机器学习——朴素贝叶斯

 在考虑一个结果的概率时候，要考虑众多的属性，贝叶斯算法利用所有可能的数据来进行修正预测，如果大量的特征产生的影响较小，放在一起，组合的影响较大，适合于朴素贝叶斯分类

导入类库

from sklearn.datasets import fetch_20newsgroups
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.naive_bayes import MultinomialNB

代码

def article_category():
    categories = ['alt.atheism', 'soc.religion.christian', 'comp.graphics', 'sci.med']
    twenty_train = fetch_20newsgroups(subset='train', categories=categories)
    # print(twenty_train)
    print(twenty_train.data)
    print(twenty_train.target)
    # 将x训练集词频向量化
    tfidf_transformer = TfidfVectorizer()
    X_train_tfidf = tfidf_transformer.fit_transform(twenty_train.data)

    # 贝叶斯训练
    clf = MultinomialNB(alpha=1.0).fit(X_train_tfidf, twenty_train.target)
    docs_new = ['Chemical reaction', 'Intel CPU is good']

    # 将要预测的数据词频向量化
    X_new_tfidf = tfidf_transformer.transform(docs_new)
    # 预测
    predicted = clf.predict(X_new_tfidf)
    print(predicted)
    for doc, category in zip(docs_new, predicted):
        print('%r => %s' % (doc, twenty_train.target_names[category]))