Comparing strings in cell arrays

Question

I'm trying to find the most frequent word in a list of words. Here is my code so far:

uniWords = unique(lower(words));
for i = 1:length(words)
    for j = 1:length(uniWords)
        if (uniWords(j) == lower(words(i)))
            freq(j) = freq(j) + 1;
        end
    end
end

When I try to run the script, I get the following error:

Undefined function 'eq' for input arguments of
type 'cell'.

Error in Biweekly3 (line 106)
    if (uniWords(j) == lower(words(i)))

Any help is appreciated!

chappjc · Answer 1 · 2013-11-09T00:20:50.067

2

You need to extract the contents of the cell with {}:

strcmpi(uniWords{j},words{i})

Also, I suggest comparing strings with strcmp or in this case strcmpi, which ignores case so you do not need to call lower.

Be careful when using == on strings because they must be the same length or you will get an error:

>> s1='first';
>> s2='second';
>> s1==s2
Error using  == 
Matrix dimensions must agree.

edited Nov 09 '13 at 00:20

answered Nov 08 '13 at 23:25

chappjc

30,359
6
75
132

+1 Wise advices. But you forgot an important one: vectorize! :-P – Luis Mendo Nov 09 '13 at 00:33
1

Yeah, yeah, yeah... ;) – chappjc Nov 09 '13 at 00:34

score 2 · Answer 2 · edited May 23 '17 at 12:21

2

I guess you need to do:

if (uniWords{j} == lower(words{i}))

Also, I suggest not using i and j as variables in MATLAB.

Update

As Chappjc points out, it is better to use strcmp (or in your case strcmpi and skip lower), since you want to ignore cases.

edited May 23 '17 at 12:21

Community

1
1

answered Nov 08 '13 at 23:25

Stewie Griffin

14,889
11
39
70

There is another issue with the code in the question. For strings of different length, using `==` will give `Matrix dimensions must agree.` – chappjc Nov 09 '13 at 00:23

Luis Mendo · Answer 3 · 2013-11-09T00:39:31.410

No need for loops. unique gives you a unique identifier for each word, and you can then sum occurrences of each identifier with sparse. From that you easily find the maximum, and the maximizing word(s):

[~, ~, jj ] = unique(lower(words));
freq = full(sparse(ones(1,length(jj)),jj,1)); % number of occurrences of each word
m = max(freq);
result = lower(words(jj(freq==m))); % return more than one word if there's a tie

For example, with

words = {'The','hello','one','bye','the','one'}

the result is

>> result

result = 

    'one'    'the'

Comparing strings in cell arrays

3 Answers3

Linked