Monthly Archives: April 2014

Duck performance

So, here it is:

Below you'll see how the groups located the duck from the respective corner points (colour coded by origin).

Seems that there is still some room for improvement…. especially precision seems to be a little low (it is hard to judge accuracy, as we don't have an absolute reference for the location of the duck).

p <- xyplot(duckies.y ~ duckies.x | group, groups = origin, data = duckies,
            as.table = TRUE, auto.key = TRUE, asp = "iso",
            xlim = range(corners$x), ylim = range(corners$y))
corners[5, ] <- corners[1, ]
l <- xyplot(corners$y ~ corners$x, type = "l", asp = "iso")
p + as.layer(l)

plot of chunk unnamed-chunk-2

The implications of errors like these could be that we locate trees where they are not and that we then see patterns that are not real and draw wrong conclusions from the statistical analysis (Ripley's K etc…).

So please, in the forest, be as accurate AND precise as possible with your measurements.

In case you are confused about the difference between accuracy and precision, here's a little image to clarify:

accuracy vs. precision

(image from: