Algorithm：实现LDA的Gibbs Gauss采样(绘制多图subplot)

输出结果

实现代码

import numpy as np

import matplotlib.pyplot as plt

N = 1000

# 初始化y, 可以任选一个值

y = 0

xs = []

ys = []

for i in range(N):

   # 更新x_t

   x = np.random.normal(0.8*y, 0.6)

   # 更新y_t

   y = np.random.normal(0.8*x, 0.6)

   xs.append(x)

   ys.append(y)

xs2, ys2 = np.random.multivariate_normal( [0, 0], [[1,0.8],[0.8,1]], N ).T

plt.subplot(211)

plt.title('gibbs Gauss')

plt.scatter(xs, ys)

plt.subplot(212)

plt.scatter(xs2, ys2)

plt.show()