iterate over a vector in a for loop

I am trying to loop through a matrix - my model has no problem looping through the bandwidths "bw" but i get an error when looping through the priors, because prior has vectors of values

the input of prior works in the format when not using a for loop...

prior = [0.427,0.226,0.347];

but i want the model to run through 3 different types of prior probabilities, as below

the error i get is: Error using classreg.learning.classif.FullClassificationModel.processPrior (line 264)

Prior probabilities must be a vector of length 3.

when i print prior i get

priors =

0.3300 0.3300 0.3300

0.4270 0.2260 0.3470

0.2000 0.2000 0.6000

bandwidths = [0.2, 0.5, 1];

priors = [0.33 0.33 0.33; 0.427 0.226 0.347; 0.2 0.2 0.6];

global mymatrix

for f=1:length(bandwidths);

for l=1:length(priors);

bw= bandwidths(f);

p = priors(l)

Mdl = fitcnb(x_train_crossval,y_labels_train_crossval,'ClassNames', class_names,'PredictorNames',predictor_names,'Prior',p,'Width',bw);

Answer 1

[Jan]

Avoid length, but use size with a specified dimension.

for f = 1:size(bandwidths, 2)
  for l = 1:size(priors, 1)
    bw = bandwidths(f);
    p  = priors(l, :);
    

priors(l) is a scalar, but the error message explains, that a [1 x 3] vector is required.

Answer 2

[Umar]

Hi Catherine,

To resolve the error and correctly loop through different sets of prior probabilities, you need to adjust how the priors variable is accessed within the loop. Instead of iterating over rows of the priors matrix, you should extract a single row vector for each iteration.

Here is the corrected code snippet:

global mymatrix

for f = 1:length(bandwidths) for l = 1:size(priors, 1) % Iterate over rows of priors bw = bandwidths(f); p = priors(l, :); % Extract a single row of priors Mdl = fitcnb(x_train_crossval, y_labels_train_crossval, 'ClassNames', class_names, 'PredictorNames', predictor_names, 'Prior', p, 'Width', bw); end end

By using p = priors(l, :), you can access each row of the priors matrix as a single row vector, satisfying the requirements of the fitcnb function.

Hope this will help resolve your issue.

[Umar]

Sorry for unformatted code