为每个 S.D 绘制具有百分比或数据计数的正态分布

Question

l = {31.2: 1, 35.1: 4, 39.0: 13, 42.9: 33, 46.8: 115, 50.7: 271, 54.6: 363, 58.5: 381, 62.4: 379, 66.3: 370, 70.2: 256, 74.1: 47, 78.0: 2}
# convert the dictionary to a list
l_list = [k for k, v in l.items() for _ in range(v)]

我想要符合上述数据的正态分布曲线以及打印在每个区域的样本数量，如下图所示。

Answer 1

这是一种计算和绘制适合数据的高斯法线的方法。请注意，数据已经预先分组在一起，因此无法再计算真实的均值和标准差。

import numpy as np
import matplotlib.pyplot as plt
from scipy.stats import gaussian_kde, norm

l = {31.2: 1, 35.1: 4, 39.0: 13, 42.9: 33, 46.8: 115, 50.7: 271, 54.6: 363, 58.5: 381, 62.4: 379, 66.3: 370, 70.2: 256, 74.1: 47, 78.0: 2}
# convert the dictionary to a list
l_array = np.array( [k for k, v in l.items() for _ in range(v)])


s = sum(l.values())
bin_width = 3.9
bin_centers = list(l.keys())
heights = [v/s/bin_width for v in l.values()]
plt.bar(bin_centers, heights, width=bin_width*0.9, color='dodgerblue', label='histogram')
for c, h, v in zip(bin_centers, heights, l.values()):
    plt.text(c, h, v, ha='center', va='bottom')
plt.xticks(bin_centers)
mean = l_array.mean()
sdev = l_array.std()
for i in range(-3, 4):
    plt.axvline(mean+i*sdev, color='limegreen', ls='--', lw=2, label='mean + i*sdev' if i == 0 else None)

x = np.linspace(mean-4*sdev, mean+4*sdev, 500)
plt.plot(x, norm.pdf(x, mean, sdev), color='orange', lw=3, label='gaussian normal')
plt.autoscale(enable=True, axis='x', tight=True)
plt.legend()
plt.show()

为每个 S.D 绘制具有百分比或数据计数的正态分布

Plotting normal distribution with percentage or counts of data for each S.D

python-3.x

pandas

matplotlib

seaborn

normal-distribution