manuscript/bin/support_plots.R
In sahirbhatnagar/sail: Sparse Additive Interaction Learning
## ---- load-support ----
pacman::p_load(psych)
pacman::p_load(here)
load(here::here("manuscript/results/support_methods_comparison_error_summary.RData"))
par(family="serif")
## ---- support-error-cross ----
error.crosses(errorSummary[c(6:10),],
errorSummary[c(1:5),],
labels=unique(errorSummary$group1),
xlab="Number of Active Variables",
main = "SUPPORT Data: Means (+/- 1 SD) from 200 Train/Validate/Test Splits",
sd = TRUE,
cex.lab = 1.4,
cex.axis = 1.4,
cex.main = 1.5,
offset = -0.8,
xlim = c(0, 120),
ylab="Test Set AUC",
colors = sail:::cbbPalette[c(1,3,4,7,2)],
pch=16,cex=1)
## ---- dzclass-interaction ----
load(here::here("manuscript/results/support_sail200Bootstrap.RData"))
par(mfrow=c(1,2), tcl=-0.5, family="serif",
cex.lab = 1.6, font.lab = 1.6, cex.axis = 1.6,
cex.main = .1,
omi=c(0.2,0.2,0,0),
mar = c(4, 4, 1, 1.1) + 0.1)
object = sailfittrain
x = usedatcomplete
varname = "num.co"
xvar = paste0("bs(",varname,", degree = 3)", 1:3)
design = designmat2
s = optLambda
e = usedatcomplete$dzclass
ind <- which(object$vnames %in% xvar)
ctPred <- list()
ARFPred <- list()
for (ite in 1:200) {
allCoefs <- COEFlist[[ite]]
a0 <- allCoefs[1, ]
betas <- as.matrix(allCoefs[object$main.effect.names[ind], , drop = FALSE])
alphas <- as.matrix(allCoefs[object$interaction.names[ind], , drop = FALSE])
betaE <- as.matrix(allCoefs["E", , drop = FALSE])
design.mat.main <- design[, object$main.effect.names[ind], drop = FALSE]
design.mat.int <- design[, object$main.effect.names[ind], drop = FALSE] * e
originalE <- e
originalX <- x[,varname]
f.hat <- drop(originalE * as.vector(betaE) + design.mat.main %*% betas + design.mat.int %*% alphas)
ylims <- range(f.hat)
control = drop(unique(originalE))[1]
ARF = drop(unique(originalE))[2]
cont_index <- which(originalE==control)
ARF_index <- which(originalE==ARF)
cont_pred <- f.hat[cont_index]
ARF_pred <- f.hat[ARF_index]
cont_pred[order(originalX[cont_index])] -> ctPred[[ite]]
ARF_pred[order(originalX[ARF_index])] -> ARFPred[[ite]]
}
matrix(unlist(ctPred),byrow=T,nrow=200) -> ctPredmat
matrix(unlist(ARFPred),byrow=T,nrow=200) -> ARFPredmat
apply(ctPredmat,2,function(x) quantile(x,0.95)) -> ctPredmax
apply(ctPredmat,2,function(x) quantile(x,0.05)) -> ctPredmin
apply(ARFPredmat,2,function(x) quantile(x,0.95)) -> ARFPredmax
apply(ARFPredmat,2,function(x) quantile(x,0.05)) -> ARFPredmin
originalX[cont_index][order(originalX[cont_index])] -> polygonXct
originalX[ARF_index][order(originalX[ARF_index])] -> polygonXarf
allCoefs <- coef(object, s = s)
a0 <- allCoefs[1, ]
betas <- as.matrix(allCoefs[object$main.effect.names[ind], , drop = FALSE])
alphas <- as.matrix(allCoefs[object$interaction.names[ind], , drop = FALSE])
betaE <- as.matrix(allCoefs["E", , drop = FALSE])
design.mat.main <- design[, object$main.effect.names[ind], drop = FALSE]
design.mat.int <- design[, object$main.effect.names[ind], drop = FALSE] * e
originalE <- e # this is the centered E
originalX <- x[,varname]
f.hat <- drop(originalE * as.vector(betaE) + design.mat.main %*% betas + design.mat.int %*% alphas)
ylims <- range(f.hat)
control = drop(unique(originalE))[1]
ARF = drop(unique(originalE))[2]
cont_index <- which(originalE==control)
