## Analyzing results ## Computing distances MUD <- PUD <- matrix(NA, nrow = Draws, ncol = SS) for(i in 1:SS){ MUD[, i] <- (UMEU - MU[, i]) / abs(UMEU) PUD[, i] <- (UMEU - PU[, i]) / abs(UMEU) } ## Graph PuDMean <- apply(PUD, 2, function(x) mean(x)) MuDMean <- apply(MUD, 2, function(x) mean(x)) PuDMin <- apply(PUD, 2, function(x) min(x)) MuDMin <- apply(MUD, 2, function(x) min(x)) PuDMax <- apply(PUD, 2, function(x) max(x)) MuDMax <- apply(MUD, 2, function(x) max(x)) ylims <- range(na.omit(c(PuDMax, MuDMax, PuDMin, MuDMin))) plot(cbind(1:SS, PuDMean), type = "p", col = "blue", pch = 17, cex = 1.2, ylim = ylims, ylab = "Relative deviations from 'true' utility", xlab = "Sample Sizes", axes = FALSE) points(1:SS, PuDMax, type = "p", col = "blue", cex = 1.2, pch = 22) points(1:SS, PuDMin, type = "p", col = "blue", cex = 1.2, pch = 22) points(1:SS, MuDMean, type = "p", col = "darkred", cex = 1.2, pch = 19) points(1:SS, MuDMax, type = "p", col = "darkred", cex = 1.2, pch = 22) points(1:SS, MuDMin, type = "p", col = "darkred", cex = 1.2, pch = 22) arrows(x0 = 1:SS, y0 = PuDMin, y1 = PuDMax, code = 3, col = "blue", length = 0.0, angle = 90, lwd = 2) arrows(x0 = 1:SS, y0 = MuDMin, y1 = MuDMax, code = 3, col = "darkred", length = 0.0, angle = 90, lwd = 1.0) axis(1, at = 1:SS, Samples) axis(2, pretty(ylims), pretty(ylims), las = 2) box() legend("topright", legend = c("PU mean deviation", "PU min/max deviation", "MU mean deviation", "MU min/max deviation"), pch = c(17, 22, 19, 22), col = c(rep("blue", 2), rep("darkred", 2))) abline(h = 0, col = "gray")