Calculation of y axis values or cdf values from given x axis values using another plot of best fit cdf

I have calculated the cdf of best fit distribution. Now I have some x axis values. I want to get their corresponding y axis values or rather cdf values using the same cdf plot of best fit distribution as mentioned above. how to do so?

[Torsten]

In which form do you have the cdf ? As a function depending on x like f(x) = x^2 ? As a list of (x/y) tuples ? ...

[Payel]

best fit distribution has been found to be weibull. then cdf in a matrix format has been calculated using y=cdf(pd,x).

[Torsten]

Then - given an x-value of your choice - you should be able to compute the corresponding y-value just as y = cdf(pd,x), shouldn't you ?

Answer 1

[Star Strider]

It would help to have your code and data.

Assuming it is similar to this approach, this may do what you want —

[f,x] = ecdf(wblrnd(2,6,[1 500]));                      % Calculate Empirical CDF
[xs,idx] = unique(x);                                   % Unique Points (Required For 'interp1')
fs = f(idx);                                            % Unique Points (Required For 'interp1')

xq = 3*rand(1,10)                                       % Random Query Points (Can Be Any Values Within The Range of 'xs')

xq = 1×10

2.2552 1.2215 1.1666 2.8139 0.1724 0.1016 1.4187 2.6146 2.1569 1.2920

yq = interp1(xs, fs, xq, 'linear')                      % Interpolated CDF Values

yq = 1×10

0.8649 0.0489 0.0356 NaN NaN NaN 0.1095 0.9974 0.7893 0.0657

figure
plot(x, f, 'DisplayName','CDF')
hold on
stem(xq, yq, '.r', 'DisplayName','Values At Query Points')
hold off
grid
legend('Location','best')

prms = wblfit(wblrnd(2,6,[1 500]));                     % Estimate Weibull Parameters
x = linspace(0, 3, 150);                                % Independent Variable Vector (For 'plot')
f = wblcdf(x,prms(1),prms(2));
q = 2.5*rand(1,10)                                      % Random Query Points (Can Be Any Values Within The Range of 'xs')

q = 1×10

1.8311 0.1284 1.7376 1.1860 0.9805 2.0554 1.7172 0.3337 0.9212 0.7660

yq = interp1(x, f, xq, 'linear')                        % Interpolated CDF Values

yq = 1×10

0.8782 0.0541 0.0414 0.9996 0.0000 0.0000 0.1263 0.9936 0.8013 0.0746

figure
plot(x, f, 'DisplayName','CDF')
hold on
stem(xq, yq, '.r', 'DisplayName','Values At Query Points')
hold off
grid
legend('Location','best')

yq = wblcdf(xq,prms(1),prms(2));                        % Use 'wblcdf' To Calculate & Plot Query Points

figure
plot(x, f, 'DisplayName','CDF')
hold on
stem(xq, yq, '.r', 'DisplayName','Values At Query Points')
hold off
grid
legend('Location','best')

Experiment with this approach with your data.

EDIT — (13 Aug 2023 at 16:36)

Expanded to include specific distribution, additional options.

.

[Payel]

Thanks

[Star Strider]

My pleasure!

If my Answer helped you solve your problem, please Accept it!

.