Cellfun for mean squared error

Question

Phoebe Daphne 2021-5-23

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/837548-cellfun-for-mean-squared-error

评论： Phoebe Daphne 2021-5-23

I have a 450x2 cell data structure called data, with each cell containing 10000 x 1 Doubles.

I want to calculate the mean squared error (via the immse function) between for each row in the cell, i.e. immse(data{i,1},data{i,2}) with i indicating one of the 450 rows. Then I want to add the calculated mean squared error to a variable called data_sum (My overall goal is to calculate the average mean squared error).

So far, I am solving this by looping over data:

data_sum = 0;
for p = 1:size(data,1)
    data_sum = data_sum + immse(data{p,1},data{p,2});
end 
data_avg = data_sum / size(data,1);

I want to optimize my code and use cellfun instead. So far I have tried the following:

C = cellfun(@immse, data{:,1}, data{:,2}, 'UniformOutput', false);

However, this gives me the error:

Error using cellfun
Input #2 expected to be a cell array, was double instead.

I read online that I should convert double to cell to solve this problem. Since I am using cellfun because, well, I am applying it to a cell, it makes me assume that I am indexing data in a wrong way, because otherwise I wouldn't have to convert it to a cell first. Or should I be using a different function than cellfun?