ARF_index <- which(originalE==ARF)
cont_pred <- f.hat[cont_index]
ARF_pred <- f.hat[ARF_index]
min.length.top <- range(f.hat)[1] ; max.length.top <- range(f.hat)[2]
cbbPalette <- c("#999999", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")
color = cbbPalette[c(6,4,7)]
plot(originalX, f.hat,
pch = 19,
col = color[1],
bty="n",
xlim=c(0,10),
ylim=c(-0.4,3),
xaxt="n",
type = "n",
main = "",
xlab = "Number of comorbidities",
ylab = "Marginal risk")
axis(1, labels = T)
ctPredmat <- ctPredmat[apply(ctPredmat,1,function(x) sum(x > ctPredmax)==0),]
ctPredmat <- ctPredmat[apply(ctPredmat,1,function(x) sum(x < ctPredmin)==0),]
ARFPredmat <- ARFPredmat[apply(ARFPredmat,1,function(x) sum(x > ARFPredmax)==0),]
ARFPredmat <- ARFPredmat[apply(ARFPredmat,1,function(x) sum(x < ARFPredmin)==0),]
rgb(red=0,green=114,blue=178,max=255,alpha=25) -> Col1
rgb(red=0,green=158,blue=115,max=255,alpha=25) -> Col2
apply(ctPredmat,1,function(x) lines(polygonXct,-x,col=Col1,lty=1))
apply(ARFPredmat,1,function(x) lines(polygonXarf,-x,col=Col2,lty=1))
lines(polygonXct, -apply(ctPredmat,2,median), col = color[1], lwd = 3)
lines(polygonXarf, -apply(ARFPredmat,2,median), col = color[2], lwd = 3)
# graphics::rug(originalX, side = 1)
object = sailfittrain
x = usedatcomplete
varname = "age"
xvar = paste0("bs(",varname,", degree = 3)", 1:3)
design = designmat2
s = optLambda
e = usedatcomplete$dzclass
ind <- which(object$vnames %in% xvar)
ctPred <- list()
ARFPred <- list()
for (ite in 1:200) {
allCoefs <- COEFlist[[ite]]
a0 <- allCoefs[1, ]
betas <- as.matrix(allCoefs[object$main.effect.names[ind], , drop = FALSE])
alphas <- as.matrix(allCoefs[object$interaction.names[ind], , drop = FALSE])
betaE <- as.matrix(allCoefs["E", , drop = FALSE])
design.mat.main <- design[, object$main.effect.names[ind], drop = FALSE]
design.mat.int <- design[, object$main.effect.names[ind], drop = FALSE] * e
originalE <- e
originalX <- x[,varname]
f.hat <- drop(originalE * as.vector(betaE) + design.mat.main %*% betas + design.mat.int %*% alphas)
ylims <- range(f.hat)
control = drop(unique(originalE))[1]
ARF = drop(unique(originalE))[2]
cont_index <- which(originalE==control)
ARF_index <- which(originalE==ARF)
cont_pred <- f.hat[cont_index]
ARF_pred <- f.hat[ARF_index]
cont_pred[order(originalX[cont_index])] -> ctPred[[ite]]
ARF_pred[order(originalX[ARF_index])] -> ARFPred[[ite]]
}
matrix(unlist(ctPred),byrow=T,nrow=200) -> ctPredmat
matrix(unlist(ARFPred),byrow=T,nrow=200) -> ARFPredmat
apply(ctPredmat,2,function(x) quantile(x,0.95)) -> ctPredmax
apply(ctPredmat,2,function(x) quantile(x,0.05)) -> ctPredmin
apply(ARFPredmat,2,function(x) quantile(x,0.95)) -> ARFPredmax
apply(ARFPredmat,2,function(x) quantile(x,0.05)) -> ARFPredmin
originalX[cont_index][order(originalX[cont_index])] -> polygonXct
originalX[ARF_index][order(originalX[ARF_index])] -> polygonXarf
allCoefs <- coef(object, s = s)
a0 <- allCoefs[1, ]
betas <- as.matrix(allCoefs[object$main.effect.names[ind], , drop = FALSE])
alphas <- as.matrix(allCoefs[object$interaction.names[ind], , drop = FALSE])
betaE <- as.matrix(allCoefs["E", , drop = FALSE])
design.mat.main <- design[, object$main.effect.names[ind], drop = FALSE]
design.mat.int <- design[, object$main.effect.names[ind], drop = FALSE] * e
