多项式函数的奇怪图像

问题描述

我在绘制具有置信界限的多项式时遇到了一些麻烦。这是我的代码：

import matplotlib.pyplot as plt
import numpy as np
X = np.array([-5.965215369881319,-40.41538208207736,-15.584956679988448,-6.073510488327594,-11.784890508714675,-7.754674907524617,-17.482364474520395,2.4347468437246667,-16.133111795228572,-15.815302670890363,5.9730059659614305,-19.249139823392717,4.044936045002517,-7.102200416593474,5.035187269390729,-23.543269648523623,-12.593308808761405,-21.08859785268947,-24.712871361819676,-21.028901691001877,7.208054914877421,-29.6589088548177])
Y = np.array([-2.6822693448184607,-23.168555618313547,-3.6166894384329806,-3.5137320916685866,-3.770179381108618,-12.788411352407874,-15.698803377485447,1.9978332067376703,-11.838042662997829,-8.377671546754629,8.109573809406804,-14.749849913813343,2.8160696371542833,-3.3810722874645083,5.560322978176329,-16.710386872172883,-6.795050134412731,-9.855604995547115,-25.386715163603533,-11.828949808296766,-8.402106796338003,7.631911984593458,-18.155638519731614])
#
plt.plot(X,Y,'.')

poly_degree = 5
sd_cutoff = 1 # 2 keeps everything

coeffs = np.polyfit(X,poly_degree)
poly_eqn = np.poly1d(coeffs)

Y_hat = poly_eqn(X)
delta = Y - Y_hat
sd_p = np.std(delta)

ok = abs(delta) < sd_p * sd_cutoff

plt.scatter(X,color=np.where(ok,'k','r'))
plt.fill_between(
    X,Y_hat - sd_p * sd_cutoff,Y_hat + sd_p * sd_cutoff,color='#00000020')
plt.plot(X,Y_hat)

为什么我的多项式看起来这么奇怪？ https://imgur.com/a/hf4gY3P

解决方法

由于链接的问题没有提供排序解决方案，这里是对 X-Y 对进行排序的代码：

...
ind = np.argsort(X)
X = X[ind]
Y = Y[ind]
...

输出：