How to produce new samples

[Kafkaica]

Hi

I am trying to find the best Bivariate fit for my data and produce new samples. When I choose the Clayton model, I receive decent data. However, when I choose Frank or Gumbel, the produced data turns out like the below figure. I was wondering if someone could help me with that.

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from copulas.bivariate.base import Bivariate, CopulaTypes
from copulas.bivariate.clayton import Clayton
from copulas.bivariate.frank import Frank
from copulas.bivariate.gumbel import Gumbel
import scipy.stats as stats


""" Preparing data """
with open('Maximum Yearly Discharge.txt', 'r') as f:
    file = f.read()
lst = file.split('\n')
x = np.array([])
for i in lst:
    x = np.append(x, float(i))

with open('Prediction.txt', 'r') as f:
    file = f.read()
lst = file.split('\n')
y = np.array([])
for i in lst:
    y = np.append(y, float(i))

z = np.append(x, y)
z = np.reshape(z, (int(len(z)/2), 2), order='F')


copula = Bivariate(CopulaTypes.FRANK)
copula.fit(z)


""" Producing Samples"""
samples = copula.sample(1000)

normalized_x = (x-min(x))/(max(x)-min(x))
normalized_y = (y-min(y))/(max(y)-min(y))
plt.scatter(samples[:, 0], samples[:, 1], color='0.75', label='Simulated Data')
plt.scatter(normalized_x, normalized_y, label='Empirical Data', color='blue')
plt.xlabel('Maximum Yearly Discharge (Scaled)')
plt.ylabel('Associated Tidal height (Scaled)')
plt.legend(loc='lower right')
plt.savefig('Simulated Data.jpg')
plt.show()

Maximum Yearly Discharge.txt
Prediction.txt

[Kafkaica]

Using Clayton for sampling

using Gumbel or Frank for sampling:

[csala]

Hi @Kafkaica thanks for bringing this up, and for the detailed example!
We will have a look at it as soon as we can and provide a response.

[JDTheRipperPC]

Hi @Kafkaica Upon reviewing it, we detected a bug in the way the samples were produced inside the Frank and Gumbel classes.
We just fixed it and a new release with this bug fix will be created soon.

Thanks for the heads up!