【随机梯度下降】理解与使用

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline

from sklearn.linear_model import LinearRegression

data = pd.read_csv('./training data.txt')
data.head(5) 

# 数据映射
for col in data.columns[0:1]:  # 遍历所有类名
#     print(col) 
    u = data[col].unique()  #  得出每个分类下面的种类名称

    def convert(x):  # 将上面得出的u 进行索引映射

        return np.argwhere(u == x)[0,0]  #  将上面得出的u 进行索引映射
    
    data[col] = data[col].map(convert)  # 将上面得出的u 进行索引映射
data.head(5)

# 数据分类   data.iloc[[行],[列]
X = data.iloc[:,[1,2]]
X.head()
print(type(X))
X.head()

# 数据分类
y = data['Gender']
print(type(y))
y.head()

# 切分训练集跟测试集
from sklearn.model_selection import train_test_split

X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.2)  # 切分
print("训练集大小",X_train.shape,y_train.shape)
print("测试集大小",X_test.shape,y_test.shape)

# 2: 梯度下降
from sklearn.linear_model import LinearRegression
from pylab import mpl
mpl.rcParams['font.sans-serif'] = ['FangSong'] # 指定默认字体
mpl.rcParams['axes.unicode_minus'] = False # 解决保存图像是负号'-'显示为方块的问题


# 随机梯度下降  要先对数据进行归一化处理
from sklearn.linear_model import SGDRegressor

from sklearn.preprocessing import StandardScaler

# 归一化数据
std = StandardScaler()
std.fit(X_train) # 计算训练集X数据

X_train_std = std.transform(X_train) # 得出归一化训练集X上的归一化值
X_test_std = std.transform(X_test)   # 得出归一化测试集X上的归一化值

print(X_train_std,X_test_std)

# n_iter代表浏览多少次，默认是5
sgd_reg = SGDRegressor(n_iter=100) # 梯度下降对象实例

sgd_reg.fit(X_train_std, y_train) # 归一化值X训练集 与  y训练集 进行计算训练

res = sgd_reg.score(X_test_std, y_test) # 比较y预测值跟训练之的对比值
print("随机梯度下降预测准确率为:",res)

y_test_std = sgd_reg.predict(X_test_std)  # 预测值
y_test_std

# 梯度下降预测集图
y_data = np.array(y_test_std)
x_data = np.arange(1,len(y_test)+1)

z=np.arange(1,len(y_test)+1)
m=np.array([0.5]*len(y_test))
plt.plot(z,m,label="分割线",color='red')

plt.xlabel("人数")
plt.ylabel("预测性别")
plt.title("大于0.5的是男性，小于0.5的是女性")


plt.scatter(x_data,y_data)