Numpy课后习题

Numpy课后习题

Generate matrices A, with random Gaussian entries, B, a Toeplitz matrix，where A ∈Rn×m and B ∈Rm×m, for n = 200, m = 500.
由于n = 200， m = 500过大，难以显示，使用n = 10， m = 15替代

1.类的创建和初始化，创建矩阵A、B，A矩阵利用randn生成高斯分布的元素值，B矩阵利用toeplitz函数生成toeplitz矩阵

import numpy as np
import numpy.matlib as npm
import numpy.linalg as npl
from scipy.linalg import toeplitz
import scipy
import time
import random
import math

class Exercise():

    def __init__(self):
        self.n = 10
        self.m = 15
        t = npm.rand(1, self.m)
        self.A = npm.randn(self.n, self.m)
        self.B = toeplitz(t)
        print("Matrix A:")
        print(self.A)
        print("Matrix B:")
        print(self.B)
        print()

2、Exercise 9.1:Calculate A + A, AAT,ATA and AB. Write a function that computes A(B−λI) for any λ.

做法：直接调用其内置的函数，实现加法和内乘。使用属性T获取转置矩阵。

    def Exercise_9_1(self, k):
        """
        Exercise 9.1:Calculate A + A, AAT,ATA and AB. 
        Write a function that computes A(B−λI) for any λ.
        """
        print("Exercise_9_1")
        print("A + A")
        print(self.A + self.A)
        print("A*A^T")
        print(self.A * self.A.T)
        print("A^T*A")
        print(self.A.T * self.A)
        print("AB")
        print(self.A * self.B)
        print("A(B-kI) for k = " + str(k))
        print(self.A * (self.B - k * npm.identity(self.m)))
        print()

3、Exercise 9.2:Generate a vector b with m entries and solve Bx = b.

做法：使用rand随机生成b矩阵，利用numpy.linalg.solve求解Bx=b得到x,最后带入Bx与b比较是否相等验证结果。

    def Exercise_9_2(self):
        """
        Exercise 9.2:Generate a vector b with 
        m entries and solve Bx = b.
        """
        print("Exercise_9_2")
        b = npm.rand(self.m, 1)
        print("matrix b")
        print(b)
        x = npl.solve(self.B, b)
        print("x for B*x = b")
        print(x)
        print("nunpy.allclose(self.B * x, b) = " + 
            str(np.allclose(self.B * x, b)))
        print()

4、Exercise 9.3:Compute the Frobenius norm of A: ||A||F and the infinity norm of B: ||B||∞. Also find the largest and smallest singular values of B.

做法：使用norm和eigvals函数求解范数和特征值。

    def Exercise_9_3(self, F):
        """
        Exercise 9.3:Compute the Frobenius norm of A: ||A||F and the 
        infinity norm of B: ||B||∞. Also find the largest 
        and smallest singular values of B.
        """
        print("Exercise_9_3")
        print("||A||F : " + str(npl.norm(self.A, F)))
        print("||B||∞ : " + str(npl.norm(self.B, np.inf)))
        beig = npl.eigvals(self.B)
        print("Max eigval of B : " + str(max(beig)))
        print("Min eigval of B : " + str(min(beig)))
        print()

5、Exercise 9.4:Generate a matrix Z, n × n, with Gaussian entries, and use the power iteration to find the largest eigenvalue and corresponding eigenvector of Z. How many iterations are needed till convergence?

做法：根据幂法求解的过程写成python代码，输出结果

    def Exercise_9_4(self, n):
        """
        Exercise 9.4:Generate a matrix Z, n × n, with Gaussian entries, 
        and use the power iteration to find the largest eigenvalue and 
        corresponding eigenvector of Z. How many iterations are needed 
        till convergence?
        Optional: use the time.clock() method to compare computation 
        time when varying n.
        """
        print("Exercise_9_4")
        tstart = time.clock()
        Z = npm.randn(n, n)
        u = npm.rand(n, 1)
        k1 = 0
        k2 = 1
        count = 0
        print("Matrix Z:")
        print(Z)
        while math.fabs(k1 - k2) > 1e-10:
            count = count + 1
            v = Z * u
            k1 = k2
            k2 = max(np.abs(v))
            u = v / k2
        print("count : " + str(count))
        print("the largest eigenvalue : " + str(count))
        tend = time.clock();
        print("Time used : " + str(tend - tstart))
        print()

对于n=25,迭代了589次，用了0.19s的时间。

6、Exercise 9.5 Generate an n×n matrix, denoted by C, where each entry is 1 with probability p and 0 otherwise. Use the linear algebra library of Scipy to compute the singular values of C. What can you say about the relationship between n, p and the largest singular value?

做法：利用random.random()生成[0,1)的浮点数与p比较，小于p则该位置为1，否则为0，据此生成C矩阵，用svd分解函数scipy.linalg.svd得到s

    def Exercise_9_5(self, n, p):
        """
        Generate an n×n matrix, denoted by C, where each entry is 1 with 
        probability p and 0 otherwise. Use the linear algebra library of
        Scipy to compute the singular values of C. 
        What can you say about the relationship between n, p 
        and the largest singular value?
        """
        print("Exercise_9_5")
        C = npm.zeros((n, n))
        for i in range(0, n):
            for j in range(0, n):
                if random.random() < p:
                    C[i,j] = 1
        (u,s,v) = scipy.linalg.svd(C)
        print("n : " + str(n) + "    p : " + str(p))
        print(s)
        print()

p越接近0.5，其奇异值越大。

7、Exercise9.6Write a function that takes a value z and an array A and finds the element in A that is closest to z. The function should return the closest value, not index.

Hint: Use the built-in functionality of Numpy rather than writing code to find this value manually. In particular, use brackets and argmin.

做法：先得到差值矩阵A-z,将矩阵内元素取绝对值，元素则表示z到A矩阵该位置元素的距离，再利用numpy.argmin函数得到最小值下标，再取元素值则为答案。

    def Exercise_9_6(self, A, z):
        """
        Write a function that takes a value z and an array A and finds 
        the element in A that is closest to z. The function should 
        return the closest value, not index.
        Hint: Use the built-in functionality of Numpy rather than 
        writing code to find this value manually. In particular, 
        use brackets and argmin.
        """
        print("Exercise_9_6")
        print("Matrix A :")
        print(A)
        print("z : " + str(z))
        print(A[0, np.argmin(np.abs(A-z))])
        print()

对于z = 0.473448798866156找到的值为0.5070130468728579