机器学习实战第4章

sumaoqing123

于 2018-03-11 20:29:40 发布

阅读量375

点赞数

本文链接：https://blog.youkuaiyun.com/sumaoqing123/article/details/79519755

版权

本章深入探讨了支持向量机（SVM）的概念，解释了最大间隔分类和核技巧。通过实例展示了如何使用SVM进行二分类及多分类问题的解决，并讨论了SVM在实际应用中的优缺点。同时，章节还涵盖了SVM参数调优和避免过拟合的策略。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

贝叶斯决策理论：选择具有最高概率的决策。

条件概率：
p(gray|bucketB)表示：在已知石头出自B桶的条件下，取出灰色石头的概率。
p(gray|bucketB) = p(gray and bucketB) / p(bucketB)

贝叶斯准则：
p(c|x) = p(x|c) * p(c) / p(x)

import numpy as np
def loadDataSet():
        postingList =[["my", "dog", "has", "flea", "problems", "help", "please"],
             ["maybe", "not", "take", "him", "to", "dog", "park", "stupid"],
             ["my", "dalmation", "is", "so", "cute", "I", "love", "him"],
             ["stop", "posting", "stupid", "worthless", "garbage"],
             ["mr", "licks", "ate", "my", "steak", "how", "to", &#