机器学习
文章平均质量分 91
SnailDove
在学本科生
展开
-
决策树学习的特征选择
此博客停止更新迁移至SnailDove's Blog,查看本文点击此处决策树学习决策树学习通常包含三个方面:特征选择、决策树生成和决策树剪枝。决策树学习思想主要来源于:Quinlan在1986年提出的ID算法、在1993年提出的C4.5算法和Breiman等人在1984年提出的CART算法。特征选择为了解释清楚各个数学概念,引入例子表5.1 贷款申原创 2015-05-02 00:40:01 · 17760 阅读 · 0 评论 -
吴恩达《机器学习》课程总结
The content table of Machine LearningThis course is a coursera version teached by Andrew NG, AP of Stanford University, which corresponds to the full-time campus version CS229 at Stanford university...原创 2018-05-30 09:59:09 · 2975 阅读 · 1 评论 -
机器学习的信息论基础
本文会随着工作学习持续不定期更新,欢迎点击本文原文链接,也欢迎关注个人独立域名博客自信息1在信息论中,自信息(英语:self-information),由克劳德·香农提出,是与概率空间中的单一事件或离散随机变量的值相关的信息量的量度。它用信息的单位表示,例如 bit、nat或是hart,使用哪个单位取决于在计算中使用的对数的底。自信息的期望值就是信息论中的熵,它反映了随机变量采样时的平均不确定...原创 2018-10-09 11:50:54 · 915 阅读 · 0 评论 -
机器学习的微积分基础
本文会一直随工作学习一直更新,欢迎关注个人博客:SnaiDove,以及本文原文链接梯度一个点的切线的斜率与法线的斜率相乘等于-1证明:斜率 k1=tanθk_1=tan\thetak1=tanθ,θ\thetaθ 是倾斜角,对应的法线的倾斜角为 θ+90\theta+90θ+90,那么k1∗k2=tanθ∗tan(θ+90)=tanθ∗(−cotθ)=−1k_1 * k_2=tan\th...原创 2018-10-09 11:50:40 · 693 阅读 · 0 评论 -
PCA(Principal Component Analysis) —— Python Data Science Cookook
Tip before learning about the following example , we need to have the notion ofthe principle of PCA the principle of PCA in english refers to https://en.wikipedia.org/wiki/Principal_component...原创 2017-02-11 15:29:40 · 1142 阅读 · 0 评论 -
《统计学习方法》第9章 EM/GMM/F-MM/GEM
本文原文将书上所有证明给出,点击此处查看原文, 个人技术博客:SnailDove前言EM(期望最大)算法有很多的应用,最广泛的就是混合高斯模型、聚类、HMM等等,本质上就是一种优化算法,不断迭代,获得优值,与梯度下降、牛顿法、共轭梯度法都起到同一类的作用。本文是对李航《统计学习方法》的第9章复习总结,主要内容如下EM(期望最大)算法证明有跳跃性的地方全部事无巨细地写出来,清晰地梳...原创 2018-10-28 18:26:01 · 1399 阅读 · 2 评论 -
《统计学习方法》第4章朴素贝叶斯法与贝叶斯估计
前言写本文章主要目的是复习(毕竟之前看纸质版做的笔记), 对于证明比较跳跃和勘误的地方我都做了注解,以便初学者和以后复习地时候快速阅读理解不会卡住。本文原文将书上所有证明给出,由于CSDN的公式编辑器公式支持不全,有些公式没法正常显示,欢迎点击此处查看原文, 个人技术博客:SnailDove文章目录前言朴素贝叶斯法4.1 朴素贝叶斯法的学习与分类4.1.1 基本方法4.1.2 后验概率最大化...原创 2018-10-28 19:16:11 · 819 阅读 · 0 评论 -
《统计学习方法》第8章 提升方法之AdaBoost\BoostingTree\GBDT
文章目录前沿正文8.1 提升方法AdaBoost算法8.1.1 提升方法的基本思路提升方法的思想历史背景提升方法的核心问题和思想8.1.2 AdaBoost算法8.1.3 AdaBoost的例子8.2 AdaBoost算法的训练误差分析8.3 AdaBoost算法的解释8.3.1 前向分步算法8.3.2 前向分步算法与AdaBoost8.4 提升树8.4.1 提升树模型8.4.2 提升树算法8.4...原创 2018-10-26 15:20:25 · 830 阅读 · 0 评论 -
coursera机器学习公开课笔记:11 machine-learning-system-design
NoteThis personal note is written after studying the opening course on the coursera website, Machine Learning by Andrew NG . And images, audios of this note all comes from the opening course. Tabl...原创 2018-04-27 00:07:08 · 385 阅读 · 0 评论 -
coursera机器学习公开课笔记10: advice-for-applying-machine-learning
NoteThis personal note is written after studying the opening course on the coursera website, Machine Learning by Andrew NG . And images, audios of this note all comes from the opening course. 01_e...原创 2018-04-27 00:06:46 · 298 阅读 · 0 评论 -
