Hi ,

I have purchased the book

‘Classification,Parameter Estimation and State Estimation--an engineering approach using MATLAB’

List 7.6:

List 7.7

The scripts in the book are checked with a previous version of PRTools, which can be obtained from the website. Newer versions of PRTools are sometimes not fully compatible. Here we have an example:

94 by 2 dataset with 0 classes: []

??? Error using ==> dataset.setlabtype
Soft labeled datasets should contain at least one class

The dataset z is unlabeled. In previous versions unlabeled objects were either
neglected, or, in case the total dataset was unlabeled, given the label ‘1’.
In the newer versions this was removed in order the give the user the responsibility
to handle unlabeled objects in his own way. As the dataset z has no labels, the label
type cannot be changed and an error is generated.

This can easily be fixed, in case you want to run the example with a newer version
of PRTools, by labeling the dataset:

Bob Duin

Hi, professor Duin!
Thank you very much for your detailed reply!

After consindering your fixing method to this problem carefully, I got another fixing method in Prtool4.1.4 as follows. The codes of this method is just more compact.

z = setlabtype(z,’soft’); ——> z = setlabtype(z,’soft’,[]);

There is no difference in the result.

Kun Feng

BobDuin - 18 December 2008 10:51 PM

The scripts in the book are checked with a previous version of PRTools, which can be obtained from the website. Newer versions of PRTools are sometimes not fully compatible. Here we have an example:

load nutsbolts_unlabeled; % Load the data set z
z

94 by 2 dataset with 0 classes: []

z = setlabtype(z,'soft'); % Set probabilistic labels

??? Error using ==> dataset.setlabtype
Soft labeled datasets should contain at least one class

The dataset z is unlabeled. In previous versions unlabeled objects were either
neglected, or, in case the total dataset was unlabeled, given the label ‘1’.
In the newer versions this was removed in order the give the user the responsibility
to handle unlabeled objects in his own way. As the dataset z has no labels, the label
type cannot be changed and an error is generated.

This can easily be fixed, in case you want to run the example with a newer version
of PRTools, by labeling the dataset:

load nutsbolts_unlabeled; % Load the data set z
z = setlabels(z,1);
z = setlabtype(z,'soft'); % Set probabilistic labels
[lab,w] = emclust(z,qdc,4); % Cluster using EM
figure; clf; scatterd(z);
plotm(w,[],0.2:0.2:1); % Plot results

Bob Duin

BobDuin - 18 December 2008 10:51 PM

The scripts in the book are checked with a previous version of PRTools, which can be obtained from the website. Newer versions of PRTools are sometimes not fully compatible. Here we have an example:

load nutsbolts_unlabeled; % Load the data set z
z

94 by 2 dataset with 0 classes: []

z = setlabtype(z,'soft'); % Set probabilistic labels

??? Error using ==> dataset.setlabtype
Soft labeled datasets should contain at least one class

The dataset z is unlabeled. In previous versions unlabeled objects were either
neglected, or, in case the total dataset was unlabeled, given the label ‘1’.
In the newer versions this was removed in order the give the user the responsibility
to handle unlabeled objects in his own way. As the dataset z has no labels, the label
type cannot be changed and an error is generated.

This can easily be fixed, in case you want to run the example with a newer version
of PRTools, by labeling the dataset:

load nutsbolts_unlabeled; % Load the data set z
z = setlabels(z,1);
z = setlabtype(z,'soft'); % Set probabilistic labels
[lab,w] = emclust(z,qdc,4); % Cluster using EM
figure; clf; scatterd(z);
plotm(w,[],0.2:0.2:1); % Plot results

Bob Duin

hello dear
i have trainset.mat ( 1599x6 118346 dataset object),and fullim.mat( 16384x18 2359296 double array) iwant write code for classification image involve 5 class, when write this code:
load trainset;
z=dataset(trainset);
z=setlabtype(z,’soft’[.9 .9 .9 .9 .9 ];
w=qdc(z);
load fullim.mat
nrow=412;
ncol=412;
T=dataset(fullim);
x=T*w*classc %%ERROR
??? Error using ==> dataset.setfeatsize

The total feature size does not fit with the real number of features.
This might be caused when mappings are applied to datafiles with objects of different size.

what can i do?