TEST1-pytorch

Ir1.py

@@ -0,0 +1,99 @@
+import numpy as np
+import torch
+import torch.nn.functional as F
+import csv
+import pandas as pd
+from torch.autograd import Variable
+
+
+j = 0
+j2= 0
+right=0;
+X=np.zeros((469,30))*2
+test=np.zeros((100,30))*2
+df = pd.read_csv('data.csv',header=None)
+list1=df.values.tolist()
+
+#训练集
+for i in list1[1:470]: 
+  X[j]=i[2:32]
+  j=j+1
+for i in range(X.shape[1]):
+  mean = np.mean(X[:,i])
+  std=np.std(X[:,i])
+  X[:,i]=(X[:,i]-mean)/std
+print(X)
+
+#测试集
+for i in list1[471:570]:
+  test[j2]=i[2:32]
+  j2=j2+1
+for i in range(test.shape[1]):
+  mean = np.mean(test[:,i])
+  std=np.std(test[:,i])
+  test[:,i]=(test[:,i]-mean)/std
+print(test)
+
+# labels: first N/2 are 0, last N/2 are 1
+T = np.array([0]*(189) + [1]*(280)).reshape(469,1)
+
+x_data = Variable(torch.Tensor(X))
+y_data = Variable(torch.Tensor(T))
+print(x_data)
+print(y_data)
+class Model(torch.nn.Module):
+    def __init__(self):
+        super(Model, self).__init__()
+        self.linear = torch.nn.Linear(30, 1) # 2 in and 1 out
+
+    def forward(self, x):
+        #y_pred = torch.sigmoid(self.linear(x))
+        y_pred = self.linear(x).sigmoid()
+        return y_pred
+
+# Our model    
+model = Model()
+
+criterion = torch.nn.BCELoss(reduction="mean")
+optimizer = torch.optim.SGD(model.parameters(), lr=0.01)
+
+# Training loop
+for epoch in range(100):
+    # Forward pass: Compute predicted y by passing x to the model
+    y_pred = model(x_data)
+    #print(y_pred)
+    # Compute and print loss
+    loss = criterion(y_pred, y_data)
+    print(epoch, loss.data.item())
+
+    # Zero gradients, perform a backward pass, and update the weights.
+    optimizer.zero_grad()
+    loss.backward()
+    optimizer.step()
+
+for f in model.parameters():
+    print('data is')
+    print(f.data)
+    print(f.grad)
+
+w = list(model.parameters())
+w0 = w[0].data.numpy()
+w1 = w[1].data.numpy()
+import matplotlib.pyplot as plt
+
+print("Final gradient descend:", w)
+j3 = 0
+for i in range(10000):
+  if (j3<=23)and (torch.matmul(w[0],Variable(torch.Tensor(test[j3])))<0.5): right=right+1
+  elif(j3>23)and (torch.matmul(w[0],Variable(torch.Tensor(test[j3])))>0.5): right=right+1
+  j3=j3+1
+print(right/100)
+#plot the data and separating line
+plt.scatter(X[:,0], X[:,1], c=T.reshape(len(x_data)), s=469, alpha=0.5)
+x_axis = np.linspace(-6, 6, 469)
+y_axis = -(w1[0] + x_axis*w0[0][0]) / w0[0][1]
+line_up, = plt.plot(x_axis, y_axis,'r--', label='gradient descent')
+plt.legend(handles=[line_up])
+plt.xlabel('X(1)')
+plt.ylabel('X(2)')
+plt.show()

README.txt

@@ -0,0 +1,4 @@
+������ʦ�ã�һ��ʼ�����õ�matlab��ʵ�ֵ�Բpacking�����⣬Ȼ��������ͬ��˼·��python��ʵ�֣�
+������matlab��ʮ�ּ򵥵�ʵ�ֺ��ҷ���ʹԲ�ƶ�������λ�õ��ݶȺ���ȡ����������ȷ����������һ��
+ȷ����ֵ�����Խ����������������ʮ�����⡣��python��ʵ��ʱ������һЩ���⣬����һ�³���ԭ�����
+�Ƿ��ŵ����⣬����û���ܹ��޸���ȷȻ����ȷ���ܳ������ļ��и�����һЩ��ͼ������matlab��ʵ�ֵ�ͼ��

