Weak11 Numpy Homework

最新推荐文章于 2021-01-15 00:12:38 发布

GH_Loeng

最新推荐文章于 2021-01-15 00:12:38 发布

阅读量334

点赞数 1

分类专栏： Python Homework

本文链接：https://blog.csdn.net/qq_36475045/article/details/80374717

版权

Python 同时被 2 个专栏收录

20 篇文章 0 订阅

订阅专栏

Homework

16 篇文章 0 订阅

订阅专栏

这里写图片描述

为了显示方便，n=4,m=10
代码：

import numpy as np
from scipy.linalg import toeplitz
import random
n=4
m=10
A = np.random.normal(0,1,n*m)
A = A.reshape(n,m)
B = toeplitz(list(range(1,m+1)))
print("A=",A)
print("B=",B)
print("A+A=",A+A)  #A+A
print("A*A.T=",A@A.T)    #A*A.T
print("A.T*A=",A.T@A)    #A.T*A
print("AB=",A@B)  #AB

def cal(A,B,lamb):
    return  A@(B-lamb*np.eye(m))  #A(B−λI)
print("A(B−λI)=",cal(A,B,random.random()))

输出：

A= [[-0.14127543 -0.18536895 -0.3919833   0.3917631   0.27335622 -1.95794558
   1.1441752  -0.63235459  0.1720679  -0.97279904]
 [-1.38390472 -0.16494005 -0.85358231 -0.31362912 -0.44616255  0.26444861
   0.08132372  1.16820247 -1.91757158  1.07926125]
 [-0.2395814  -0.35344047  0.27283204  1.05136257 -1.504381   -0.19087067
  -1.41278342  0.17866861 -0.49925697  0.81361346]
 [ 0.38664661 -0.01007491  1.69917109  1.72546369 -0.28064128  0.15060185
   0.84807271  1.42510812 -1.42309945 -0.50420029]]
B= [[ 1  2  3  4  5  6  7  8  9 10]
 [ 2  1  2  3  4  5  6  7  8  9]
 [ 3  2  1  2  3  4  5  6  7  8]
 [ 4  3  2  1  2  3  4  5  6  7]
 [ 5  4  3  2  1  2  3  4  5  6]
 [ 6  5  4  3  2  1  2  3  4  5]
 [ 7  6  5  4  3  2  1  2  3  4]
 [ 8  7  6  5  4  3  2  1  2  3]
 [ 9  8  7  6  5  4  3  2  1  2]
 [10  9  8  7  6  5  4  3  2  1]]
A+A= [[-0.28255086 -0.37073789 -0.78396659  0.78352621  0.54671244 -3.91589116
   2.28835041 -1.26470919  0.3441358  -1.94559809]
 [-2.76780944 -0.32988011 -1.70716461 -0.62725824 -0.89232511  0.52889721
   0.16264745  2.33640494 -3.83514316  2.15852249]
 [-0.4791628  -0.70688094  0.54566408  2.10272513 -3.00876199 -0.38174135
  -2.82556684  0.35733721 -0.99851394  1.62722692]
 [ 0.77329322 -0.02014982  3.39834218  3.45092739 -0.56128256  0.3012037
   1.69614542  2.85021623 -2.84619889 -1.00840059]]
A*A.T= [[ 6.95467871 -2.22745554 -2.24005637 -0.09962909]
 [-2.22745554  9.25155372  2.37724554  1.55859419]
 [-2.24005637  2.37724554  6.60081763  1.93879692]
 [-0.09962909  1.55859419  1.93879692 11.14503533]]
A.T*A= [[ 2.14204587  0.33523173  1.82826745  0.79394409  0.83074066  0.01459668
   0.39219294 -1.0191372   2.19880351 -1.74603612]
 [ 0.33523173  0.18658854  0.09990256 -0.40986868  0.55745488  0.38526826
