predict.RNAmf: prediction of the RNAmf emulator with multiple fidelity levels.

Description

The function computes the posterior mean and variance of RNA models with multiple fidelity levels by fitted model from RNAmf.

Usage

# S3 method for RNAmf
predict(object, x = NULL, nimpute = 50, ...)

Value

mu: A list of vectors containing the predictive posterior mean at each fidelity level.
sig2: A list of vectors containing the predictive posterior variance at each fidelity level.
time: A scalar indicating the computation time.

Arguments

object: An object of class RNAmf fitted by RNAmf.
x: A vector or matrix of new input locations for prediction.
nimpute: Number of imputations for non-nested designs. Default is 50.
...: Additional arguments for compatibility with generic method predict.

Details

The predict.RNAmf function internally calls closed_form_RNA to recursively compute the closed-form posterior mean and variance at each level.

From the fitted model from RNAmf, the posterior mean and variance are calculated based on the closed-form expression derived by a recursive fashion. The formulas depend on its kernel choices. For further details, see Heo and Sung (2025, <tools:::Rd_expr_doi("https://doi.org/10.1080/00401706.2024.2376173")>).

Examples

Run this code

# \donttest{
### two levels example ###

### Perdikaris function ###
f1 <- function(x) {
  sin(8 * pi * x)
}

f2 <- function(x) {
  (x - sqrt(2)) * (sin(8 * pi * x))^2
}

### training data ###
n1 <- 13
n2 <- 8

### fix seed to reproduce the result ###
set.seed(1)

### generate initial nested design ###
X <- NestedX(c(n1, n2), 1)
X1 <- X[[1]]
X2 <- X[[2]]

y1 <- f1(X1)
y2 <- f2(X2)

### n=100 uniform test data ###
x <- seq(0, 1, length.out = 100)

### fit an RNAmf ###
fit.RNAmf <- RNAmf(list(X1, X2), list(y1, y2), kernel = "sqex", constant=TRUE)

### predict ###
predy <- predict(fit.RNAmf, x)$mu[[2]]
predsig2 <- predict(fit.RNAmf, x)$sig2[[2]]

### RMSE ###
print(sqrt(mean((predy - f2(x))^2)))

### visualize the emulation performance ###
plot(x, predy,
  type = "l", lwd = 2, col = 3, # emulator and confidence interval
  ylim = c(-2, 1)
)
lines(x, predy + 1.96 * sqrt(predsig2 * length(y2) / (length(y2) - 2)), col = 3, lty = 2)
lines(x, predy - 1.96 * sqrt(predsig2 * length(y2) / (length(y2) - 2)), col = 3, lty = 2)

curve(f2(x), add = TRUE, col = 1, lwd = 2, lty = 2) # high fidelity function

points(X1, y1, pch = 1, col = "red") # low-fidelity design
points(X2, y2, pch = 4, col = "blue") # high-fidelity design

### three levels example ###
### Branin function ###
branin <- function(xx, l){
  x1 <- xx[1]
  x2 <- xx[2]
  if(l == 1){
    10*sqrt((-1.275*(1.2*x1+0.4)^2/pi^2+5*(1.2*x1+0.4)/pi+(1.2*x2+0.4)-6)^2 +
    (10-5/(4*pi))*cos((1.2*x1+0.4))+ 10) + 2*(1.2*x1+1.9) - 3*(3*(1.2*x2+2.4)-1) - 1 - 3*x2 + 1
  }else if(l == 2){
    10*sqrt((-1.275*(x1+2)^2/pi^2+5*(x1+2)/pi+(x2+2)-6)^2 +
    (10-5/(4*pi))*cos((x1+2))+ 10) + 2*(x1-0.5) - 3*(3*x2-1) - 1
  }else if(l == 3){
    (-1.275*x1^2/pi^2+5*x1/pi+x2-6)^2 + (10-5/(4*pi))*cos(x1)+ 10
  }
}

output.branin <- function(x, l){
  factor_range <- list("x1" = c(-5, 10), "x2" = c(0, 15))

  for(i in 1:length(factor_range)) x[i] <- factor_range[[i]][1] + x[i] * diff(factor_range[[i]])
  branin(x[1:2], l)
}

### training data ###
n1 <- 20; n2 <- 15; n3 <- 10

### fix seed to reproduce the result ###
set.seed(1)

### generate initial nested design ###
X <- NestedX(c(n1, n2, n3), 2)
X1 <- X[[1]]
X2 <- X[[2]]
X3 <- X[[3]]

y1 <- apply(X1,1,output.branin, l=1)
y2 <- apply(X2,1,output.branin, l=2)
y3 <- apply(X3,1,output.branin, l=3)

### n=10000 grid test data ###
x <- as.matrix(expand.grid(seq(0, 1, length.out = 100),seq(0, 1, length.out = 100)))

### fit an RNAmf ###
fit.RNAmf <- RNAmf(list(X1, X2, X3), list(y1, y2, y3), kernel = "sqex", constant=TRUE)

### predict ###
pred.RNAmf <- predict(fit.RNAmf, x)
predy <- pred.RNAmf$mu[[3]]
predsig2 <- pred.RNAmf$sig2[[3]]

### RMSE ###
print(sqrt(mean((predy - apply(x,1,output.branin, l=3))^2)))

### visualize the emulation performance ###
x1 <- x2 <- seq(0, 1, length.out = 100)
oldpar <- par(mfrow=c(1,2))
image(x1, x2, matrix(apply(x,1,output.branin, l=3), ncol=100),
zlim=c(0,310), main="Branin function")
image(x1, x2, matrix(predy, ncol=100),
zlim=c(0,310), main="RNAmf prediction")
par(oldpar)

### predictive variance ###
print(predsig2)# }