot2)
#########
softmax <- function(x){
  m <- max(x)
  u <- x-m
  exp(u)/sum(exp(u))
}

logsumexp <- function(x){