   0.26528249 -0.15297102  0.47518353 -0.28017082]
 [ 1.82826745  0.09990256  3.84387337  3.33284709 -0.61361549  0.74557598
   0.53765399  1.72096451 -0.9849453  -1.17466007]
 [ 0.79394409 -0.40986868  3.33284709  4.33442976 -1.81886575 -0.79080588
   0.40071121  2.03270229 -2.31159037 -0.73417109]
 [ 0.83074066  0.55745488 -0.61361549 -1.81886575  2.61570636 -0.40832655
   2.16384413 -1.36279608  2.05303761 -1.82993184]
 [ 0.01459668  0.38526826  0.74557598 -0.79080588 -0.40832655  3.96259649
  -1.82134659  1.72756672 -0.96302661  1.95886828]
 [ 0.39219294  0.26528249  0.53765399  0.40071121  2.16384413 -1.82134659
   4.03093476  0.32765339 -0.46061807 -2.60234112]
 [-1.0191372  -0.15297102  1.72096451  2.03270229 -1.36279608  1.72756672
   0.32765339  3.82742496 -4.4661919   1.30277685]
 [ 2.19880351  0.47518353 -0.9849453  -2.31159037  2.05303761 -0.96302661
  -0.46061807 -4.4661919   5.98115767 -1.92562321]
 [-1.74603612 -0.28017082 -1.17466007 -0.73417109 -1.82993184  1.95886828
  -2.60234112  1.30277685 -1.92562321  3.02732762]]
AB= [[-15.73079285 -13.71297924 -12.06590353 -11.20279441  -9.55615909
   -7.36281132  -9.08535471  -8.51954769  -9.21844986  -9.57321624]
 [ -2.72381528  -3.00507044  -3.6162057   -5.93450558  -8.88006369
  -12.71794692 -16.02693294 -19.1732715  -19.98320512 -24.62828191]
 [ -9.4069582   -8.00228375  -7.30449023  -6.06103263  -2.7148499
   -2.37742916  -2.42174976  -5.29163721  -7.80418744 -11.31525162]
 [ 11.35374551   8.10999059   4.84608584   4.98052327   8.56588809
   11.58997036  14.91525632  19.93668771  27.80833533  32.83378406]]
A(B−λI)= [[-15.62625848 -13.57581863 -11.77586213 -11.49267289  -9.7584244
   -5.91406263  -9.93196778  -8.05164762  -9.34576859  -8.85341005]
 [ -1.69981841  -2.88302584  -2.98461192  -5.70244102  -8.54993327
  -12.91362119 -16.08710705 -20.03766314 -18.56433049 -25.426863  ]
 [ -9.229684    -7.74076146  -7.50636768  -6.83897056  -1.60170864
   -2.23619763  -1.37638458  -5.42384002  -7.4347707  -11.91727113]
 [ 11.0676529    8.11744534   3.58881296   3.70379562   8.77354386
   11.47853507  14.28773927  18.88220308  28.86133367  33.20685853]]

这里写图片描述
代码：

import numpy as np
from scipy.linalg import toeplitz
import random
n=4
m=10
A = np.random.normal(0,1,n*m)
A = A.reshape(n,m)
B = toeplitz(list(range(1,m+1)))
b = np.array(range(0,m))
print("A=",A)
print("B=",B)
print("b=",b)
print("B\\b=",np.linalg.solve(B,b))
print("Bx=",B@np.linalg.solve(B,b))

