XploRe Help : quantilecl

Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Library:	stats
See also:	quantile cdfcl mediancl

Quantlet:		quantilecl
Description:		Computes the empirical quantiles of a classified data set.

Usage:	`quan = quantilecl(gr, fr, q)`
Input:
	`gr`	m x n grid matrix, each column must be ordered ascendently
	`fr`	(m-1) x n matrix of frequencies for each interval in each column
	`q`	n x 1 vector of levels for the quantiles for each column. It has to hold 0 <= q[i] <=1
Output:
	`quan`	1 x n row vector of quantiles for each column

Note:

Each column of the gr matrix represents a grid (must be ordered ascendently) and each corresponding column of fr matrix represents the numbers of observations in the corresponding intervals (given by gr matrix). E.g., for gr=#(1,2,3)~#(10,15,20) and fr=#(7,4)~#(14,16) the first dataset contains 7 observations in the interval [1,2) and 4 observations in the interval [2,3). The same is true for the second column.

Example:

library("stats") randomize(666) ; make four columns each with 15 random intervals gr = sort(normal(5,4)|5+normal(5,4)|10+normal(5,4),0) gr[,3] = 10 + gr[,3] fr = ceil(30*uniform(14)~20*uniform(14,2)~2*uniform(14)) q = #(0, 1, 0.5, 0.25) quantilecl(gr,fr,q)

Result:

Contents of quan
[1,]  -1.2676   11.271   15.313  0.41783

Author: K. Komorad, W. Haerdle, 20010716. license MD*Tech