pkg/R/plot_valse.R

   1 #' Plot
   2 #'
   3 #' It is a function which plots relevant parameters
   4 #'
   5 #' @param X matrix of covariates (of size n*p)
   6 #' @param Y matrix of responses (of size n*m)
   7 #' @param model the model constructed by valse procedure
   8 #' @param n sample size
   9 #' @return several plots
  10 #'
  11 #' @export
  12 plot_valse <- function(X, Y, model, n, comp = FALSE, k1 = NA, k2 = NA)
  13 {
  14   require("gridExtra")
  15   require("ggplot2")
  16   require("reshape2")
  17   require("cowplot")
  18
  19   K <- length(model$pi)
  20   ## regression matrices
  21   gReg <- list()
  22   for (r in 1:K)
  23   {
  24     Melt <- melt(t((model$phi[, , r])))
  25     gReg[[r]] <- ggplot(data = Melt, aes(x = Var1, y = Var2, fill = value)) +
  26       geom_tile() + scale_fill_gradient2(low = "blue", high = "red", mid = "white",
  27       midpoint = 0, space = "Lab") + ggtitle(paste("Regression matrices in cluster", r))
  28   }
  29   print(gReg)
  30
  31   ## Differences between two clusters
  32   if (comp)
  33   {
  34     if (is.na(k1) || is.na(k))
  35       print("k1 and k2 must be integers, representing the clusters you want to compare")
  36     Melt <- melt(t(model$phi[, , k1] - model$phi[, , k2]))
  37     gDiff <- ggplot(data = Melt, aes(x = Var1, y = Var2, fill = value))
  38       + geom_tile()
  39       + scale_fill_gradient2(low = "blue", high = "red", mid = "white", midpoint = 0,
  40         space = "Lab")
  41       + ggtitle(paste("Difference between regression matrices in cluster",
  42         k1, "and", k2))
  43     print(gDiff)
  44   }
  45
  46   ### Covariance matrices
  47   matCov <- matrix(NA, nrow = dim(model$rho[, , 1])[1], ncol = K)
  48   for (r in 1:K)
  49     matCov[, r] <- diag(model$rho[, , r])
  50   MeltCov <- melt(matCov)
  51   gCov <- ggplot(data = MeltCov, aes(x = Var1, y = Var2, fill = value)) + geom_tile()
  52     + scale_fill_gradient2(low = "blue", high = "red", mid = "white", midpoint = 0,
  53       space = "Lab")
  54     + ggtitle("Covariance matrices")
  55   print(gCov)
  56
  57   ### Proportions
  58   gam2 <- matrix(NA, ncol = K, nrow = n)
  59   for (i in 1:n)
  60     gam2[i, ] <- c(model$proba[i, model$affec[i]], model$affec[i])
  61
  62   bp <- ggplot(data.frame(gam2), aes(x = X2, y = X1, color = X2, group = X2))
  63     + geom_boxplot()
  64     + theme(legend.position = "none")
  65     + background_grid(major = "xy", minor = "none")
  66   print(bp)
  67
  68   ### Mean in each cluster
  69   XY <- cbind(X, Y)
  70   XY_class <- list()
  71   meanPerClass <- matrix(0, ncol = K, nrow = dim(XY)[2])
  72   for (r in 1:K)
  73   {
  74     XY_class[[r]] <- XY[model$affec == r, ]
  75     if (sum(model$affec == r) == 1) {
  76       meanPerClass[, r] <- XY_class[[r]]
  77     } else {
  78       meanPerClass[, r] <- apply(XY_class[[r]], 2, mean)
  79     }
  80   }
  81   data <- data.frame(mean = as.vector(meanPerClass),
  82     cluster = as.character(rep(1:K, each = dim(XY)[2])), time = rep(1:dim(XY)[2], K))
  83   g <- ggplot(data, aes(x = time, y = mean, group = cluster, color = cluster))
  84   print(g + geom_line(aes(linetype = cluster, color = cluster))
  85     + geom_point(aes(color = cluster)) + ggtitle("Mean per cluster"))
  86 }