输出：

A= [[ 1.10072277 -1.74155593 -0.87292138  2.36295768  1.06313563 -1.09282835
  -1.65235618 -1.29662255  1.02735804  0.92605323]
 [-0.20442559 -1.13772172 -0.09905539 -2.16642607  0.05472793  1.26894958
   0.02942699  0.06011498  1.61455627 -0.53479065]
 [ 0.09988037  0.78840479 -0.22336113  1.17640885  0.65951768 -0.35056981
   0.27483197 -1.74387682 -0.21052275 -0.51640518]
 [-0.82507688 -0.2172729  -1.12598322  1.04701727  0.46405799 -1.51862459
   1.44115554 -1.92401456  0.08674978 -0.15373378]]
B= [[ 1  2  3  4  5  6  7  8  9 10]
 [ 2  1  2  3  4  5  6  7  8  9]
 [ 3  2  1  2  3  4  5  6  7  8]
 [ 4  3  2  1  2  3  4  5  6  7]
 [ 5  4  3  2  1  2  3  4  5  6]
 [ 6  5  4  3  2  1  2  3  4  5]
 [ 7  6  5  4  3  2  1  2  3  4]
 [ 8  7  6  5  4  3  2  1  2  3]
 [ 9  8  7  6  5  4  3  2  1  2]
 [10  9  8  7  6  5  4  3  2  1]]
b= [0 1 2 3 4 5 6 7 8 9]
B\b= [ 9.09090909e-01  5.04646829e-17 -1.11022302e-16  1.11022302e-16
 -2.77555756e-16  1.11022302e-16  3.60822483e-16  5.55111512e-17
 -4.71844785e-16 -9.09090909e-02]
Bx= [4.4408921e-16 1.0000000e+00 2.0000000e+00 3.0000000e+00 4.0000000e+00
 5.0000000e+00 6.0000000e+00 7.0000000e+00 8.0000000e+00 9.0000000e+00]

这里写图片描述
代码：

import numpy as np
from scipy.linalg import toeplitz
import random
n=4
m=10
A = np.random.normal(0,1,n*m)
A = A.reshape(n,m)
B = toeplitz(list(range(1,m+1)))
print("A=",A)
print("B=",B)
print("||A||F=",np.linalg.norm(A,'fro')) #2-范数
print("||B||∞=",np.linalg.norm(B,np.inf))   #无穷范数
u,sigma,v = np.linalg.svd(B)    #奇异值分解
print("Max singular value = ",max(sigma))
print("Min singular value = ",min(sigma))

输出：

A= [[ 1.76739889 -0.61062597  0.02890403 -0.73851563 -0.67586851  0.63724716
   0.08316141 -1.5604136   1.55553185  0.22876358]
 [-0.35941143 -0.80682195  0.09212975 -1.36546868  1.06930886  1.71811841
  -0.82373042 -0.83853979 -1.7913538  -0.26047562]
 [ 0.90643659 -0.74680685 -1.00215034  0.45642739  1.76741504  1.23191851
   2.03571596  0.2671933   1.85416188  0.49919575]
 [-1.46066207  0.28280865 -0.41617868  1.73892589 -0.66836749  0.06630803
  -0.03678402 -0.0268019   1.28013338  0.27449861]]
B= [[ 1  2  3  4  5  6  7  8  9 10]
 [ 2  1  2  3  4  5  6  7  8  9]
 [ 3  2  1  2  3  4  5  6  7  8]
 [ 4  3  2  1  2  3  4  5  6  7]
 [ 5  4  3  2  1  2  3  4  5  6]
 [ 6  5  4  3  2  1  2  3  4  5]
 [ 7  6  5  4  3  2  1  2  3  4]
 [ 8  7  6  5  4  3  2  1  2  3]
 [ 9  8  7  6  5  4  3  2  1  2]
 [10  9  8  7  6  5  4  3  2  1]]
||A||F= 6.628749769033752
||B||∞= 55.0
Max singular value =  44.076292527938
Min singular value =  0.5125428154684579

