pytorch實(shí)現(xiàn)邏輯回歸

# 使用邏輯回歸完成一個(gè)二分類任務(wù)
# 數(shù)據(jù)準(zhǔn)備
import torch
import matplotlib.pyplot as plt

x1 = torch.randn(365)+1.5   # randn():輸出一個(gè)形狀為size的標(biāo)準(zhǔn)正態(tài)分布Tensor
x2 = torch.randn(365)-1.5
#print(x1.shape)  # torch.Size([365])
#print(x2.shape)  # torch.Size([365])
data = zip(x1.data.numpy(),x2.data.numpy())  # 創(chuàng)建一個(gè)聚合了來自每個(gè)可迭代對(duì)象中的元素的迭代器。 x = [1,2,3]

pos = []
neg = []
def classification(data):
    for i in data:
        if (i[0] > 1.5+0.1*torch.rand(1).item()*(-1)**torch.randint(1,10,(1,1)).item()):
            pos.append(i)
        else:
            neg.append(i)

classification(data)
# 將正、負(fù)兩類數(shù)據(jù)可視化
pos_x = [i[0] for i in pos]
pos_y = [i[1] for i in pos]
neg_x = [i[0] for i in neg]
neg_y = [i[1] for i in neg]
plt.scatter(pos_x,pos_y,c = 'r',marker = "*")
plt.scatter(neg_x,neg_y,c = 'b',marker = "^")
plt.show()

# 構(gòu)造正、負(fù)兩類數(shù)據(jù)可視化結(jié)果如上圖所示

# 構(gòu)建模型
import torch.nn as nn
class LogisticRegression(nn.Module):
    def __init__(self):
        super(LogisticRegression, self).__init__()
        self.linear = nn.Linear(2,1)
        self.sigmoid = nn.Sigmoid()

    def forward(self,x):
        return self.sigmoid(self.linear(x))

model = LogisticRegression()
criterion = nn.BCELoss()
optimizer = torch.optim.SGD(model.parameters(),0.01)
epoch = 5000
features = [[i[0],i[1]] for i in pos]
features.extend([[i[0],i[1]] for i in neg])   #extend 接受一個(gè)參數(shù)，這個(gè)參數(shù)總是一個(gè) list，并且把這個(gè) list 中的每個(gè)元素添加到原 list 中
features = torch.Tensor(features)   # torch.Tensor 生成單精度浮點(diǎn)類型的張量

label = [1 for i in range(len(pos))]
label.extend(0 for i in range(len(neg)))
label = torch.Tensor(label)
print(label.shape)

for i in range(500000):
    out = model(features)
    #print(out.shape)
    loss = criterion(out.squeeze(1),label)
    optimizer.zero_grad()
    loss.backward()
    optimizer.step()
    # 分類任務(wù)準(zhǔn)確率
    acc = (out.ge(0.5).float().squeeze(1)==label).sum().float()/features.size()[0]
    if (i % 10000 ==0):
        plt.scatter(pos_x, pos_y, c='r', marker="*")
        plt.scatter(neg_x, neg_y, c='b', marker="^")
        weight = model.linear.weight[0]
        #print(weight.shape)
        wo = weight[0]
        w1 = weight[1]
        b = model.linear.bias.data[0]
        # 繪制分界線
        test_x = torch.linspace(-10,10,500)   # 500個(gè)點(diǎn)
        test_y = (-wo*test_x - b) / w1
        plt.plot(test_x.data.numpy(),test_y.data.numpy(),c="pink")
        plt.title("acc:{:.4f},loss:{:.4f}".format(acc,loss))
        plt.ylim(-5,3)
        plt.xlim(-3,5)
        plt.show()