TensorTest2D (version 1.1.2)

plot.tsglm: Plot Effective Image Pixels for A "tsglm" Object

Description

plot method for self-defined class "tsglm".

Usage

# S3 method for tsglm
plot(
  x,
  method = p.adjust.methods,
  alpha = NULL,
  type = c("coef", "tval"),
  background = NULL,
  showlabels = TRUE,
  plot.legend = TRUE,
  ...
)

Arguments

x

an object of class "tsglm".

method

p-value correction method. See p.adjust. Its default value is "none".

alpha

double. The value of significance level. Its default value is NULL. If specify real-valued alpha, for example 0.05, the image plot marks the pixels with p-values smaller than alpha.

type

string. The type of values shown on the image pixels when background = NULL. Set type = 'coef' for showing the values of estimated coefficients of the \(B\) matrix. Set type = 'tval' for showing the t-statistics of the coefficients of the \(B\) matrix. If background is not NULL, the plot will neglect the choice for type and show the background image as per user's interest.

background

an image data that used as the background of the effectiveness markers. If background = NULL, the background color shows the effect size of the each pixel according to the setting in type.

showlabels

boolean. For showlabels = TRUE, if the row and column names of the image exist, the row and column names are shown on the sides of the image plot; otherwise, the row and column indices are shown.

plot.legend

boolean. Set plot.legend = TRUE if the colorbar legend is needed. The default is TRUE.

...

further arguments passed to the image function.

Author

Ping-Yang Chen

See Also

tensorReg2D, draw.coef

Examples

# Simulation data
n <- 500 # number of observations
n_P <- 3; n_G <- 16 # dimension of 3-D tensor variables.
n_d <- 1 # number of numerical variable, if n_d == 1,  numerical variable equals to intercept.
beta_True <- rep(1, n_d)
B_True <- c(1, 1, 1)%*%t(rnorm(n_G)) + c(0, .5, .5)%*%t(rnorm(n_G))
B_True <- B_True / 10
W <- matrix(rnorm(n*n_d), n, n_d); W[,1] <- 1
X <- array(rnorm(n*n_P*n_G), dim=c(n_P, n_G, n))

# Binomial Responses
p_B <- exp(W%*%beta_True + X%hp%B_True); p_B <- p_B/(1+p_B)
y_B <- rbinom(n, 1, p_B)
DATA_B <- list(y = y_B, W = W, X = X)

# Binomial Model
result_B <- tensorReg2D(y = DATA_B$y, X = DATA_B$X, W=NULL, n_R = 1,
family = "binomial", opt = 1, max_ite = 100, tol = 10^(-7) )

# Plot the effect size of the pixels
plot(result_B, method = "fdr", alpha = 0.05, type = "coef")
# Plot the t-statistics of the coefficients of the pixels
plot(result_B, method = "fdr", alpha = 0.05, type = "tval")

# Plot the effective pixels with data image as the background
x0 <- DATA_B$X[,,which(DATA_B$y == 0)]
x1 <- DATA_B$X[,,which(DATA_B$y == 1)]
m0 <- m1 <- matrix(0, dim(DATA_B$X)[1], dim(DATA_B$X)[2])
for (i in 1:dim(x0)[3]) m0 <- m0 + x0[,,i]/dim(x0)[3]
for (i in 1:dim(x1)[3]) m1 <- m1 + x1[,,i]/dim(x1)[3]
par(mfrow = c(1, 2), mar = c(2, 2, 2, 2))
plot(result_B, method = "fdr", alpha = 0.05,
background = m0, col = gray(seq(0, 1, 0.05)))
title("Category 0")
plot(result_B, method = "fdr", alpha = 0.05,
background = m1, col = gray(seq(0, 1, 0.05)))
title("Category 1")