sonnetsCounts.mat file

Does anyone know how the sonnetsCounts.mat file was created on the following MATLAB page: https://uk.mathworks.com/help/textanalytics/ref/ldamodel.predict.html

Predict Top LDA Topics of Word Count Matrix

Load the example data. sonnetsCounts.mat contains a matrix of word counts and a corresponding vocabulary of preprocessed versions of Shakespeare's sonnets.

load sonnetsCounts.mat

size(counts)

ans = 1×2

154 3092

When I open the sonnetsCounts.mat file, it has the following data

val =

(1,1) 1

(106,1) 1

(131,1) 2

(154,1) 1

(1,2) 1

(143,2) 1

I presume the second column in the frequency of words. But I'm not sure if the vector in the first column represents two words?

Peter

Answer 1

[Walter Roberson]

The counts is a sparse matrix.

(143,2)        1

means that sonnet #143, unique word #2, had a count of 1.

[Peter Mayhew]

Thanks Walter for your response, I've spent more time looking into this since your reply.

Would I be correct in saying that the sonnet sparse matrix would be created using the encode command. i.e. encode(bag,documents)

Peter

[Walter Roberson]

No. It is the Counts property of the bag directly not the result of encoding an additional document against the bag.

[Peter Mayhew]

OK, so if I understand correctly. I would perform the following command

bag = bagOfWords(documents);

Then check the counts property of variable bag.

[Walter Roberson]

Counts with a capital C, but Yes.