The actual warning: "warning: :RHO and :TRIES are not a known argument keywords."
;;
;; Contributed by Joel J. Adamson adamsonj@email.unc.edu
;;
;; 2012-12-28: Fix the declaim, avoid incf before tail-recursion, and use multiple return values instead
;; of a list. The last change modifies the interface, but, then again, the list was never compatible with the
;; declaim. Paul Khuong
(declaim (ftype (function (double-float
double-float
&key (:rho double-float)
(:error-limit integer)
(:tries integer))
(values double-float double-float &optional))
random-normal-bivariate))
(defun random-normal-bivariate (sigma-x sigma-y &key (rho 0.0d0) (error-limit 500) (tries 0))
"Return a pair of numbers with specific correlation coefficent rho
and with specified variances sigma-x and sigma-y; a direct port of
gsl_ran_bivariate_gaussian from the GNU Scientific Library:
do
{
/* choose x,y in uniform square (-1,-1) to (+1,+1) */
u = -1 + 2 * gsl_rng_uniform (r);
v = -1 + 2 * gsl_rng_uniform (r);
/* see if it is in the unit circle */
r2 = u * u + v * v;
}
while (r2 > 1.0 || r2 == 0);
scale = sqrt (-2.0 * log (r2) / r2);
_x = sigma_x * u * scale;
y = sigma_y * (rho * u + sqrt(1 - rho_rho) * v) \ scale;
}
I need a good test for this one.
"
(let* ((u (- 1.0d0 (+ (* 2.0d0 (random 1.0d0)))))
;; give me a pair of random numbers from the built-in
;; uniform random number generator
;;
;; dividing by 100.0 ensures that I get a number
;; between 0 and 1
(v (- 1.0 (+ (* 2.0 (random 1.0d0)))))
;; the next two terms test to see if the pair is
;; within the unit circle
(radius (+ (* u u) (* v v)))
(scale (sqrt (/ (* -2.0d0 (log radius)) radius))))
;; if this pair is not within the unit circle, try again
(cond ((or (not (<= radius 1.0d0))
(= radius 0.0d0))
(if (< tries error-limit)
;; increment the number of tries
;;
;; tries should not be specified by the user
(random-normal-bivariate sigma-x sigma-y :rho rho
:tries (1+ tries))
;; too many tries --- sorry!
(error "Unsuccessful bivariate-gaussian after ~S tries" tries)))
;; otherwise return the two values, calculated as in
;; the C-code above
(t (values (* sigma-x u scale)
(* sigma-y (+ (* rho u)
(* v
(sqrt (- 1.0d0
(* rho rho)))))
scale))))))
fmitha
commented
11 years ago
I'm getting an error when compiling with sbcl 1.1.2 in Debian squeeze.
See http://paste.lisp.org/display/134316 for the error (second annotation) and suggested fix (first annotation). Pasted also below.
Error is:
; file: /usr/local/share/common-lisp/source/cl-randist/normal.lisp ; in: DEFUN RANDOM-NORMAL-BIVARIATE ; (RANDOM-DISTRIBUTIONS:RANDOM-NORMAL-BIVARIATE RANDOM-DISTRIBUTIONS::SIGMA-X ; RANDOM-DISTRIBUTIONS::SIGMA-Y :RHO RANDOM-DISTRIBUTIONS::RHO :TRIES ; (INCF RANDOM-DISTRIBUTIONS::TRIES)) ; ; caught WARNING: ; :RHO and :TRIES are not a known argument keywords.
Suggested fix is:
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
The actual warning: "warning: :RHO and :TRIES are not a known argument keywords."
;; ;; Contributed by Joel J. Adamson adamsonj@email.unc.edu ;; ;; 2012-12-28: Fix the declaim, avoid incf before tail-recursion, and use multiple return values instead ;; of a list. The last change modifies the interface, but, then again, the list was never compatible with the ;; declaim. Paul Khuong
(declaim (ftype (function (double-float double-float &key (:rho double-float) (:error-limit integer) (:tries integer)) (values double-float double-float &optional)) random-normal-bivariate)) (defun random-normal-bivariate (sigma-x sigma-y &key (rho 0.0d0) (error-limit 500) (tries 0)) "Return a pair of numbers with specific correlation coefficent rho and with specified variances sigma-x and sigma-y; a direct port of gsl_ran_bivariate_gaussian from the GNU Scientific Library:
void gsl_ran_bivariate_gaussian (const gsl_rng * r, double sigma_x, double sigma_y, double rho, double x, double y) { double u, v, r2, scale;
do { /* choose x,y in uniform square (-1,-1) to (+1,+1) */
} while (r2 > 1.0 || r2 == 0);
scale = sqrt (-2.0 * log (r2) / r2);
_x = sigma_x * u * scale; y = sigma_y * (rho * u + sqrt(1 - rho_rho) * v) \ scale; }
I need a good test for this one. " (let* ((u (- 1.0d0 (+ (* 2.0d0 (random 1.0d0))))) ;; give me a pair of random numbers from the built-in ;; uniform random number generator ;; ;; dividing by 100.0 ensures that I get a number ;; between 0 and 1 (v (- 1.0 (+ (* 2.0 (random 1.0d0))))) ;; the next two terms test to see if the pair is ;; within the unit circle (radius (+ (* u u) (* v v))) (scale (sqrt (/ (* -2.0d0 (log radius)) radius)))) ;; if this pair is not within the unit circle, try again (cond ((or (not (<= radius 1.0d0)) (= radius 0.0d0)) (if (< tries error-limit) ;; increment the number of tries ;; ;; tries should not be specified by the user (random-normal-bivariate sigma-x sigma-y :rho rho :tries (1+ tries)) ;; too many tries --- sorry! (error "Unsuccessful bivariate-gaussian after ~S tries" tries))) ;; otherwise return the two values, calculated as in ;; the C-code above (t (values (* sigma-x u scale) (* sigma-y (+ (* rho u) (* v (sqrt (- 1.0d0 (* rho rho))))) scale))))))