normalization issue pdf gmm matlab
显示 更早的评论
I plotted a histogram normalized by pdf. I den fitted GMM model using fitgmdist. I then plotted the pdf of gmm over histrogram . I found pdf way too higher than the normalized histogram values. kindly help. I am attaching the data
options = statset('Display','final');
gmdist = fitgmdist(data',3,'CovType','diagonal','Regularize', 1e-5)
gmsigma = sqrt(gmdist.Sigma);
gmmu = gmdist.mu;
gmwt = gmdist.ComponentProportion;
histogram(data, 'Normalization', 'pdf', 'EdgeColor', 'none')
x = -5:0.0001:30;
xlim([-5 30])
hold on;
plot(x', pdf(gmdist, x'), 'k')
hold on;
2 个评论
Torsten
2022-6-3
I think the histogram is much too low ...
Torsten
2022-6-3
@intesaaf ashraf comment moved here:
ok but why histogram is so low then
采纳的回答
Tom Lane
2022-6-3
You have a substantial amount of missing (NaN) data. I believe the histogram function produces a density that integrates to the proportion of missing values. That is, part of the pdf is allocated to the NaN values so is not shown. Try this instead:
histogram(data(~isnan(data)), 'Normalization', 'pdf', 'EdgeColor', 'none')
1 个评论
Torsten
2022-6-3
@intesaaf ashraf comment moved here
thanks a lot
更多回答（0 个）
类别
在 帮助中心 和 File Exchange 中查找有关 Gaussian Mixture Models 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
Translated by 
