| Group: | Statistical Data Analysis |
| Topic: | Nonparametric Methods |
| See also: | sker isoreg rmed lowess |
| Function: | l1line | |
| Description: | l1line computes the least absolute deviation line from scatterplot data. It gives the estimate b0 and b1 that minimizes sum_i=1,n |y_i - b0 - b1 x_i |. |
| Usage: | b = l1line (x, y) | |
| Input: | ||
| x | n x 1 vector | |
| y | n x 1 vector | |
| Output: | ||
| b | 2 x 1 vector | |
; read the geyser data
x = read ("geyser")
y=x[,2]
x=x[,1]
; compute the estimate
b = l1line (x, y)
b
gives coefficient b0=34.73 and b1=10.417.
; compute the least square line
yh = (matrix(rows(x))~x[,1])*b
; create a display for plotting
d = createdisplay(1,1)
; show the data and the fitted line
x = read ("geyser")
show (d, 1, 1, x~y, x~yh)
| Group: | Statistical Data Analysis |
| Topic: | Nonparametric Methods |
| See also: | sker isoreg rmed lowess |