originalE <- e # this is the centered E
originalX <- x[,varname]
f.hat <- drop(originalE * as.vector(betaE) + design.mat.main %*% betas + design.mat.int %*% alphas)
ylims <- range(f.hat)
control = drop(unique(originalE))[1]
ARF = drop(unique(originalE))[2]
cont_index <- which(originalE==control)
ARF_index <- which(originalE==ARF)
cont_pred <- f.hat[cont_index]
ARF_pred <- f.hat[ARF_index]
min.length.top <- range(f.hat)[1] ; max.length.top <- range(f.hat)[2]
cbbPalette <- c("#999999", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")
color = cbbPalette[c(6,4,7)]
plot(originalX, f.hat,
pch = 19,
col = color[1],
bty="n",
ylim=c(-0.1,1.3),
xaxt="n",
type = "n",
main = "",
xlab = "Age",
ylab = "")
axis(1, labels = T)
ctPredmat <- ctPredmat[apply(ctPredmat,1,function(x) sum(x > ctPredmax)==0),]
ctPredmat <- ctPredmat[apply(ctPredmat,1,function(x) sum(x < ctPredmin)==0),]
ARFPredmat <- ARFPredmat[apply(ARFPredmat,1,function(x) sum(x > ARFPredmax)==0),]
ARFPredmat <- ARFPredmat[apply(ARFPredmat,1,function(x) sum(x < ARFPredmin)==0),]
rgb(red=0,green=114,blue=178,max=255,alpha=25) -> Col1
rgb(red=0,green=158,blue=115,max=255,alpha=25) -> Col2
apply(ctPredmat,1,function(x) lines(polygonXct,-x,col=Col1,lty=1))
apply(ARFPredmat,1,function(x) lines(polygonXarf,-x,col=Col2,lty=1))
lines(polygonXct, -apply(ctPredmat,2,median), col = color[1], lwd = 3)
lines(polygonXarf, -apply(ARFPredmat,2,median), col = color[2], lwd = 3)
legend("topleft", c("non-ARF/MOSF", "ARF/MOSF"),
col = color[1:3], lwd = c(3,3), cex = 1.6, bty = "n")
# graphics::rug(originalX, side = 1)
sahirbhatnagar/sail documentation built on July 17, 2021, 5:10 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
## ---- load-support ----
pacman::p_load(psych)
pacman::p_load(here)
load(here::here("manuscript/results/support_methods_comparison_error_summary.RData"))
par(family="serif")
## ---- support-error-cross ----
error.crosses(errorSummary[c(6:10),],
errorSummary[c(1:5),],
labels=unique(errorSummary$group1),
xlab="Number of Active Variables",
main = "SUPPORT Data: Means (+/- 1 SD) from 200 Train/Validate/Test Splits",
sd = TRUE,
cex.lab = 1.4,
cex.axis = 1.4,
cex.main = 1.5,
offset = -0.8,
xlim = c(0, 120),
ylab="Test Set AUC",
colors = sail:::cbbPalette[c(1,3,4,7,2)],
pch=16,cex=1)
## ---- dzclass-interaction ----
load(here::here("manuscript/results/support_sail200Bootstrap.RData"))
par(mfrow=c(1,2), tcl=-0.5, family="serif",
cex.lab = 1.6, font.lab = 1.6, cex.axis = 1.6,
cex.main = .1,
omi=c(0.2,0.2,0,0),
mar = c(4, 4, 1, 1.1) + 0.1)
object = sailfittrain
x = usedatcomplete
varname = "num.co"
xvar = paste0("bs(",varname,", degree = 3)", 1:3)
design = designmat2
s = optLambda
e = usedatcomplete$dzclass
ind <- which(object$vnames %in% xvar)
ctPred <- list()
ARFPred <- list()
for (ite in 1:200) {
allCoefs <- COEFlist[[ite]]
a0 <- allCoefs[1, ]
betas <- as.matrix(allCoefs[object$main.effect.names[ind], , drop = FALSE])
alphas <- as.matrix(allCoefs[object$interaction.names[ind], , drop = FALSE])
betaE <- as.matrix(allCoefs["E", , drop = FALSE])
design.mat.main <- design[, object$main.effect.names[ind], drop = FALSE]
design.mat.int <- design[, object$main.effect.names[ind], drop = FALSE] * e
originalE <- e
originalX <- x[,varname]
