What is Newton's method when there is a lot of input?
2 Ansichten (letzte 30 Tage)
Ältere Kommentare anzeigen
I am trying to estimate 3parameters of mixed weibull distribution. And I think the Nweton method for it.
But, error happend
'Error: exit :There are too many output arguments.
Error: etimation_weibull> accelerated_newton (line 30)
if ~ exit ('n_iter','var')
Error: etimation_weibull (line 21)
[x, x_err, x_h] = accelerated_newton (func, dfunc, x0, conv_tol, n_iter);
I think that the error is related to the part where the @ function is used. Is it not possible if there are multiple inputs?
Also, is there any other method for estimating the three parameters of the mixed Weibull distribution?
%% Estimation 3parameters in weibull distribution
load('weibul.mat')
X = Euclidean_distance; % input data
%% Define first-order and second-order derivatives of Weibull distribution
a = 0.1;
b = 0.1;
c = 0.1;
func = @(x,a,b,c) 1-exp((x-c)/a).^b;
dfunc = @(x,a,b,c) (x>c).*(b/a).*(((x-c)/a).^(b-1)).*exp(-(x-c)/a).^b;
x0 = Euclidean_distance(1);
conv_tol = 1e-10;
n_iter = 30;
[x, x_err, x_h] = accelerated_newton(func,dfunc,x0,conv_tol,n_iter); %● error happend
%% newton-methods function
function [x, x_err, x_h] = accelerated_newton(func,dfunc,x0,conv_tol,n_iter)
if ~isa(func, 'function_handle') || ~isa(dfunc,'function_handle')
error('function must be a handle (add"@" before the function name')
end
if ~exit('n_iter', 'var')
n_iter = 30;
end
if ~exist('conv_tol', 'var')
conv_tol = 1e-15;
end
x = x0;
x_h = nan(n_iter + 1, 1);
x_h(1) = x;
for k=1:n_iter
f = func(x);
df = dfunc(x);
delta = f./df;
if k==1
df_23 = dfunc(x - 2.*delta./3);
FF = 0.5 + sqrt(max(0 ,0.75.*df_23 - 0.5));
else
d2f = (4.*df + 2.*df_old - 6.*(f - f_old)./(x - x_old))./(x-x_old);
A2_hat = -.5.*d2f./df;
FF = 0.5 + sqrt(max(0, 0.25 - A2_hat.*delta));
end
x_old = x;
x = x_old - delta./FF;
x_h(k+1) = x;
x_err = abs(x_old - x);
if ((x_err < conv_tol) && (abs(f) - conv_tol))
x_h(k+2:n_iter+1) = [];
break;
end
df_old = df;
f_old = f;
if k == n_iter
disp('The function did not converge')
end
end
end
0 Kommentare
Antworten (0)
Siehe auch
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!