demo1：使用tfidf 处理新闻数据, 并制作模型

# demo1. 使用tfidf 处理新闻数据, 并制作模型

# 导入数据集
from sklearn.datasets import fetch_20newsgroups
# 导入分割数据集函数
from sklearn.model_selection import train_test_split
# 导入特征工程的类
from sklearn.feature_extraction.text import TfidfVectorizer
# 导入估计器对象
from sklearn.naive_bayes import MultinomialNB
# 导入波士顿房价数据
from sklearn.datasets import load_boston
# 导入标准化的数据集
from sklearn.preprocessing import StandardScaler
# 导入线性回归模型
from sklearn.linear_model import  LinearRegression
# 导入损失方程
from sklearn.metrics import mean_squared_error


def mlbcls():

    """
        tfidf 处理数据,制作朴素贝叶斯算法
    :return: 
    """

    # 1.获取数据集
    new = fetch_20newsgroups(subset="all")

    # 2.确认好特征值与目标值
    x = new.data
    y = new.target

    # 3.分割数据集
    x_train,x_test, y_train, y_test = train_test_split(x, y,test_size=0.3)

    # 4.特征工程处理数据
    tfidf = TfidfVectorizer()

    x_train = tfidf.fit_transform(x_train)
    x_test = tfidf.transform(x_test)

    # 5.导入估计器对象制作模型
    mlb = MultinomialNB()

    mlb.fit(x_train, y_train)

    # 6. 计算准确度
    score = mlb.score(x_test, y_test)
    print("模型准确度: ",score)


    return None

# mlbcls()


def line():

    """
        线性回归,预测波士顿房价
    :return: 
    """
    # 1.获取数据集
    lb = load_boston()

    # 2. 确定特征值与目标值
    x = lb.data
    y = lb.target

    # 3. 分割数据集
    x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.3)

    # 4. 特征工程处理数据
    std = StandardScaler()

    x_train = std.fit_transform(x_train)
    x_test = std.transform(x_test)

    # 5. 创建估计器
    liner = LinearRegression(fit_intercept=True)

    liner.fit(x_train, y_train)

    print("模型的回归系数:", liner.coef_)
    print("模型的偏置:", liner.intercept_)

    # 6.计算模型的损失
    y_predict = liner.predict(x_test)
    error = mean_squared_error(y_test, y_predict)
    print("模型的损失:", error)
    return None

line()