# Plot a section view of a prediction model or function, including design points if available. ```r ## S3 method for class '`function`' sectionview( fun, vectorized = FALSE, dim = NULL, center = NULL, lty_center = 2, col_center = "black", axis = NULL, npoints = 101, col_fun = if (!is.null(col)) col else "blue", col = NULL, col_fading_interval = 0.5, mfrow = NULL, Xlab = NULL, ylab = NULL, Xlim = NULL, ylim = NULL, title = NULL, add = FALSE, ... ) ## S3 method for class 'matrix' sectionview( X, y, center = NULL, lty_center = 2, col_center = "black", axis = NULL, col_points = if (!is.null(col)) col else "red", col = NULL, col_fading_interval = 0.5, bg_fading = 5, mfrow = NULL, Xlab = NULL, ylab = NULL, Xlim = NULL, ylim = NULL, title = NULL, add = FALSE, ... ) ## S3 method for class 'character' sectionview(eval_str, axis = NULL, mfrow = NULL, ...) ## S3 method for class 'km' sectionview( km_model, type = "UK", center = NULL, axis = NULL, npoints = 101, col_points = if (!is.null(col)) col else "red", col_fun = if (!is.null(col)) col else "blue", col = NULL, conf_level = 0.95, conf_fading = 0.5, bg_fading = 5, mfrow = NULL, Xlab = NULL, ylab = NULL, Xlim = NULL, ylim = NULL, title = NULL, add = FALSE, ... ) ## S3 method for class 'Kriging' sectionview( Kriging_model, center = NULL, axis = NULL, npoints = 101, col_points = if (!is.null(col)) col else "red", col_fun = if (!is.null(col)) col else "blue", col = NULL, conf_level = 0.95, conf_fading = 0.5, bg_fading = 5, mfrow = NULL, Xlab = NULL, ylab = NULL, Xlim = NULL, ylim = NULL, title = NULL, add = FALSE, ... ) ## S3 method for class 'NuggetKriging' sectionview( NuggetKriging_model, center = NULL, axis = NULL, npoints = 101, col_points = if (!is.null(col)) col else "red", col_fun = if (!is.null(col)) col else "blue", col = NULL, conf_level = 0.95, conf_fading = 0.5, bg_fading = 5, mfrow = NULL, Xlab = NULL, ylab = NULL, Xlim = NULL, ylim = NULL, title = NULL, add = FALSE, ... ) ## S3 method for class 'NoiseKriging' sectionview( NoiseKriging_model, center = NULL, axis = NULL, npoints = 101, col_points = if (!is.null(col)) col else "red", col_fun = if (!is.null(col)) col else "blue", col = NULL, conf_level = 0.95, conf_fading = 0.5, bg_fading = 5, mfrow = NULL, Xlab = NULL, ylab = NULL, Xlim = NULL, ylim = NULL, title = NULL, add = FALSE, ... ) ## S3 method for class 'glm' sectionview( glm_model, center = NULL, axis = NULL, npoints = 101, col_points = if (!is.null(col)) col else "red", col_fun = if (!is.null(col)) col else "blue", col = NULL, conf_level = 0.95, conf_fading = 0.5, bg_fading = 5, mfrow = NULL, Xlab = NULL, ylab = NULL, Xlim = NULL, ylim = NULL, title = NULL, add = FALSE, ... ) ## S3 method for class 'list' sectionview( modelFit_model, center = NULL, axis = NULL, npoints = 101, col_points = if (!is.null(col)) col else "red", col_fun = if (!is.null(col)) col else "blue", col = NULL, bg_fading = 5, mfrow = NULL, Xlab = NULL, ylab = NULL, Xlim = NULL, ylim = NULL, title = NULL, add = FALSE, ... ) sectionview(...) ``` ## Arguments - `fun`: a function or 'predict()'-like function that returns a simple numeric, or an interval, or mean and standard error: list(mean=...,se=...). - `vectorized`: is fun vectorized? - `dim`: input variables dimension of the model or function. - `center`: optional coordinates (as a list or data frame) of the center of the section view if the model's dimension is > 2. - `lty_center`: line type for the section center of the plot (if any). - `col_center`: color for the section center of the plot (if any). - `axis`: optional matrix of 2-axis combinations to plot, one by row. The value `NULL` leads to all possible combinations i.e. `choose(D, 2)`. - `npoints`: an optional number of points to discretize plot of response surface and uncertainties. - `col_fun`: color of the function plot. - `col`: color of the object (use col_* for specific objects). - `col_fading_interval`: an optional factor of alpha (color channel) fading used to plot confidence intervals. - `mfrow`: an optional list to force `par(mfrow = ...)` call. The default value `NULL` is automatically set for compact view. - `Xlab`: an optional list of string to overload names for X. - `ylab`: an optional string to overload name for y. - `Xlim`: an optional list to force x range for all plots. The default value `NULL` is automatically set to include all design points. - `ylim`: an optional list to force y range for all plots. - `title`: an optional overload of main title. - `add`: to print graphics on an existing window. - `...