MATLAB: Finding unique phrases and their frequency in a cell array

unique

Hi,

I usually deal with numbers but never with text. I have a cell array of phrases(attached the mat file) . I want to know what are these uniques phrases and the frequency of each phrase (how many time it appears). Any help would be greatly appreciated.

Best Answer

Try this:

D = load('matlab.mat');
nbEmployees = D.nbEmployees;
[uE,~,ix] = unique(nbEmployees);
tally = accumarray(ix, 1);
Out = table(uE, tally, 'VariableNames',{'Phrase','Frequency'})

producing:

Out =
  13×2 table
                    Phrase                    Frequency
    ______________________________________    _________
    {'Between 1,000 and  2,999 employees'}        1    
    {'Between 100 and 149 employees'     }        5    
    {'Between 15 and  24 employees'      }       11    
    {'Between 150 and 199 employees'     }        1    
    {'Between 200 and 249 employees'     }        1    
    {'Between 25 and  49 employees'      }        9    
    {'Between 250 and 499 employees'     }        1    
    {'Between 5 and 14 employees'        }       25    
    {'Between 50 and  74 employees'      }        9    
    {'Between 500 and  749 employees'    }       10    
    {'Between 75 and  99 employees'      }        4    
    {'I don't know'                      }        3    
    {'Under 5 employees'                 }       70

Related Solutions

MATLAB: Replacing -999 in a table to NaN regardless of the type of the column

Let's take some sample data.

x = randi([-10 10], 1, 10);

Change three locations to -999. I also want to make a copy of x to keep around so I can show two techniques and have the original data for comparison.

x(randperm(10, 3)) = -999
x = 1×10
  -999  -999    -9    -9     4     6     8     7     0  -999
xcopy = x;

You could use logical indexing.

x(x == -999) = NaN
x = 1×10
   NaN   NaN    -9    -9     4     6     8     7     0   NaN

You could also use standardizeMissing.

y = standardizeMissing(xcopy, -999)
y = 1×10
   NaN   NaN    -9    -9     4     6     8     7     0   NaN

Now compare the original data and the two modified copies.

[xcopy; x; y]
ans = 3×10
  -999  -999    -9    -9     4     6     8     7     0  -999
   NaN   NaN    -9    -9     4     6     8     7     0   NaN
   NaN   NaN    -9    -9     4     6     8     7     0   NaN

MATLAB: How do i check the frequency of an element in a multi level cell array

The following loads and concatenates all elements of your cell array, including the 1 x 2 string in result2{1,1}(7), into a single column string array called elements:

load asd
elements = vertcat(result2{:});
elements = [elements{:}].';

The next step is to get a unique list of names and the index iNames which can be used to get frequency as follows:

[NameList,~,iNames] = unique(elements,'stable');  
Frequency = histcounts(iNames,numel(NameList)).'; 
T = table(NameList, Frequency)

The second input to the unique function is used to maintain the order of the input array, you might also like to have a look at the documentation on unique regarding the third output.

Edit: Modified unique to maintain the order and included some description of the code.

Best Answer

Related Solutions

MATLAB: Replacing -999 in a table to NaN regardless of the type of the column

MATLAB: How do i check the frequency of an element in a multi level cell array

Related Question