f.hat <- drop(originalE * as.vector(betaE) + design.mat.main %*% betas + design.mat.int %*% alphas)
ylims <- range(f.hat)
control = drop(unique(originalE))[1]
ARF = drop(unique(originalE))[2]
cont_index <- which(originalE==control)
ARF_index <- which(originalE==ARF)
cont_pred <- f.hat[cont_index]
ARF_pred <- f.hat[ARF_index]
cont_pred[order(originalX[cont_index])] -> ctPred[[ite]]
ARF_pred[order(originalX[ARF_index])] -> ARFPred[[ite]]
}
matrix(unlist(ctPred),byrow=T,nrow=200) -> ctPredmat
matrix(unlist(ARFPred),byrow=T,nrow=200) -> ARFPredmat
apply(ctPredmat,2,function(x) quantile(x,0.95)) -> ctPredmax
apply(ctPredmat,2,function(x) quantile(x,0.05)) -> ctPredmin
apply(ARFPredmat,2,function(x) quantile(x,0.95)) -> ARFPredmax
apply(ARFPredmat,2,function(x) quantile(x,0.05)) -> ARFPredmin
originalX[cont_index][order(originalX[cont_index])] -> polygonXct
originalX[ARF_index][order(originalX[ARF_index])] -> polygonXarf
allCoefs <- coef(object, s = s)
a0 <- allCoefs[1, ]
betas <- as.matrix(allCoefs[object$main.effect.names[ind], , drop = FALSE])
alphas <- as.matrix(allCoefs[object$interaction.names[ind], , drop = FALSE])
betaE <- as.matrix(allCoefs["E", , drop = FALSE])
design.mat.main <- design[, object$main.effect.names[ind], drop = FALSE]
design.mat.int <- design[, object$main.effect.names[ind], drop = FALSE] * e
originalE <- e # this is the centered E
originalX <- x[,varname]
f.hat <- drop(originalE * as.vector(betaE) + design.mat.main %*% betas + design.mat.int %*% alphas)
ylims <- range(f.hat)
control = drop(unique(originalE))[1]
ARF = drop(unique(originalE))[2]
cont_index <- which(originalE==control)
ARF_index <- which(originalE==ARF)
cont_pred <- f.hat[cont_index]
ARF_pred <- f.hat[ARF_index]
min.length.top <- range(f.hat)[1] ; max.length.top <- range(f.hat)[2]
cbbPalette <- c("#999999", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")
color = cbbPalette[c(6,4,7)]
plot(originalX, f.hat,
pch = 19,
col = color[1],
bty="n",
xlim=c(0,10),
ylim=c(-0.4,3),
xaxt="n",
type = "n",
main = "",
xlab = "Number of comorbidities",
ylab = "Marginal risk")
axis(1, labels = T)
ctPredmat <- ctPredmat[apply(ctPredmat,1,function(x) sum(x > ctPredmax)==0),]
ctPredmat <- ctPredmat[apply(ctPredmat,1,function(x) sum(x < ctPredmin)==0),]
ARFPredmat <- ARFPredmat[apply(ARFPredmat,1,function(x) sum(x > ARFPredmax)==0),]
ARFPredmat <- ARFPredmat[apply(ARFPredmat,1,function(x) sum(x < ARFPredmin)==0),]
rgb(red=0,green=114,blue=178,max=255,alpha=25) -> Col1
rgb(red=0,green=158,blue=115,max=255,alpha=25) -> Col2
apply(ctPredmat,1,function(x) lines(polygonXct,-x,col=Col1,lty=1))
apply(ARFPredmat,1,function(x) lines(polygonXarf,-x,col=Col2,lty=1))
lines(polygonXct, -apply(ctPredmat,2,median), col = color[1], lwd = 3)
lines(polygonXarf, -apply(ARFPredmat,2,median), col = color[2], lwd = 3)
# graphics::rug(originalX, side = 1)
object = sailfittrain
x = usedatcomplete
varname = "age"
xvar = paste0("bs(",varname,", degree = 3)", 1:3)
design = designmat2
s = optLambda
e = usedatcomplete$dzclass
ind <- which(object$vnames %in% xvar)
ctPred <- list()
ARFPred <- list()
for (ite in 1:200) {
allCoefs <- COEFlist[[ite]]
a0 <- allCoefs[1, ]
betas <- as.matrix(allCoefs[object$main.effect.names[ind], , drop = FALSE])