`: arguments of the `sectionview.km`, `sectionview.glm`, `sectionview.Kriging` or `sectionview.function` function - `X`: the matrix of input design. - `y`: the array of output values (two columns means an interval). - `col_points`: color of points. - `bg_fading`: an optional factor of alpha (color channel) fading used to plot design points outside from this section. - `eval_str`: the expression to evaluate in each subplot. - `km_model`: an object of class `"km"`. - `type`: the kriging type to use for model prediction. - `conf_level`: an optional list of confidence intervals to display. - `conf_fading`: an optional factor of alpha (color channel) fading used to plot confidence intervals. - `Kriging_model`: an object of class `"Kriging"`. - `NuggetKriging_model`: an object of class `"Kriging"`. - `NoiseKriging_model`: an object of class `"Kriging"`. - `glm_model`: an object of class `"glm"`. - `modelFit_model`: an object returned by DiceEval::modelFit. ## Details If available, experimental points are plotted with fading colors. Points that fall in the specified section (if any) have the color specified `col_points` while points far away from the center have shaded versions of the same color. The amount of fading is determined using the Euclidean distance between the plotted point and `center`. ## Examples ```r x1 <- rnorm(15) x2 <- rnorm(15) y <- x1 + x2 + rnorm(15) model <- lm(y ~ x1 + x2) sectionview(function(x) sum(x), dim=2, center=c(0,0), Xlim=cbind(range(x1),range(x2)), col='black') sectionview(function(x) { x = as.data.frame(x) colnames(x) <- all.vars(model$call)[-1] p = predict.lm(model, newdata=x, se.fit=TRUE) cbind(p$fit-1.96 * p$se.fit, p$fit+1.96 * p$se.fit) }, vectorized=TRUE, dim=2, center=c(0,0), Xlim=cbind(range(x1),range(x2)), add=TRUE) X = matrix(runif(15*2),ncol=2) y = apply(X,1,branin) sectionview(X,y, center=c(.5,.5)) x1 <- rnorm(15) x2 <- rnorm(15) y <- x1 + x2^2 + rnorm(15) model <- glm(y ~ x1 + I(x2^2)) sectionview(model, center=c(.5,.5)) sectionview("abline(h=5)") if (requireNamespace("DiceKriging")) { library(DiceKriging) X = matrix(runif(15*2),ncol=2) y = apply(X,1,branin) model <- km(design = X, response = y, covtype="matern3_2") sectionview(model, center=c(.5,.5)) } if (requireNamespace("rlibkriging")) { library(rlibkriging) X = matrix(runif(15*2),ncol=2) y = apply(X,1,branin) model <- Kriging(X = X, y = y, kernel="matern3_2") sectionview(model, center=c(.5,.5)) } if (requireNamespace("rlibkriging")) { library(rlibkriging) X = matrix(runif(15*2),ncol=2) y = apply(X,1,branin) + 5*rnorm(15) model <- NuggetKriging(X = X, y = y, kernel="matern3_2") sectionview(model, center=c(.5,.5)) } if (requireNamespace("rlibkriging")) { library(rlibkriging) X = matrix(runif(15*2),ncol=2) y = apply(X,1,branin) + 5*rnorm(15) model <- NoiseKriging(X = X, y = y, kernel="matern3_2", noise=rep(5^2,15)) sectionview(model, center=c(.5,.5)) } x1 <- rnorm(15) x2 <- rnorm(15) y <- x1 + x2^2 + rnorm(15) model <- glm(y ~ x1 + I(x2^2)) sectionview(model, center=c(.5,.5)) if (requireNamespace("DiceEval")) { library(DiceEval) X = matrix(runif(15*2),ncol=2) y = apply(X,1,branin) model <- modelFit(X, y, type = "StepLinear") sectionview(model, center=c(.5,.5)) } ## A 2D example - Branin-Hoo function sectionview(branin, center= c(.5,.5), col='black') ## Not run: ## a 16-points factorial design, and the corresponding response d <- 2; n <- 16 design.fact <- expand.grid(seq(0, 1, length = 4), seq(0, 1, length = 4)) design.fact <- data.frame(design.fact); names(design.fact) <- c("x1", "x2") y <- branin(design.fact); names(y) <- "y" if (requireNamespace("DiceKriging")) { library(DiceKriging) ## model: km model <- DiceKriging::km(design = design.fact, response = y) sectionview(model, center= c(.5,.5)) sectionview(branin, center= c(.5,.5), col='red', add=TRUE) } if (requireNamespace("rlibkriging")) { library(rlibkriging) ## model: Kriging model <- Kriging(X = as.matrix(design.fact), y = as.matrix(y), kernel="matern3_2") sectionview(model, center= c(.5,.5)) sectionview(branin, center= c(.5,.5), col='red', add=TRUE) } ## model: glm model <- glm(y ~ 1+ x1 + x2 + I(x1^2) + I(x2^2) + x1*x2, data=cbind(y,design.fact)) sectionview(model, center= c(.5,.5)) sectionview(branin, center= c(.5,.5), col='red', add=TRUE) if (requireNamespace("DiceEval")) { library(DiceEval) ## model: StepLinear model <- modelFit(design.fact, y, type = "StepLinear") sectionview(model, center= c(.5,.5)) sectionview(branin, center= c(.5,.5), col='red', add=TRUE) } ## End(Not run) ``` ## See Also `sectionview.function` for a section plot, and `sectionview3d.function` for a 2D section plot. `sectionview.matrix` for a section plot, and `sectionview3d.matrix` for a 2D section plot. `sectionview.matrix` for a section plot, and `sectionview3d.matrix` for a 2D section plot. `sectionview.km` for a section plot, and `sectionview3d.km` for a 2D section plot. `sectionview.Kriging` for a section plot, and `sectionview3d.Kriging` for a 2D section plot. `sectionview.NuggetKriging` for a section plot, and `sectionview3d.NuggetKriging` for a 2D section plot. `sectionview.NoiseKriging` for a section plot, and `sectionview3d.NoiseKriging` for a 2D section plot. `sectionview.glm` for a section plot, and `sectionview3d.glm` for a 2D section plot. `sectionview.glm` for a section plot, and `sectionview3d.glm` for a 2D section plot. ## Author(s) Yann Richet, IRSN

sectionview function