11.1
代码如下:
import numpy as np
import matplotlib.pyplot as plt
X = np.arange(0, 2.0, 0.02)
Y = (np.sin(X - 2)**2 ) * np.exp(-X ** 2)
plt.plot(X, Y)
plt.title("11.1")
plt.xlabel("X")
plt.ylabel("Y")
plt.axis([0, 2, 0, 1])
plt.show()
效果图如下:
11.2
注释:这里我们认为使用最小二乘得到的多元线性回归模型中不含常数项,否则X应该为20*11,首项均为1的矩阵才对
代码如下:
import matplotlib.pyplot as plt
import numpy as np
X = np.random.randint(-5,5,200).reshape(20,10)
b = np.random.randint(-2,2,10).reshape(10,1)
z = np.random.normal(size = (20,1))
y = np.dot(X,b)+z
estib = np.linalg.lstsq(X, y,None)[0]
t = np.arange(0,10,1)
plt.scatter(t,b,c = 'r',marker = 'x')
plt.scatter(t,estib,c = 'b')
plt.legend(['True coefficients','Estimated coefficients'])
plt.xlim(0,9)
plt.ylim(-2.2,2.2)
plt.hlines(0,0,9,linestyles = 'dashed')
plt.show()
结果如下:
11.3
import matplotlib.pyplot as plt
import numpy as np
from scipy import stats
z = np.random.randn(10000)
gkde = stats.gaussian_kde(z)
ind = np.linspace(4,-4,100)
kdepf = gkde.evaluate(ind)
plt.hist(z,bins=25,normed = 1,color = 'b',edgecolor = 'k')
plt.plot(ind,kdepf,label='kde',color='r')
plt.xlabel('value')
plt.ylabel('frequency')
plt.title('11.3')
plt.show()
效果图: