高斯分组数据

时间:2013-06-18 20:18:56

标签: matlab scaling gaussian binning

我只是有一个关于如何对数据点进行高斯分级的简单问题。让我们说在X = 100时我检测到5000个电子但我的FWHM就像4个点。在matlab中是否有可能用以X = 100为中心的高斯分组来容纳5000个电子。像X = 99和X = 101之间的2500个电子和95到105之间的5000个电子?

1 个答案:

答案 0 :(得分:1)

听起来您在单个点(X=100e=5000)进行单次测量,并且还知道FWHM(FWHM = 4)的值。

如果情况确实如此,您可以像这样计算standard deviation sigma

sigma = FWHM/ 2/sqrt(2*log(2));

你可以像这样制作垃圾箱:

[N, binCtr] = hist(sigma*randn(e,1) + X, Nbins);

其中N是每个箱子中的电子数量,binCtr是箱子中心,Nbins是您想要使用的箱子数量。

如果电子数量,则可能会耗尽内存。在这种情况下,最好做同样的事情,但是在较小的批次中,如下:

% Example data
FWHM = 4;
e = 100000;
X = 100;
Nbins = 100;

% your std. dev.
sigma = FWHM/ 2/sqrt(2*log(2));

% Find where to start with the bin edges. That is, find the point 
% where the PDF indicates that at most 1 electron is expected to fall 
f = @(x, mu, sigma) exp(-0.5*((x-mu)/sigma).^2)/sigma/sqrt(2*pi);
g = @(y) quadgk(@(x)f(x,X,sigma), -inf, y)*e - 1; 
h = fzero(g, X-FWHM*3);

% Create initial bin edges
binEdges = [-inf  linspace(h, 2*X-h, Nbins-2)  +inf];

% Bin electrons in batches
c = e;
done = false;
step = 5e3;
Nout = zeros(Nbins,1);
while ~done

    % electrons still to be binned
    c = c - step;

    % Last step
    if c <= 0
        step = c+step;
        c = 0;
        done = true;
    end

    % Bin the next batch
    N = histc(sigma*randn(step,1) + X, binEdges);   
    Nout = Nout + N;

end

% Bin edges must now be re-defined
binEdges =[...
    2*binEdges(2)-binEdges(3),... 
    binEdges(2:end-1),...
    2*binEdges(end-1)-binEdges(end-2)];