ERROR:NOT ENOUGH INPUT ARGUMENT IN LINE n = length(y);

% parse arguments
params = {2, 0.7, 10, 8e-4, 1e-2, 'quadprog'};
i = ~cellfun(@isempty, varargin);
params(i) = varargin(i);
[tau0, tau1, delta_knot, alpha, gamma, solver] = deal(params{:});

n = length(y);
y = y(:);

% bateman ARMA model
a1 = 1/min(tau1, tau0); % a1 > a0
a0 = 1/max(tau1, tau0);
ar = [(a1*delta + 2) * (a0*delta + 2), 2*a1*a0*delta^2 - 8, ...
       (a1*delta - 2) * (a0*delta - 2)] / ((a1 - a0) * delta^2);
ma = [1 2 1];

% matrices for ARMA model
i = 3:n;
A = sparse([i i i], [i i-1 i-2], repmat(ar, n-2, 1), n, n);
M = sparse([i i i], [i i-1 i-2], repmat(ma, n-2, 1), n, n);

% spline
delta_knot_s = round(delta_knot / delta);
spl = [1:delta_knot_s delta_knot_s-1:-1:1]'; % order 1
spl = conv(spl, spl, 'full');
spl = spl / max(spl);
% matrix of spline regressors
i = bsxfun(@plus, (0:length(spl)-1)'-floor(length(spl)/2), 1:delta_knot_s:n);
nB = size(i, 2);
j = repmat(1:nB, length(spl), 1);
p = repmat(spl(:), 1, nB);
valid = i >= 1 & i <= n;
B = sparse(i(valid), j(valid), p(valid));

% trend
C = [ones(n,1) (1:n)'/n];
nC = size(C, 2);

% Solve the problem:
% .5*(M*q + B*l + C*d - y)^2 + alpha*sum(A,1)*p + .5*gamma*l'*l
% s.t. A*q >= 0

if strcmpi(solver, 'quadprog')
    % Use Matlab's quadprog
    H = [M'*M, M'*C, M'*B; C'*M, C'*C, C'*B; B'*M, B'*C, B'*B+gamma*speye(nB)];
    f = [alpha*sum(A,1)'-M'*y; -(C'*y); -(B'*y)];

    [z, obj] = quadprog(H, f, [-A zeros(n,length(f)-n)], zeros(n, 1), ...
        [], [], [], [], [], optimset('Algorithm', 'interior-point-convex', ...
        'TolFun', 1e-13));
    %z = qp([], H, f, [], [], [], [], zeros(n,1), [A zeros(n,length(f)-n)], []); 
    obj = obj + .5 * (y' * y);
elseif strcmpi(solver, 'sedumi')
    % Use SeDuMi 
    U = [A, sparse(n,nC), -speye(n), sparse(n,n+nB+4); ...
         M, C, sparse(n,n+2), -speye(n), sparse(n,2), B; ...
         sparse(1,2*n+nC), 1, sparse(1,n+nB+3); ...
         sparse(1,3*n+nC+2), 1, sparse(1,nB+1)];
    b = [sparse(n,1); y; 1; 1];
    c = sparse([n+nC+(1:n), 2*n+nC+2, 3*n+nC+4], ...
               1, [alpha*ones(1,n), 1, gamma], 3*n+nC+nB+4, 1);
    K = struct('f', n+nC, 'l', n, 'r', [2+n 2+nB]);
    pars.eps = 1e-6;
    pars.chol.maxuden = 1e2;
    z = sedumi(U, b, c, K, pars);
    obj = c' * z;
    %objd = b' * s;
end

l = z(end-nB+1:end);
d = z(n+1:n+nC);
t = B*l + C*d;
q = z(1:n);
p = A * q;
r = M * q;
e = y - r - t;

end

Answer 1

[Cris LaPierre]

Based on what you've shared, you are trying to take the length of a variable that doesn't yet exist (y). That would give a different error, though, so perhaps you've left something out?

I suggest also sharing the entire error message (all the red text).

[NUR BATRISYIA HANNANI ZAIN AZMI]

function [r, p, t, l, d, e, obj] = cvxEDA(y, delta, varargin)
%CVXEDA Convex optimization approach to electrodermal activity processing
%   This function implements the cvxEDA algorithm described in "cvxEDA: a
%   Convex Optimization Approach to Electrodermal Activity Processing"
%   (http://dx.doi.org/10.1109/TBME.2015.2474131 also available from the
%   authors' homepages).
%
%   Syntax:
%   [r, p, t, l, d, e, obj] = cvxEDA(y, delta, tau0, tau1, delta_knot,
%                                    alpha, gamma, solver)
%
%   where:
%      y: observed EDA signal (we recommend normalizing it: y = zscore(y))
%      delta: sampling interval (in seconds) of y
%      tau0: slow time constant of the Bateman function (default 2.0)
%      tau1: fast time constant of the Bateman function (default 0.7)
%      delta_knot: time between knots of the tonic spline function (default 10)
%      alpha: penalization for the sparse SMNA driver (default 0.0008)
%      gamma: penalization for the tonic spline coefficients (default 0.01)
%      solver: sparse QP solver to be used, 'quadprog' (default) or 'sedumi'
%
%   returns (see paper for details):
%      r: phasic component
%      p: sparse SMNA driver of phasic component
%      t: tonic component
%      l: coefficients of tonic spline
%      d: offset and slope of the linear drift term
%      e: model residuals
%      obj: value of objective function being minimized (eq 15 of paper)