TEST1-README.md

@@ -0,0 +1,4 @@
+非常抱歉超过了提交时间，在此深表悔意！！（⭐）
+需要配合自我自行整理后的包含在该文件夹中的data.csv文件共同使用。
+训练集梯度下降良好
+测试集矩阵相乘的相关代码不熟悉，导致算出的正确率不符合预期，只有54％〜64％左右，需要在此做进一步的改进

homework/aaa.txt

@@ -0,0 +1 @@
+a biney cindy ca t

homework/count_word.py

@@ -0,0 +1,11 @@
+def count_char(file):
+       import os.path
+       if os.path.isfile(file):
+          with open(file,'r') as  opf:
+            total =0
+            for line in opf:
+             for char in('#','\n'):
+               line=line.replace(char," ")
+               word=line.split()
+               total +=len(word)
+             return total

homework/test.txt

@@ -0,0 +1 @@
+It's freezing, snowing and getting dark. This is the last day of the year - New Year's Eve. On this cold and dark night, a cute little girl was walking barefoot in the street. She came out of the house wearing a pair of slippers, but what's the use?

packing.py

@@ -0,0 +1,65 @@
+import matplotlib.pyplot as plt 
+
+from matplotlib.patches import Rectangle  
+
+class Annotate(object):
+
+    def __init__(self):
+
+        self.ax = plt.gca()
+
+        self.rect = Rectangle((-1,-1), 0, 0)
+
+        self.x0 = None         
+
+        self.y0 = None         
+
+        self.x1 = None         
+
+        self.y1 = None         
+
+        self.ax.add_patch(self.rect)
+
+        self.ax.figure.canvas.mpl_connect('button_press_event', self.on_press)
+
+        self.ax.figure.canvas.mpl_connect('button_release_event', self.on_release)
+
+
+
+
+
+    def on_press(self, event):
+
+        print ('press')         
+
+        self.x0 = event.xdata         
+
+        self.y0 = event.ydata      
+
+
+
+
+
+    def on_release(self, event):         
+
+        print ('release')         
+
+        self.x1 = event.xdata         
+
+        self.y1 = event.ydata         
+
+        self.rect.set_width(self.x1 - self.x0)
+
+        self.rect.set_height(self.y1 - self.y0)         
+
+        self.rect.set_xy((self.x0, self.y0))         
+
+        self.ax.figure.canvas.draw()  
+
+
+
+
+
+a = Annotate() 
+
+plt.show()

unit4/tests/.travis.yml

@@ -0,0 +1,14 @@
+language: python
+python:
+  - "3.6"      # current default Python on Travis CI
+  - "3.7"
+  - "3.8"
+  - "3.8-dev"  # 3.8 development branch
+  - "nightly"  # nightly build
+# command to install dependencies
+# install:
+
+# command to run tests
+script:
+  - pytest ./unit4/tests/
+

unit4/tests/README.txt

@@ -0,0 +1,5 @@
+i have met some problem about the system which can't install the module automatically in the virtual machine, and this module
+in my code is matplotlib, i don't know how to operate the travis CI to pip install it so i have to wipe some relative codes out and 
+i can't have some picture in that website, but it doesn't mean there are some mathmatic wrong in my code, because my computer
+has installed the tensflow and matplotlib so if i use the pytest directly, it will work, so i suggest run the pytest directly instead of the
+travis, sincerely.

unit4/tests/test_xmsj_matlab.py

@@ -0,0 +1,48 @@
+import numpy as np
+import random as rd
+import math
+import matplotlib
+from math import pi
+from numpy import cos, sin
+from matplotlib import pyplot as plt
+
+m=5
+r=0
+e=0.01
+Px=2*np.random.rand(m)-1
+Py=2*np.random.rand(m)-1
+U=np.zeros([2,5])
+Dijx=0
+Dijy=0
+dijx=0
+dijy=0
+while (abs((4 - m*pi*r*r) / 4) > 0.1):
+  for i in range(m):
+   for j in range(m):
+    if i != j:
+      dijx=(Px[i] - Px[j]) / abs(Px[i] - Px[j]) * max(2 * r - math.sqrt((Px[i] - Px[j]) ** 2 + (Py[i] - Py[j]) ** 2),0)
+      dijy=(Py[i] - Py[j]) / abs(Py[i] - Py[j]) * max(2 * r - math.sqrt((Px[i] - Px[j]) ** 2 + (Py[i] - Py[j]) ** 2),0)
+      Dijx=Dijx + dijx
+      Dijy=Dijy + dijy
+   X_dijboundary=- max(Px[i] + r - 1,0) + max(- (Px[i] - r + 1),0)
+   Y_dijboundary=- max(Py[i] + r - 1,0) + max(- (Py[i] - r + 1),0)
+   U[0][i]=Dijx + X_dijboundary
+   U[1][i]=Dijy + Y_dijboundary
+   print(U)
+  if U.all() < 0.001:
+   r=r + 0.0001
+  for h in range(m):
+   Px[h]=Px[h] + e*U[0][h]
+   Py[h]=Py[h] + e*U[1][h]
+  print(r)
+  print(abs(4 - m*pi*r*r / 4))
+theta=[i*pi/180 for i in range(0,360)]
+fig=plt.figure()
+for q in range(m):
+   x=sin(theta)
+   y=cos(theta)
+   plt.plot(Px[q]+r*x,Py[q]+r*y,'r')
+
+plt.ion()
+plt.pause(5)
+plt.close()

