Error in Value of 'support' parameter must have two elements.

Error:

Error using ksdensity>compute_finite_support (line 181)

Value of 'support' parameter must have two elements.

Code:

clear

tic

disp('--- start ---')

distr='normal';

distr='kernel';

% read data

data = importdata('LDA data.mat')%import the data

features=data(:,1:end-1); %split data without labels

lable=data(:,end);

X = features;%Take up to row 11

Y = lable;%Take row 12 which is the label

% Create a cvpartition aka cross validation

c = cvpartition(Y,'holdout',.2);

% Create a training set

x = X(training(c,1),:);

y = Y(training(c,1));

% test set

u=X(test(c,1),:);

v=Y(test(c,1),:);

yu=unique(y);%find the unique elements. The resulting vector yu is sorted in ascending order and its elements are of the same class as y.

nc=length(yu); % number of classes

ni=size(x,2); % independent variables

ns=length(v); % test set

% compute class probability

for i=1:nc

    fy(i)=sum(double(y==yu(i)))/length(y);

end

switch distr

    case 'normal'

        % normal distribution

        % parameters from training set

        for i=1:nc

            xi=x((y==yu(i)),:);

            mu(i,:)=mean(xi,1);

            sigma(i,:)=std(xi,1);

        end

        % probability for test set

        for j=1:ns

            fu=normcdf(ones(nc,1)*u(j,:),mu,sigma);

            P(j,:)=fy.*prod(fu,2)';

        end

    case 'kernel'

        % kernel distribution

        % probability of test set estimated from training set

        for i=1:nc

            for k=1:ni

                xi=x(y==yu(i),k);

                ui=u(:,k);

                fuStruct(i,k).f = ksdensity(xi,ui);

            end

        end

        % re-structure

        for i=1:ns

            for j=1:nc

                for k=1:ni

                    fu(j,k)=fuStruct(j,k).f(i);

                end

            end

            P(i,:)=fy.*prod(fu,2)';

        end

    otherwise

        disp('invalid distribution stated')

        return

end

 %get predicted output for test set

[pv0,id]=max(P,[],2);

for i=1:length(id)

    pv(i,1)=yu(id(i));

end

% compare predicted output with actual output from test data

confMat=myconfusionmat(v,pv);

disp('confusion matrix:')

disp(confMat)

conf=sum(pv==v)/length(pv);

disp(['accuracy = ',num2str(conf*100),'%'])

toc

LDA data is a data with seven features to compare and its has a last column as labels.

[Walter Roberson]

duplicates http://www.mathworks.co.uk/matlabcentral/answers/59845-value-of-support-parameter-must-have-two-elements

[Lester Lim]

Oh okay, my bad...