etas-declustered

library("ETAS")
df.2 <- (df1[,c("time", "latitude", "longitude", "mag")])

time_convert <- function(x){
  dates <- strsplit(x$time, split = "T")
  dates1<-unlist(dates)
  k1<-seq(1, length(dates1), 2)
  dates2<-dates1[k1]
  dates3 <- unlist(strsplit(dates1[-k1], split = "Z"))
  x$date <- dates2
  x$time <- dates3
  return(x)
}

nepal.data <- time_convert(df.2)
colnames(nepal.data)[c(2,3)] <- c("lat", "long")
tail(nepal.data)
gregion <- list(lat = c(26.3, 26.3, 30.4, 30.4),
                   long = c(79, 88.2, 88.2, 79))
nepal.cat <- catalog(nepal.data, study.start = "1990-01-11",
                      study.end = "2022-05-20", 
                      region.poly = gregion, 
                      mag.threshold = 5)
plot(nepal.cat)

param01 <- c(0.5, 0.2, 0.05, 2.7, 1.2, 0.02, 2.3, 0.03)

nepal.fit <- etas(nepal.cat, param0 = param01)
plot(nepal.fit)
rates(nepal.fit)

pr <- probs(nepal.fit)

plot(nepal.cat$longlat.coord[pr$target & (1 - pr$prob > 0.95), 1:2])
points(nepal.cat$longlat.coord[pr$target & (pr$prob > 0.95), 1:2],
          pch = 3, col = 2)
map("world", add = TRUE, col = "grey")
legend("bottomleft", c("background", "triggered"), pch = c(1, 3),
          col = 1:2)

nepal.res <- resid.etas(nepal.fit)
summary(nepal.res$tres)

summary(na.omit(nepal.res$sres$z))

ks.test(nepal.res$U, punif)