Log-Normal fit of histogram

12 次查看（过去 30 天）
De Ar
De Ar 2019-10-29
回答： Jeff Miller 2019-10-29
I have a skewed Gaussian distribution in my histogram that I am trying to fit using a log-normal distribution, but my code is giving me a flat fit. Any suggestions to what I am doing wrong?
skewedGauss.png
Attempt
yyaxis left
bar(binLocations, counts);
xlabel('Intensity')
ylabel('Pixel Counts')
% Log-Normal fit
pd = fitdist(counts, 'LogNormal');
pd = makedist('LogNormal', 'mu', pd.mu, 'sigma', pd.sigma);
counts_fit = pdf(pd, binLocations);
counts_fit = counts_fit./sum(counts_fit);
yyaxis right
plot(binLocations, counts_fit, '-', 'LineWidth', 2)
xlim([min(binLocations(:)) max(binLocations(:))])

请先登录，再回答此问题。

回答（1 个）

Jeff Miller
Jeff Miller 2019-10-29
fitdist wants the x values that were tabulated to get the histogram, not the bin counts. From your histogram, it looks like your x should be a vector of length about 10^6 containing numbers in the range 0.4--0.8.

类别

在 Help Center 和 File Exchange 中查找有关 Lognormal Distribution 的更多信息

标签

产品


版本

R2018a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by