这里写图片描述
题目是想算出主特征值和主特征向量。算法是用幂法求解。
算法：

输入：A；输出：λ和X
u:=随机向量
While 不满足判停准则 do
    v:=Au
    λ:=max(v)
    u:=v/λ
End
x:=u

max(v)为向量v的绝对值最大的分量，即max（v)=vj，其中j满足|vj|=max(1≤k≤n)|vk|,若j的值不唯一，则取最小的那个，并且称u=v/max(v)为向量v的规格化向量。
代码：

import numpy as np
from scipy.linalg import toeplitz
import random
import time
n=100
Z = np.random.normal(0,1,n*n)
Z = Z.reshape(n,n)
def powerIteration(A):
    begintime = time.clock()
    u = np.random.normal(0,1,n)
    lamb=0
    lastlamb=0
    count=0
    while abs(lastlamb-lamb)>=1e-6 or count==0:
        v = A@u
        lastlamb=lamb
        j = np.argmax(abs(v))
        lamb = v[j]
        u = v/lamb
        count += 1
    x = u
    endtime=time.clock()
    return x,lamb,count,endtime-begintime
x,lamb,count,time=powerIteration(Z)
print("主特征值=",lamb)
#print("主特征向量=",x)
print("迭代次数=",count)
print("花费时间=",time)

可以通过修改n来测试算法
输出：
精度要求6位小数
由于数据是随机的，因此时间上没有可比性

矩阵大小= 10
主特征值= -3.6392824996637034
主特征向量= [ 1.          0.14361591 -0.19157489  0.03119712  0.52552813 -0.36625306
  0.94472046 -0.72621135 -0.14788755 -0.44033934]
迭代次数= 69
花费时间= 0.00036187682913724775

矩阵大小= 50
主特征值= 7.766544319679567
迭代次数= 238
花费时间= 0.0013120010366427944

矩阵大小= 100
主特征值= 10.500415332054505
迭代次数= 320
花费时间= 0.0225031288913611

矩阵大小= 1000
主特征值= 32.54980745398664
迭代次数= 447
花费时间= 0.14958747621726515

这里写图片描述
代码：

import numpy as np
from scipy.linalg import toeplitz
import random
import time
for n in (10,50,100):
    for p in (0.1,0.5,0.9):
        print("n=%d,p=%.1f"%(n,p))
        C = np.random.binomial(1,p,n*n)
        C = C.reshape(n,n)
        u,sigma,v = np.linalg.svd(C)
        print("最大特异值=",max(sigma))

输出：

n=10,p=0.1
最大特异值= 1.9021130325903073
n=10,p=0.5
最大特异值= 4.859185150253102
n=10,p=0.9
最大特异值= 9.101742974321443
n=50,p=0.1
最大特异值= 6.189969662651178
n=50,p=0.5
最大特异值= 25.767586355786325
n=50,p=0.9
最大特异值= 44.688202529805025
n=100,p=0.1
最大特异值= 10.733560377216394
n=100,p=0.5
最大特异值= 49.55564032221308
n=100,p=0.9
最大特异值= 90.75654035220786

猜测关系是最大特异值=n*p

这里写图片描述

argmin返回最小值的下标，因此求出两个数的距离的绝对值再调用argmin就知道是哪个数了。
代码：

import numpy as np
from scipy.linalg import toeplitz
import random
import time
z = 5
A = np.random.normal(5,5,10)
print("A=",A)
print("z=",z)
print("|A-z|=",abs(A-z))
print(np.argmin(abs(A-z)))

输出：

A= [ 3.40970621  3.3574148   7.01400765 -1.56122245  7.15815137  7.22773429
 11.59661275  3.49448966 -2.19594911  1.23482191]
z= 5
|A-z|= [1.59029379 1.6425852  2.01400765 6.56122245 2.15815137 2.22773429
 6.59661275 1.50551034 7.19594911 3.76517809]
7

GH_Loeng

关注

1
点赞
踩
0

收藏

觉得还不错? 一键收藏
0
评论
Weak11 Numpy Homework

为了显示方便，n=4,m=10 代码：import numpy as npfrom scipy.linalg import toeplitzimport randomn=4m=10A = np.random.normal(0,1,n*m)A = A.reshape(n,m)B = toeplitz(list(range(1,m+1)))print(&amp;amp;quot;A=&amp;amp;quot;,A)pr...
复制链接

扫一扫