Info

# How to perform repeated calculations on a different bin

1 view (last 30 days)
Charlie Finnie on 25 Oct 2016
Closed: MATLAB Answer Bot on 20 Aug 2021
I have split my data into bins (5 million year spacing through time) and then performed calculations to remove NaNs, calculate the mean and calculate the standard deviation of the first bin (0-5million years).
I was just wandering if there is an easy way of repeating all of these calculations for the remaining bins (5-10myr, 10-15myr....etc)?
Thanks Charlie
##### 2 CommentsShowHide 1 older comment
Chaya N on 25 Oct 2016
How are you storing each of these bin data? (Arrays/ cells/ structures)

Steven Lord on 25 Oct 2016
Use discretize to bin your data then use the output from discretize as the subs input to accumarray.
% Sample data
n = 10;
x = n*rand(100, 1);
y = (1:100).';
% Bin it and accumulate the data in each bin
bin = discretize(x, 0:n);
A1 = accumarray(bin, y, [n 1], @mean);
% Accumulate the data in the bins using a FOR loop
A2 = zeros(n, 1);
for k = 1:n
A2(k) = mean(y(bin == k));
end
% These two results should be the same
[A1, A2, A1-A2]

Chaya N on 25 Oct 2016
Edited: Chaya N on 25 Oct 2016
Put all the bin data into a cell array and use cellfun
all_bins_raw = {bin1,bin2,....,binN}; % put all raw data in one cell array
all_bins = cellfun(@(x) x(~isnan(x)),all_bins_raw,'UniformOutput',0); % Remove NaN's
bin_means = cellfun(@mean, all_bins);
bin_sds = cellfun(@std, all_bins);