Coefficients reported with limited cubic spline in R-band rms

Question

I am trying to understand part of the output of a package rmsin R, where I use limited cubic splines in quantile regression. The problem that I have in understanding the output probably has more in common with a limited cubic spline than with the fact that I use them in quantile regression. I am applying a bounded cubic spline function with k = 3 nodes as follows:

fitted.model <- Rq(y ~ rcs(x, 3), x=TRUE, y=TRUE, tau=0.50, data=d)

I am confused by the fact that

coef(fitted.model) # shows ...

Interception xx '

39.67901296 0.31760226 0.01875437

then

Function(fitted.model) # shows ...

(x = 56.15) {39.679013 + 0.31760226 * x + 1.25554621e-05 * pmax (x-38.45.0) ^ 3-2.3161627e-05 * pmax (x-56.15 , 0) ^ 3 + 1,0607006e -05 * pmax (x-77.1,0) ^ 3}

2.24 2.25 RMS, , & beta; ₀, & beta; ₁ & beta; ₂ ( .. & beta; _k-1), coef(fitted.model), (.. k-1) , Formula(fitted.model). (& beta; ₀) X (& beta; ₁), , , & beta; ₂, coef(fitted.model). ?

Answer 1

, Function coef, - . ( ) " ", , . , rcspline.eval Frank Harrell Hmisc, norm :

norm: 0 , Devlin Weeks (1986), 1 , 2 ( ). norm = 2 -.

, , , 1.2554621 & times; 10 ^-5 0.01875437 (77.1-38.45) ²= 38.65 ²= 1493.82, .. .

; , , x, . , () .