Fixunknowns function resulting logical rows

Hello everyone,

Say Num is matrix with these numbers,

       1     2   NaN     3
       4     5   NaN   NaN
     NaN     3     2     3

And if fixunknowns used on Num, it will result like this,

      1.0000    2.0000    2.0000    3.0000
      1.0000    1.0000         0    1.0000
      4.0000    5.0000    4.5000    4.5000
      1.0000    1.0000         0         0
      2.6667    3.0000    2.0000    3.0000
           0    1.0000    1.0000    1.0000

Is there any way to get rid those 1 and 0 (logical) rows? If so please kindly answer this question, or guide me to the documentation if it's available.

Thank you,

Randi

Answer 1

[Stephen23]

Where M is the output matrix:

new = M(1:2:end,:)

[Randi Pratama]

Thank you very much, Stephen.

[Randi Pratama]

Just a little bit curious, how is it works? I've tried a bigger matrix with inf number on it and it doesn't work.

[Stephen23]

"Just a little bit curious, how is it works?"

The fixunknowns help clearly describes the output matrix: you only want every second row. This is easy to do with indexing (as my answer shows). Using indexing is one of the most basic MATLAB operations, as shown in the introductory tutorials:

https://www.mathworks.com/help/matlab/getting-started-with-matlab.html

And in more detail here:

https://www.mathworks.com/help/matlab/math/matrix-indexing.html

The indexing in my answer simply uses the colon operator to generate a row index selecting every second row, starting from the first row, and going until the end row. The colon for the column index means "all columns" of the matrix.

"I've tried a bigger matrix with inf number on it and it doesn't work"

Indexing works for all MATLAB arrays, regardless of their size. I have no idea what code you tried, how you used Inf, or what "it doesn't work" means. If you want help please show the code you tried and explain what you expected.

[Randi Pratama]

I see, in my case the logical rows don't show up if there is no NaN or inf in one rows. Since it show up randomly, it will look like this.

-40.4622  -39.6940  -36.3595  -30.4689  -34.1643  -38.0808  -35.5036  -36.0544  -34.6927  -43.5385  -39.1102  -39.6138
    -Inf  -42.7852  -36.8174  -30.1588  -35.2925  -41.1073  -36.6398  -36.4354  -35.0243      -Inf  -35.7836  -39.5963
  1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000
    -Inf  -37.0857  -40.0706  -31.5189  -35.1094  -36.0910  -36.5473  -46.3665  -35.0425  -37.9092  -36.3211      -Inf
  1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000
    -Inf  -36.1978  -37.9558  -30.7696  -35.4735  -36.5567  -36.2683  -39.5703  -35.3162  -36.2137  -38.4406      -Inf
…

I've tried to replace the inf and NaN number using ~isfinite. Here's my code by the way.

konfirmasi2=input('Jumlah parameter? (>1) ');                                               % Inputing how many times user want the process to be executed
konfirmasi=cell(size(konfirmasi2));
ukuran=cell(1,konfirmasi2);
filetype='*.wav';
for ulang=1:konfirmasi2
fprintf('Silahkan masukkan parameter ke-%d\n',ulang)                                        % Translated to "Please input no. %d parameter(s)"
file = uigetfile(sprintf(filetype),'MultiSelect','on','Pilih suara yang akan digunakan');   % Select (one or multiple) .wav file(s) used to import to workspace
char = ischar(file);
z=cell(size(file));
if isequal(file,0);
    clear file;
   return;
end
if char==1;                                                                                 % Transforming char file resulting from importing only one file to cell file
    file={file};
    z=cell(1,1);
end
for k=1:length(file)
y=audioread(file{1,k},[1 2000]);                                                            % Low-spec computer
hasil=mfcc1(y);
hasil=hasil';
hasil=hasil(:)';
hasil=hasil';
z{k}=hasil;                                                                                 
ukuran{ulang}=z;
end
konfirmasi{ulang}=input('Simpan sebagai? ','s');                                            % Saving output as data struct
Data.(konfirmasi{ulang})=z;
fprintf('\n')
end
trainMatrix=cell2mat(cellfun(@(x) cell2mat(x),ukuran,'un',0));
trainMatrix=fixunknowns(trainMatrix);
trainMatrix=trainMatrix(1:2:end,:);                                                         % The new code, try
clc

Much appreciated