Select values around quantiles

Description

The function selects from a set of points along the range of x values these points that are close to a certain quantile of y. First, x values are classified in n groups. Secondly, the specified quantile of y is computed for each group. Thirdly, all y values in a group that are outside +- 0.05 of the estimated quantile are removed. See examples for an illustration.

Usage

1
SelectQuantiles(x, y, q = 0.5, n = NULL, ...)

Arguments

x

vector of x values

y

vector of y values

q

quantile value for which data should be selected

n

number of classes for x

...

further arguments

Author(s)

Matthias Forkel <matthias.forkel@geo.tuwien.ac.at> [aut, cre]

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
# x and y points
x <- 1:1000
y <- x * rnorm(1000, 1, 3)
plot(x, y)

# select points that are close to the median
q05 <- SelectQuantiles(x, y)
points(q05$x, q05$y, col="red")

# select points that are close to the 0.9 quantile
q09 <- SelectQuantiles(x, y, 0.9)
points(q09$x, q09$y, col="blue")

# select points that are close to the 0.1 quantile
q01 <- SelectQuantiles(x, y, 0.1)
points(q01$x, q01$y, col="purple")

# the selected points can be used for fits to a specific quantile
abline(lm(y ~ x, q09), col="blue")
abline(lm(y ~ x, q01), col="purple")