第七节 朴素贝叶斯算法

https://baike.baidu.com/item/%E6%9C%B4%E7%B4%A0%E8%B4%9D%E5%8F%B6%E6%96%AF/4925905

from sklearn.datasets import load_iris, fetch_20newsgroups
from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.naive_bayes import MultinomialNB  # 朴素贝叶斯API


# 朴素贝叶斯会受到特征之间相关性的影响，各特征相关性越强，预测效果越差，该算法使用简单不需要调参，有坚实的统计学基础有稳定的分类效率，对缺失数据不太敏感，常用于文本分类
def naviebayes():
    '''朴素贝叶斯'''
    news = fetch_20newsgroups(subset='all')

    # 进行训练集和测试集的分割，x特征值，y目标值
    x_train, x_test, y_train, y_test = train_test_split(news.data, news.target, test_size=0.25)

    # 对数据集进行特征抽取
    tf = TfidfVectorizer()

    # 以训练集当中的词列表进行每篇文章重要性统计['a', 'b', 'c', 'd']
    x_train = tf.fit_transform(x_train)
    x_test = tf.transform(x_test)

    # 进行朴素贝叶斯算法预测，alpha拉普拉斯平滑系数
    mlt = MultinomialNB(alpha=1.0)

    mlt.fit(x_train, y_train)
    y_predict = mlt.predict(x_test)

    sorce = mlt.score(x_test, y_test)
    
if __name__ == "__main__":
    naviebayes()