上午: Python機器學習套件與資料分析

使用tensorflow.keras
測試不同optimizer、batzh_size影響的數據結果

optimizer = ['rmsprop', 'sgd', 'adam', 'adagrad', 'adamax']
Batch_size = [8, 16, 32, 64]

Loss_test = []
MAE_test = []

#套用不同的optimizer & Batch_size
for opt in optimizer:
    Admission_Model.set_weights(Admission_Model_weights)
    Admission_Model.compile(optimizer = opt, loss = 'mse', metrics = ['mae'])
    Admission_Model_History = Admission_Model.fit(X_train, y_train, epochs = 20, batch_size = 8, verbose = 0, 
                                                 validation_data = (X_test, y_test))
    Result = Admission_Model.evaluate(X_test, y_test)
    Loss_test.append(Result[0])
    MAE_test.append(Result[1])

import matplotlib.pyplot as plt

plt.figure(dpi = 100)
plt.bar(range(len(optimizer)), Loss_test)
plt.xticks(range(len(optimizer)), optimizer)
plt.xlabel('Optimizer')
plt.ylabel('MSE')

plt.show()

下午: Pytorch 與深度學習初探

######################### step1: load data (generate) ############
import torch
import torch.nn as nn
import matplotlib.pyplot as plt
import numpy as np
from sklearn import datasets

def scatter_plot():
  plt.scatter(X[Y==0,0], X[Y==0,1]) # scatter(X1,X2) where y==0
  plt.scatter(X[Y==1,0], X[Y==1,1]) #==> scatter(X[Y==0,0],X[Y==0,1])

def plot_fit(model=0):
  if model==0:
    w1=-7
    w2=5
    b=0
  else:
    [w,b]=model.parameters()
    [w1,w2]=w.view(2)
    w1=w1.detach().item()
    w2=w2.detach().item()
    b=b[0].detach().item()

  x1=np.array([-1.5,1.5])
  x2=(w1*x1+b)/(-1*w2)
  scatter_plot()
  plt.plot(x1,x2,'r')
  plt.show()


# generate data 
n_samples=100
centers=[[-0.5,-0.5],[0.5,0.5]]
X,Y=datasets.make_blobs(n_samples=n_samples,random_state=1,centers=centers,cluster_std=0.2)
print("X=",X.shape,"Y=",Y.shape)

########################### step2: preprocessing X,Y ############
tensorX=torch.Tensor(X)
tensorY=torch.Tensor(Y.reshape(len(X),1))
print(tensorX.shape,tensorY.shape)
print(type(tensorX))

plot_fit(0)

########################### step3: build model ############
class LogRegModel(nn.Module):
 def __init__(self,numIn,numOut):
   super(LogRegModel,self).__init__()
   self.layer1=torch.nn.Linear(numIn,numOut)
 def forward(self,x):
   x=self.layer1(x)
   yPre=torch.sigmoid(x)
   return yPre


model=LogRegModel(2,1)
print(model)
plot_fit(model)

#print("########### parameters=",w,b)
criterion= nn.BCELoss()
optimizer=torch.optim.SGD(model.parameters(),lr=0.01)#advance 2:lr=0.0001

############################ step4: traing model############
torch.manual_seed(1)
epochs=5000
losses=[]

for e in range(epochs):
  preY=model.forward(tensorX)
  loss=criterion(preY,tensorY)
  losses.append(loss)

  optimizer.zero_grad()
  loss.backward()
  optimizer.step()

plt.plot(range(epochs),losses)
plt.show()

######　plot boundary 

plot_fit(model)

# step5: evaluate model (draw)