library(tseries)
salida_JB <- jarque.bera.test(modelo_estimado$residuals)
print(salida_JB)

library(fastGraph)
alpha_sign <- 0.05
JB <- salida_JB$statistic
GL <- salida_JB$parameter
VC <- qchisq(1-alpha_sign,GL,lower.tail = TRUE)
shadeDist(JB,ddist = "dchisq",
          parm1 = GL,
          lower.tail = FALSE,xmin = 0,
          sub=paste("VC:",round(VC,2)," ","JB:",round(JB,2)))

library(normtest)
jb.norm.test(modelo_estimado$residuals)

library(dplyr)  
library(gt)  
library(gtExtras)
residuos<-modelo_estimado$residuals
residuos %>%
  as_tibble() %>%
  mutate(posicion=row_number()) %>%
  arrange(value) %>%
  mutate(dist1=row_number()/n()) %>%
  mutate(dist2=(row_number()-1)/n()) %>%
  mutate(zi=as.vector(scale(value,center=TRUE))) %>%
  mutate(pi=pnorm(zi,lower.tail = TRUE)) %>%
  mutate(dif1=abs(dist1-pi)) %>%
  mutate(dif2=abs(dist2-pi)) %>%
  rename(residuales=value) -> tabla_KS

#Formato
 tabla_KS %>%
  gt() %>%
  tab_header("Tabla para calcular el Estadistico KS") %>%
  tab_source_note(source_note = "Fuente: Elaboración propia") %>%
  tab_style(
    style = list(
      cell_fill(color = "#A569BD"),
      cell_text(style = "italic")
      ),
    locations = cells_body(
      columns = dif1,
      rows = dif1==max(dif1)
    )) %>%
   tab_style(
    style = list(
      cell_fill(color = "#3498DB"),
      cell_text(style = "italic")
      ),
    locations = cells_body(
      columns = dif2,
      rows = dif2==max(dif2)
    ))

D<-max(max(tabla_KS$dif1),max(tabla_KS$dif2))
print(D)

library(gt)
data.frame(
  n = c("50", ">100*"), 
  α_0.20 = c(0.086, 0.061),
  α_0.05 = c(0.115, 0.081),
  Comparación = c("0.0754 < 0.115", "0.0754 < 0.081")
) %>% 
  gt() %>% 
  tab_footnote("*Valores para n=88")

library(nortest)
prueba_KS <- lillie.test(modelo_estimado$residuals)
print(prueba_KS)

p.value <- prueba_KS$p.value

library(dplyr)
library(gt)
residuos<-modelo_estimado$residuals
residuos %>%  
  as_tibble() %>%
  rename(residuales=value) %>%
  arrange(residuales) %>%
  mutate(pi=(row_number()-0.375)/(n()+0.25)) %>%
  mutate(mi=qnorm(pi,lower.tail = TRUE)) %>% 
  mutate(ai=0)->tabla_SW

m<-sum(tabla_SW$mi^2)
n <- nrow(residuos)
print(class(residuos))

print(n)

residuos <- as.data.frame(residuos)
n <- nrow(residuos)
theta<-1/sqrt(n)
tabla_SW$ai[n]<- -2.706056*theta^5+4.434685*theta^4-2.071190*theta^3-0.147981*theta^2+0.2211570*theta+tabla_SW$mi[n]/sqrt(m)
tabla_SW$ai[n-1]<- -3.582633*theta^5+5.682633*theta^4-1.752461*theta^3-0.293762*theta^2+0.042981*theta+tabla_SW$mi[n-1]/sqrt(m)
tabla_SW$ai[1]<- -tabla_SW$ai[n]
tabla_SW$ai[2]<- -tabla_SW$ai[n-1]
omega<-(m-2*tabla_SW$mi[n]^2-2*tabla_SW$mi[n-1]^2)/(1-2*tabla_SW$ai[n]^2-2*tabla_SW$ai[n-1]^2)
tabla_SW$ai[3:(n-2)]<-tabla_SW$mi[3:(n-2)]/sqrt(omega)

tabla_SW %>% 
  mutate(ai_ui=ai*residuales,ui2=residuales^2) ->tabla_SW

tabla_SW %>%
  gt() %>% tab_header("Tabla para calcular el Estadistico W") %>% 
  tab_source_note(source_note = "Fuente: Elaboración propia")

W<-(sum(tabla_SW$ai_ui)^2)/sum(tabla_SW$ui2)
print(W)

mu<-0.0038915*log(n)^3-0.083751*log(n)^2-0.31082*log(n)-1.5861
sigma<-exp(0.0030302*log(n)^2-0.082676*log(n)-0.4803)
Wn<-(log(1-W)-mu)/sigma
print(Wn)

p.value<-pnorm(Wn,lower.tail = FALSE)
print(p.value)

library(fastGraph)
shadeDist(Wn,ddist = "dnorm",lower.tail = FALSE)

salida_SW<-shapiro.test(modelo_estimado$residuals)
print(salida_SW)

Wn_salida<-qnorm(salida_SW$p.value,lower.tail = FALSE)
print(Wn_salida)