我想要一个非常快速的实现来计算我的时间序列上的移动窗口偏斜度，因为我必须为许多不同的窗口大小做这件事，并且应用程序是时间关键的（类似地，对于其他度量，如峰度，均方根，自相关等，也要这样做）。在这里，我专注于偏度。
我有一个时间序列数据数组state_timeseries（~ 10，000，000）和相应的时间值。我想使用窗口大小（~ 300，000 - 4，000，000）和步长（=相应窗口大小的1%= 3000 - 40000）来查找此时间序列的移动窗口偏斜度。对于以下列表中的所有窗口尺寸，均应执行此操作：

window_size_list = floor(linspace(300000, 4000000, 200));
step_size_list = floor(window_size_list * (1 / 100));

对于并行计算，我有16个工作人员，拥有约130 GB RAM和GeForce RTX 2070 Super 8 GB大小。
如果我用C来做这些计算，有可能获得很好的速度增益吗？（我不知道C，但阅读在线让我想到这一点）。然后将生成的移动窗口偏度时间序列返回MATLAB进行进一步处理。

我的当前实现：

目前，我只是简单地循环（包括窗口大小和每个窗口的移动窗口），如下所示（但这相当慢）：

% Dummy data of equal length
state_timeseries = randn(1, 10000000);
time = 1: length(state_timeseries);

total_window_count = 200; % 200 as used above

window_size_list = floor(linspace(300000, 4000000, total_window_count));
step_size_list = floor(window_size_list * (1 / 100));

for k = 1: total_window_count

    % Set window size and step size to calculate the corresponding EWS timeseries
    window_size = window_size_list(k);
    window_step = step_size_list(k);

    % Generate EWS timeseries for particular window size
    Data_skewness_timeseries{k} = sk_timeseries_func(time, state_timeseries, window_size, window_step);
    
end

function Data = sk_timeseries_func(time, state_timeseries, window_size, window_step)

    % Use GPU if available
    gpu_available = canUseGPU();
    if gpu_available == 1
        time = gpuArray(time);
        state_timeseries = gpuArray(state_timeseries);
    end
    
    % Get the indices where windows start and the corresponding times at which they end.
    % This is to specify the skewness of a window at the time value at the end of the window.
    window_idx = 1: window_step: (length(time) - window_size + 1);
    time_window_ends_idx = window_idx + window_size - 1;
    time_window_ends = time(time_window_ends_idx);
    
    sk_timeseries = zeros(1, length(time_window_ends));
    
    for i_window = 1: length(window_idx)
        timeseries_window_data = state_timeseries(window_idx(i_window): window_idx(i_window) + window_size - 1);
        
        sk_timeseries(i_window) = skewness(timeseries_window_data);
    
    end
    
    Data.time_window_ends = time_window_ends;
    Data.sk_timeseries = sk_timeseries;
    
end