CODEWITHSUNDEEP
pythonmatplotlibplot
Ada Xu
Ada Xu

Reputation: 983

Plotting a probability density

I want plot of a curve say x^5 + x^4 + x^3 + x + 1 each point x has been taken from a normal distribution. I have a vector of means and a vector of sigma values.

Using matplotlib.pyplot I can plot the mean values and I can plot variance around the mean but it doesn't look elegant and clutters the outputHere dashed yellow line is the variance and red line is the mean.

Is there any other way to plot the density function?

I used something like this :

mu = [mu1, mu2, mu3..]
sigma = [sigma1, sigma2, sigma3..]
variance1 = [mu1+sigma1, mu2+sigma2, ..]
variance2 = [mu1-sigma1, mu2-sigma2,..]


import matplotlib.pyplot as plt
plt.plot(x,mu)
plt.plot(x,variance1, ls = "--")
plt.plot(x,variance2,ls="--")

where x is an array of inputs.

Upvotes: 3

Views: 2360

Answers (1)

Joe Kington
Joe Kington

Reputation: 284870

The most common way is to use fill_between to shade the area between the confidence intervals. For example: 

import numpy as np
np.random.seed(1977)
import matplotlib.pyplot as plt

# Generate data...
x_obs = np.linspace(-2, 2, 20)
true_model = [0.2, -0.1, 4, 2, 1, 0]

noise = np.random.normal(0, 5, x_obs.shape)
y_obs = np.polyval(true_model, x_obs) + noise

# Fit to a 5-th order polynomial
fit_model = np.polyfit(x_obs, y_obs, 5)

x = np.linspace(-3, 3, 100)
y_true = np.polyval(true_model, x)
y_pred = np.polyval(fit_model, x)

# Made up confidence intervals (I'm too lazy to do the math...)
high_bound = y_pred + 3 * (0.5 * x**4 + 3)
low_bound = y_pred - 3 * (0.5 * x**4 + 3)

# Plot the results...
fig, ax = plt.subplots()
ax.fill_between(x, high_bound, low_bound, color='gray', alpha=0.5)
ax.plot(x_obs, y_obs, 'ko', label='Observed Values')
ax.plot(x, y_pred, 'k--', label='Predicted Model')
ax.plot(x, y_true, 'r-', label='True Model')
ax.legend(loc='upper left')
plt.show()

enter image description here

Upvotes: 4

Related Questions