Matlab中单元格数组的比较

Question

I have two cell arrays storing the unigram and bigrams each, which i have extracted from the text file. Now i have to compare each unigram with the bigram to find the count and later probablity of unigram present in the bigram. Can any one please help me how to sort that problem,i have already used strcmp but its not working. I am writing my code below:

for i = 1
    for j = 1:bigramRow
       bigram1 = regexp(splitBigramCellsA{j},'<s>|\w*|</s>','match');
       b1 = cellfun(@(x,y)[x], bigram1(1:end-1)','un',0)
       match = strcmp(splitUnigramCellsA, splitBigramCellsA{j,1});

        if match ==1
           bigram1count = splitbigramCellsB{j};
            unigram1count = splitUnigramCellsB{j};
            disp(bigram1count)
            disp(unigram1count)
        end
 end
end

Answer 1

If you can fit the text in memory, you could do the following:

1. create a cell array of all words (in order)
2. call unique on the cell array, and capture the third output as well. That is the original text represented as an array of indices, where each index refers to a unigram.
3. create all bigrams as bigrams = [indices(1:2:largestEven),indices(2:2:largestEven);indices(2:2:largestOdd),indices(3:2:largestOdd)] , where largestEven is 2*floor(length(indices)/2) , and largestOdd is 2*floor((length(indices)+1)/2)+1 .
4. calculate eg the frequency of each unigram in the bigrams as tabulate(bigrams(:))