alphas <- as.matrix(allCoefs[object$interaction.names[ind], , drop = FALSE])
betaE <- as.matrix(allCoefs["E", , drop = FALSE])
design.mat.main <- design[, object$main.effect.names[ind], drop = FALSE]
design.mat.int <- design[, object$main.effect.names[ind], drop = FALSE] * e
originalE <- e
originalX <- x[,varname]
f.hat <- drop(originalE * as.vector(betaE) + design.mat.main %*% betas + design.mat.int %*% alphas)
ylims <- range(f.hat)
control = drop(unique(originalE))[1]
ARF = drop(unique(originalE))[2]
cont_index <- which(originalE==control)
ARF_index <- which(originalE==ARF)
cont_pred <- f.hat[cont_index]
ARF_pred <- f.hat[ARF_index]
cont_pred[order(originalX[cont_index])] -> ctPred[[ite]]
ARF_pred[order(originalX[ARF_index])] -> ARFPred[[ite]]
}
matrix(unlist(ctPred),byrow=T,nrow=200) -> ctPredmat
matrix(unlist(ARFPred),byrow=T,nrow=200) -> ARFPredmat
apply(ctPredmat,2,function(x) quantile(x,0.95)) -> ctPredmax
apply(ctPredmat,2,function(x) quantile(x,0.05)) -> ctPredmin
apply(ARFPredmat,2,function(x) quantile(x,0.95)) -> ARFPredmax
apply(ARFPredmat,2,function(x) quantile(x,0.05)) -> ARFPredmin
originalX[cont_index][order(originalX[cont_index])] -> polygonXct
originalX[ARF_index][order(originalX[ARF_index])] -> polygonXarf
allCoefs <- coef(object, s = s)
a0 <- allCoefs[1, ]
betas <- as.matrix(allCoefs[object$main.effect.names[ind], , drop = FALSE])
alphas <- as.matrix(allCoefs[object$interaction.names[ind], , drop = FALSE])
betaE <- as.matrix(allCoefs["E", , drop = FALSE])
design.mat.main <- design[, object$main.effect.names[ind], drop = FALSE]
design.mat.int <- design[, object$main.effect.names[ind], drop = FALSE] * e
originalE <- e # this is the centered E
originalX <- x[,varname]
f.hat <- drop(originalE * as.vector(betaE) + design.mat.main %*% betas + design.mat.int %*% alphas)
ylims <- range(f.hat)
control = drop(unique(originalE))[1]
ARF = drop(unique(originalE))[2]
cont_index <- which(originalE==control)
ARF_index <- which(originalE==ARF)
cont_pred <- f.hat[cont_index]
ARF_pred <- f.hat[ARF_index]
min.length.top <- range(f.hat)[1] ; max.length.top <- range(f.hat)[2]
cbbPalette <- c("#999999", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")
color = cbbPalette[c(6,4,7)]
plot(originalX, f.hat,
pch = 19,
col = color[1],
bty="n",
ylim=c(-0.1,1.3),
xaxt="n",
type = "n",
main = "",
xlab = "Age",
ylab = "")
axis(1, labels = T)
ctPredmat <- ctPredmat[apply(ctPredmat,1,function(x) sum(x > ctPredmax)==0),]
ctPredmat <- ctPredmat[apply(ctPredmat,1,function(x) sum(x < ctPredmin)==0),]
ARFPredmat <- ARFPredmat[apply(ARFPredmat,1,function(x) sum(x > ARFPredmax)==0),]
ARFPredmat <- ARFPredmat[apply(ARFPredmat,1,function(x) sum(x < ARFPredmin)==0),]
rgb(red=0,green=114,blue=178,max=255,alpha=25) -> Col1
rgb(red=0,green=158,blue=115,max=255,alpha=25) -> Col2
apply(ctPredmat,1,function(x) lines(polygonXct,-x,col=Col1,lty=1))
apply(ARFPredmat,1,function(x) lines(polygonXarf,-x,col=Col2,lty=1))
lines(polygonXct, -apply(ctPredmat,2,median), col = color[1], lwd = 3)
lines(polygonXarf, -apply(ARFPredmat,2,median), col = color[2], lwd = 3)
legend("topleft", c("non-ARF/MOSF", "ARF/MOSF"),
col = color[1:3], lwd = c(3,3), cex = 1.6, bty = "n")
# graphics::rug(originalX, side = 1)
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.