% ______________________________________________________________________________
%
% File:                         cvxEDA.m
% Last revised:                 22 Oct 2015 r68
% ______________________________________________________________________________
%
% Copyright (C) 2014-2015 Luca Citi, Alberto Greco
%
% This program is free software; you can redistribute it and/or modify it under
% the terms of the GNU General Public License as published by the Free Software
% Foundation; either version 3 of the License, or (at your option) any later
% version.
%
% This program is distributed in the hope that it will be useful, but WITHOUT
% ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
% FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
%
% You may contact the author by e-mail ([e-mail]).
% ______________________________________________________________________________
%
% This method was first proposed in:
% A Greco, G Valenza, A Lanata, EP Scilingo, and L Citi
% "cvxEDA: a Convex Optimization Approach to Electrodermal Activity Processing"
% IEEE Transactions on Biomedical Engineering, 2015
% DOI: 10.1109/TBME.2015.2474131
%
% If you use this program in support of published research, please include a
% citation of the reference above. If you use this code in a software package,
% please explicitly inform the end users of this copyright notice and ask them
% to cite the reference above in their published research.
% ______________________________________________________________________________

% parse arguments
params = {2, 0.7, 10, 8e-4, 1e-2, 'quadprog'};
i = ~cellfun(@isempty, varargin);
params(i) = varargin(i);
[tau0, tau1, delta_knot, alpha, gamma, solver] = deal(params{:});
n = length(y);
y = y(:);

% bateman ARMA model
a1 = 1/min(tau1, tau0); % a1 > a0
a0 = 1/max(tau1, tau0);
ar = [(a1*delta + 2) * (a0*delta + 2), 2*a1*a0*delta^2 - 8, ...
       (a1*delta - 2) * (a0*delta - 2)] / ((a1 - a0) * delta^2);
ma = [1 2 1];

% matrices for ARMA model
i = 3:n;
A = sparse([i i i], [i i-1 i-2], repmat(ar, n-2, 1), n, n);
M = sparse([i i i], [i i-1 i-2], repmat(ma, n-2, 1), n, n);

% spline
delta_knot_s = round(delta_knot / delta);
spl = [1:delta_knot_s delta_knot_s-1:-1:1]'; % order 1
spl = conv(spl, spl, 'full');
spl = spl / max(spl);
% matrix of spline regressors
i = bsxfun(@plus, (0:length(spl)-1)'-floor(length(spl)/2), 1:delta_knot_s:n);
nB = size(i, 2);
j = repmat(1:nB, length(spl), 1);
p = repmat(spl(:), 1, nB);
valid = i >= 1 & i <= n;
B = sparse(i(valid), j(valid), p(valid));

% trend
C = [ones(n,1) (1:n)'/n];
nC = size(C, 2);

% Solve the problem:
% .5*(M*q + B*l + C*d - y)^2 + alpha*sum(A,1)*p + .5*gamma*l'*l
% s.t. A*q >= 0

if strcmpi(solver, 'quadprog')
    % Use Matlab's quadprog
    H = [M'*M, M'*C, M'*B; C'*M, C'*C, C'*B; B'*M, B'*C, B'*B+gamma*speye(nB)];
    f = [alpha*sum(A,1)'-M'*y; -(C'*y); -(B'*y)];

    [z, obj] = quadprog(H, f, [-A zeros(n,length(f)-n)], zeros(n, 1), ...
        [], [], [], [], [], optimset('Algorithm', 'interior-point-convex', ...
        'TolFun', 1e-13));
    %z = qp([], H, f, [], [], [], [], zeros(n,1), [A zeros(n,length(f)-n)], []); 
    obj = obj + .5 * (y' * y);
elseif strcmpi(solver, 'sedumi')
    % Use SeDuMi 
    U = [A, sparse(n,nC), -speye(n), sparse(n,n+nB+4); ...
         M, C, sparse(n,n+2), -speye(n), sparse(n,2), B; ...
         sparse(1,2*n+nC), 1, sparse(1,n+nB+3); ...
         sparse(1,3*n+nC+2), 1, sparse(1,nB+1)];
    b = [sparse(n,1); y; 1; 1];
    c = sparse([n+nC+(1:n), 2*n+nC+2, 3*n+nC+4], ...
               1, [alpha*ones(1,n), 1, gamma], 3*n+nC+nB+4, 1);
    K = struct('f', n+nC, 'l', n, 'r', [2+n 2+nB]);
    pars.eps = 1e-6;
    pars.chol.maxuden = 1e2;
    z = sedumi(U, b, c, K, pars);
    obj = c' * z;
    %objd = b' * s;
end

l = z(end-nB+1:end);
d = z(n+1:n+nC);
t = B*l + C*d;
q = z(1:n);
p = A * q;
r = M * q;
e = y - r - t;

end

the error:

i hope you can help me settle this . thank you Cris for the response.

[Walter Roberson]

You invoked the function without passing any parameters to it. What is your expectation of where it should look to find content for the variable that the function refers to as y?

[NUR BATRISYIA HANNANI ZAIN AZMI]

so what should i do?

[Walter Roberson]

y = randn(1,150);
delta = 3;

cvxEDA(y, delta)

[NUR BATRISYIA HANNANI ZAIN AZMI]

how i want to load my samples in this code to get the result?

[Cris LaPierre]

Provide the inputs you have written into your function declaration

function [r, p, t, l, d, e, obj] = cvxEDA(y, delta, varargin)
                                          ^^ ^^^^^  % These first two inputs are required

When you try to use cvxEDA, you must pass in values for y and delta.

Read more about functions here.

[NUR BATRISYIA HANNANI ZAIN AZMI]

thankyou for your response but how to write the code to pass in values of y for my samples?

[Cris LaPierre]

Walter already showed you how to do this.

You should really read though the page I linked you to. The examples show you how to write a function, and then how to use that function.