xmsj_matlab.m

@@ -0,0 +1,40 @@
+clear
+m=5;
+r=0;
+e=0.001;
+Px=2*rand(1,m)-1;
+Py=2*rand(1,m)-1;
+U(2,m)=0;
+Dijx=0;
+Dijy=0;
+while abs((4-m*pi*r*r)/4)>0.1
+for i=1:m
+    for j=1:m
+        if i~=j
+        dijx=(Px(i)-Px(j))/abs(Px(i)-Px(j))*max(2*r-sqrt((Px(i)-Px(j))^2+(Py(i)-Py(j))^2),0);
+        dijy=(Py(i)-Py(j))/abs(Py(i)-Py(j))*max(2*r-sqrt((Px(i)-Px(j))^2+(Py(i)-Py(j))^2),0);
+        Dijx=Dijx+dijx;%��Բ��֮��ĵ�������
+        Dijy=Dijy+dijy;
+        end
+    end
+    X_dijboundary=-max(Px(i)+r-1,0)+max(-(Px(i)-r+1),0);
+    Y_dijboundary=-max(Py(i)+r-1,0)+max(-(Py(i)-r+1),0);
+    %����ΪijԲ�ζԱ߽�ĵ�������
+    U(1,i)=Dijx+X_dijboundary;%�������ܣ�������Բ��֮��ͱ߽�֮��
+    U(2,i)=Dijy+Y_dijboundary;
+end
+if U<0.0001
+    r=r+0.00001;
+end
+for i=1:m
+    Px(i)=Px(i)+e*U(1,i);
+    Py(i)=Py(i)+e*U(2,i);
+end
+end
+theta=0.01:0.01:2*pi;
+for i=1:m
+    plot(Px(i)+r*sin(theta),Py(i)+r*cos(theta),'Linewidth',1);
+    hold on;
+end
+A=[1,1,-1,-1,1;1,-1,-1,1,1];
+plot(A(1,:),A(2,:))

xmsj_matlab.py

@@ -0,0 +1,37 @@
+import numpy as np
+import random as rd
+import math
+from numpy import cos, sin
+from matplotlib import pyplot as plt
+m=5
+r=0
+e=0.001
+Px=np.dot(2,np.random.randn(10)) - 1
+Py=np.dot(2,np.random.randn(10)) - 1
+U=np.ones((2,m))*0
+Dijx=0
+Dijy=0
+while abs((4 - np.dot(np.dot(np.dot(m,math.pi),r),r)) / 4) > 0.1:
+ for i in range(m):
+  for j in range(m):
+   if i != j:
+     dijx=np.dot((Px(i) - Px(j)) / abs(Px(i) - Px(j)),max(np.dot(2,r) - sqrt((Px(i) - Px(j)) ** 2 + (Py(i) - Py(j)) ** 2),0))
+     dijy=np.dot((Py(i) - Py(j)) / abs(Py(i) - Py(j)),max(np.dot(2,r) - sqrt((Px(i) - Px(j)) ** 2 + (Py(i) - Py(j)) ** 2),0))
+     Dijx=Dijx + dijx
+     Dijy=Dijy + dijy
+     X_dijboundary=- max(Px(i) + r - 1,0) + max(- (Px(i) - r + 1),0)
+     Y_dijboundary=- max(Py(i) + r - 1,0) + max(- (Py(i) - r + 1),0)
+     U[0,i]=Dijx + X_dijboundary
+     U[1,i]=Dijy + Y_dijboundary
+     if U < 0.0001:
+        r=r + 1e-05
+   for h in range(m):
+       Px[h]=Px[h] + np.dot(e,U(0,h))
+       Py[h]=Py[h] + np.dot(e,U(1,h))
+theta=[i*math.pi/180 for i in range(0,360)]
+for q in range(m):
+    plot(Px(q) + np.dot(r,sin(theta)),Py(q) + np.dot(r,cos(theta)),'r')
+    hold('on')
+plt.axis('equal')
+plt.axis('scaled')
+plt.show()