Append new column to a cell with zero & one values

I have a cell type variable with thousand rows and 8 columns. For example:

                          c3 
  A={ 12  0  69.11  13  40  100  90602  1996
      13  0  46.21  16  48  183  30502  1996
      16  0  45.34  11  37  183  40701  1996
      18  0  8.70  17  81  76  60110  1996
      26  0  78.23  13  48  254  40301  1996
      41  0  17.83  26  41  98  20501  1996
      42  0  92.02  12  75  7  20502  1996};

Taking into account the third column (c3) I would like to append a new column to variable A that wil have the value of 1 if the value in c3 corresponds to the bottom 5th percentile of all the values in c3 and zero otherwise. Basiscally append dummy variable that gives 1 for 'scores' in the bottom 5% of the distribution (c3).

Can someone help me? Thank you.

[the cyclist]

Is the notation in c3 such that the comma represents the decimal? "69,11" is a single value, equal to 69 + 11/100?

[Maria]

I am sorry, I edited the ',' to '.'. Thanks for the note.

Answer 1

[Adam]

pVal = prctile( [A{:,3}], 20 );
A(:,end+1) = {0}

A( [A{:,3}] <= pVal, end ) = {1}

I think that should do what you want.

Answer 2

[the cyclist]

A = { 12  0  69.11  13  40  100  90602  1996
      13  0  46.21  16  48  183  30502  1996
      16  0  45.34  11  37  183  40701  1996
      18  0  8.70  17  81  76  60110  1996
      26  0  78.23  13  48  254  40301  1996
      41  0  17.83  26  41  98  20501  1996
      42  0  92.02  12  75  7  20502  1996};

c3 = [A{:,3}]';

p5 = prctile(c3,5);

bot =  c3 <= p5;

A = [A,num2cell(bot)];

Answer 3

[Guillaume]

sortedcolumn = sort([A{:, 3}]);
index = floor(0.05 * numel(sortedcolumn)); %could use round as well
threshold = sortedcolumn(index);

Gets you the max of your 5 percentile.

inbottom = num2cell(A{:, 3} < threshold); tells which rows is in bottom percentile
[A{:, 9}] = inbottom{:}; %puts it in new 9th columnn

To insert in new column.

Note, this uses expansion of cell arrays to comma-separated lists. Search for comma-separated list in the matlab doc for more information.