coursera机器学习公开课笔记:09 neural-networks-learning
NoteThis personal note is written after studying the opening course on the coursera website, Machine Learning by Andrew NG. And images, audios of this note all comes from the opening course. 01_...原创 2018-04-27 00:06:18 · 255 阅读 · 0 评论 -
PLA算法总结及其证明
此博客停止更新迁移至SnailDove's Blog,查看本文点击此处PLA(Perception Learning Algorithm) 适用于二维及高维的线性可划分问题。问题的答案只有同意或者不同意。例子银行可以根据顾客的个人信息来判断是否给顾客发放信用卡。将顾客抽象为一个向量,包括姓名、年龄、年收入、负债数等。同时设定各个属性所占的权重向量为,对于正相关的属性设置相对较高的权原创 2015-04-24 00:56:57 · 4516 阅读 · 0 评论 -
coursera机器学习公开课笔记2: linear-regression-with-one-variable
NoteThis personal note is written after studying the coursera opening course, Machine Learning by Andrew NG . And images, audios of this note all comes from the opening course. Model Representatio...原创 2018-04-11 01:03:16 · 459 阅读 · 0 评论 -
coursera机器学习公开课笔记3: 03-linear-algebra-review
NoteThis personal note is written after studying the coursera opening course, Machine Learning by Andrew NG . And images, audios of this note all comes from the opening course. Matrices and Vect...原创 2018-04-11 18:34:14 · 390 阅读 · 1 评论 -
coursera机器学习公开课笔记4: linear-regression-with-multiple-variables
Multiple FeaturesNote: [7:25 - θTθTθ^T is a 1 by (n+1) matrix and not an (n+1) by 1 matrix] Linear regression with multiple variables is also known as “multivariate linear regression”. We now...原创 2018-04-11 18:49:59 · 421 阅读 · 0 评论 -
coursera机器学习公开课笔记1:Introduction
What is Machine Learning?Two definitions of Machine Learning are offered. 1. Arthur Samuel described it as: “the field of study that gives computers the ability to learn without being explicitly p...原创 2018-04-05 18:28:37 · 830 阅读 · 0 评论 -
coursera机器学习公开课笔记08: neural-networks-representation
NoteThis personal note is written after studying the opening course on the coursera website, Machine Learning by Andrew NG . And images, audios of this note all comes from the opening course. Moti...原创 2018-04-27 00:05:21 · 486 阅读 · 0 评论 -
coursera机器学习公开课笔记:06 logistic-regression
01_building-a-spam-classifierPrioritizing What to Work OnSystem Design Example: Given a data set of emails, we could construct a vector for each email. Each entry in this vector represents...原创 2018-04-27 00:04:21 · 374 阅读 · 0 评论 -
coursera机器学习公开课笔记:07 regularization
NoteThis personal note is written after studying the opening course on the coursera website, Machine Learning by Andrew NG . And images, audios of this note all comes from the opening course. Th...原创 2018-04-27 00:04:46 · 421 阅读 · 0 评论 -
入门XGBoost
欢迎访问我的技术博客:SnailDove ,本文内容有大量公式,csdn对数学公式支持不好,请访问本文原链接:一起入门xgboost前言在深度学习火起来之前,集成学习 (ensemble learning 包括 boosting: GBDT, XGBoost)是 kaggle 等比赛中的利器,所以集成学习是机器学习必备的知识点,如果提升树或者GBDT不熟悉,最好先看一下我的另一文: 《统计...原创 2018-11-10 19:02:38 · 918 阅读 · 0 评论