How to plot probability density curve?

3 次查看（过去 30 天）

显示更早的评论

Andi 2021-12-8

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1605810-how-to-plot-probability-density-curve

评论： Andi 2022-3-30

采纳的回答： Star Strider

R_0.01.csv

在 MATLAB Online 中打开

I have modify the strip as per my data but its the resulst are not expected. Why the movemedian=25 is fixed here.

X = readmatrix('R_0.01.csv');
r_a=[X(1,:)];
r_b=[X(2,:)];
r_c=[X(3,:)];
r_d=[X(4:8,:)];
r_e=[X(9:12,:)];
r_f=[X(13:16,:)];
r_g=[X(17:26,:)];
r_h=[X(27:48,:)];
r_i=[X(49:120,:)];
r_j=[X(121:186,:)];
r_aam = (r_a(~isnan(r_a)));
r_abm = r_b(~isnan(r_b));
r_acm = r_c(~isnan(r_c));
r_adm = r_d(~isnan(r_d));
r_aem = r_e(~isnan(r_e));
r_afm = r_f(~isnan(r_f));
r_agm = r_g(~isnan(r_g));
r_ahm = r_h(~isnan(r_h));
r_aim = r_i(~isnan(r_i));
 r_ajm= r_j(~isnan(r_j));
pd = makedist('Normal')
[f1,x1,flo1,fup1] = ecdf(r_aam);
[f2,x2,flo2,fup2] = ecdf(r_abm);
[f3,x3,flo3,fup3] = ecdf(r_acm);
[f4,x4,flo4,fup4] = ecdf(r_adm);
[f5,x5,flo5,fup5] = ecdf(r_aem);
[f6,x6,flo6,fup6] = ecdf(r_afm);
[f7,x7,flo7,fup7] = ecdf(r_agm);
[f8,x8,flo8,fup8] = ecdf(r_ahm);
[f9,x9,flo9,fup9] = ecdf(r_aim);
[f10,x10,flo10,fup10] = ecdf(r_ajm);
figure
plot(x, f)
grid
title('Empirical CDF')
dfdxs1 = smoothdata(gradient(f1)./gradient(x1), 'movmedian',25);
dfdxs2 = smoothdata(gradient(f2)./gradient(x2), 'movmedian',20);
dfdxs3 = smoothdata(gradient(f3)./gradient(x3), 'movmedian',25);
dfdxs4 = smoothdata(gradient(f4)./gradient(x4), 'movmedian',25);
dfdxs5 = smoothdata(gradient(f5)./gradient(x5), 'movmedian',25);
dfdxs6 = smoothdata(gradient(f6)./gradient(x6), 'movmedian',25);
dfdxs7 = smoothdata(gradient(f7)./gradient(x7), 'movmedian',25);
dfdxs8 = smoothdata(gradient(f8)./gradient(x8), 'movmedian',25);
dfdxs9 = smoothdata(gradient(f9)./gradient(x9), 'movmedian',25);
dfdxs10 = smoothdata(gradient(f10)./gradient(x10), 'movmedian',1000);
aaa1=smooth(dfdxs1)
aaa2=smooth(dfdxs2)
aaa3=smooth(dfdxs3)
aaa4=smooth(dfdxs4)
aaa5=smooth(dfdxs5)
aaa6=smooth(dfdxs6)
aaa7=smooth(dfdxs7)
aaa8=smooth(dfdxs8)
aaa9=smooth(dfdxs9)
aaa10=smooth(dfdxs10)
figure
plot(x1, aaa1)
plot(x2, aaa2)
plot(x3, aaa3)
plot(x4, aaa4)
plot(x5, aaa5)
plot(x6, aaa6)
plot(x7, aaa7)
plot(x8, aaa8)
plot(x9, aaa9)
plot(x10, aaa10)

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

采纳的回答

Star Strider 2021-12-8

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1605810-how-to-plot-probability-density-curve#answer_850145

For data with an unknown distribution, I generally use the empirical cumulative distribution (ecdf) function to get the CDF, and the use the gradient function to derive the PDF. This is generally more robust than estimating the PDF directly, at least in my experience.