i meet bollow error todaycan you help me

[tianjinwolaile]

library(INLA)
n <- 10
locations <- data.frame(x = runif(n), y = runif(n))
values <- rnorm(n)
w <- matrix(0, n, n)
for (i in 1:n) {
for (j in 1:n) {
dist <- sqrt((locations$x[i] - locations$x[j])^2 + (locations$y[i] - locations$y[j])^2)
if (dist < 0.2) {
w[i, j] <- 1
}
}
}

formula <- values ~1+f(1,model = "bym",graph = w)
undebug(INLA:::inla.interpret.formula)
result <- inla(formula, data = locations, family = "gaussian")
print(result)

is my model have wrong?think you!

[finnlindgren]

Your component specification is wrong. You likely need this:

library(INLA)
n <- 10
locations <- data.frame(index = seq_len(n), x = runif(n), y = runif(n))
values <- rnorm(n)
w <- matrix(0, n, n)
for (i in 1:n) {
  for (j in 1:n) {
    dist <- sqrt((locations$x[i] - locations$x[j])^2 + (locations$y[i] - locations$y[j])^2)
    if (dist < 0.2) {
      w[i, j] <- 1
    }
  }
}

formula <- values ~1+f(index,model = "bym",graph = w)
result <- inla(formula, data = locations, family = "gaussian")
print(result)

[tianjinwolaile]

@finnlindgren ,it work,thank you.but i find other problem,my data does't fit gaussian,so i chose poisson,
library(INLA)

set.seed(123)
n <- 100
lambda <- 3
y <- rpois(n, lambda)
data <- data.frame(y = y)

formula <- y ~ 1

result <- inla(formula, data = data, family = "poisson",
control.family = list(hyper = list(prec = list(prior = "gamma", param = c(1, 0.1)))))

i meet those error,the ,
control.family = list(hyper = list(prec = list(prior = "gamma", param = c(1, 0.1))))is shoud be deleted?thank you.i just want set a hyper
Error in inla.core.safe(formula = formula, family = family, contrasts = contrasts, :
Model poisson [ likelihood ], has none hyperparameters, but 'hyper' is list(prec = list(prior = "gamma", param = c(0.5, 0.1)))