how to create a function that counts all words in a document?

I am attempting to make a function that takes in a speech as an input argument in a text file. I would then like to read the speech and have matlab count and return the number of occurrences for every word in the speech. Any ideas?

Answer 1

[John D'Errico]

Easy, peasy.

I'd...

1. Using lower, make all words lower case, so you need not worry about capitalization.

2. Apply my allwords code to the string. Use spaces and any appropriate punctuation as separator characters.

3. Use unique to get the distinct words in that list. The third argument from unique will be important.

4. Apply accumarray to that third argument from unique to count the words.

Essentially 4 quick lines of code.

[Noah Swiston]

The allwords function only takes in a string as an input. How would i take the text file, open it, and convert it into a string so the allwords function could be used?

[John D'Errico]

textread is the simple answer. Or use fopen, then fread, or fgets or fgetl.

No need to convert anything to a string if it is a string already as read in.

Answer 2

[Sarah Palfreyman]

Get the size of the string returned by extractFileText in Text Analytics Toolbox

Answer 3

[Christopher Creutzig]

Check out the bagOfWords class in Text Analytics Toolbox. It solves exactly this problem.

(You may want to have some preprocessing to match exactly what you consider to be the same word. E.g., lowercasing, so “House” and “house” are the same. Unless you want Dr. House not to be counted as a house, of course …)