吴恩达机器学习课程讲义（翻译版）_吴恩达机器学课程讲义资源-CSDN文库

吴恩达

机器学习

5星 · 超过95%的资源需积分: 49 72 浏览量 2016-09-21 22:44:27 上传评论 5 收藏 3.99MB PDF 举报

资源推荐

资源详情

资源评论

COURSERA 机器学习课笔记

by Prof. Andrew Ng

Notes by Ryan Cheung

Ryanzjlib@gmail.com

Weibo@小小人_V

目录 ............................................................................................................................................................. 1

Week1 机器学习介绍 ............................................................................................................................. 7

1 机器学习介绍..................................................................................................................................... 7

1.1 什么是机器学习? ............................................................................................................................ 7

1.2 监督学习（Supervised Learning） ................................................................................................ 7

1.3 非监督学习（Unsupervised Learning） ........................................................................................ 9

Week1 单变量线性回归 ...................................................................................................................... 11

2 单变量线性回归（Linear Regression with One Variable） ........................................................... 11

2.1 模型表达（Model Representation） ............................................................................................ 11

2.2 代价函数（Cost Function） ......................................................................................................... 12

2.3 梯度下降（Gradient Descent） .................................................................................................... 13

2.4 对线性回归运用梯度下降法........................................................................................................ 13

Week2 多变量线性回归 ...................................................................................................................... 15

3 多变量线性回归（Linear Regression with Multiple Variables） .................................................. 15

3.1 多维特征（Multiple Features） .................................................................................................. 15

3.2 多变量梯度下降（Gradient descent for multiple variables） ..................................................... 16

3.3 特征缩放（feature scaling） ........................................................................................................ 16

3.4 学习率(Learning rate) ................................................................................................................... 17

Week2 多项式回归和正规方程 ......................................................................................................... 19

4 多项式回归和正规方程................................................................................................................... 19

4.1 多项式回归（Polynomial Regression） ...................................................................................... 19

4.2 正规方程（Normal Equation） .................................................................................................... 19

Week3 归一化 ...................................................................................................................................... 21

5 逻辑回归（Logistic Regression） .................................................................................................. 21

5.1 分类问题........................................................................................................................................ 21

5.2 分类问题建模................................................................................................................................ 21

5.3 判定边界（Decision Boundary）................................................................................................. 23

5.4 代价函数........................................................................................................................................ 24

5.5 多类分类（Multiclass Classification） ........................................................................................ 26

Week3 归一化 ...................................................................................................................................... 28

6 归一化（Regularization） ............................................................................................................... 28

6.1 过拟合问题（The Problem of Overfitting） ................................................................................ 28

6.2 归一化代价函数（Regularization Cost Function） .................................................................... 29

6.3 归一化线性回归（Regularized Linear Regression） .................................................................. 30

6.4 归一化逻辑回归（Regularized Logistic Regression） ............................................................... 31

Week4 神经网络：表达 ...................................................................................................................... 32

7 神经网络：表达 .............................................................................................................................. 32

7.1 非线性假设（Non-Linear Hypothesis） ...................................................................................... 32

7.2 神经网络介绍................................................................................................................................ 32

7.3 模型表达........................................................................................................................................ 33

7.4 神经网络模型表达........................................................................................................................ 34

7.5 正向传播 (Forward Propagation) ................................................................................................. 35

7.6 对神经网络的理解........................................................................................................................ 36

7.7 神经网络示例：二元逻辑运算符（Binary Logical Operators） ............................................... 36

7.8 多类分类........................................................................................................................................ 37

Week5 神经网络：学习 ...................................................................................................................... 39

8 神经网络：学习 .............................................................................................................................. 39

8.1 神经网络代价函数........................................................................................................................ 39

8.2 反向传播算法（Backpropagation Algorithm）........................................................................... 39

8.3 梯度检验（Gradient Checking） ................................................................................................. 42

8.4 随机初始化（Random Initialization） ......................................................................................... 43

8.5 综合起来........................................................................................................................................ 43

Week6 机器学习应用建议 .................................................................................................................. 45

9 机器学习应用建议........................................................................................................................... 45

9.1 决定下一步做什么........................................................................................................................ 45

9.2 假设的评估（Evaluating a Hypothesis） .................................................................................... 45

9.3 模型选择（交叉验证集）............................................................................................................ 46

9.4 偏倚和偏差诊断（Diagnosis Bias vs. Variance） ...................................................................... 46

9.5 归一化与偏倚/偏差 ...................................................................................................................... 47

9.6 学习曲线（Learning Curves） ..................................................................................................... 49

9.7 决定下一步做什么........................................................................................................................ 50

Week6 机器学习系统设计 .................................................................................................................. 52

10 机器学习系统设计......................................................................................................................... 52

10.1 首先要做什么.............................................................................................................................. 52

10.2 误差分析（Error Analysis） ...................................................................................................... 52

10.3 类偏斜的误差度量（Error Metrics for Skewed Classes） ........................................................ 53

10.4 查全率和查准率之间的权衡...................................................................................................... 54

10.5 机器学习的数据.......................................................................................................................... 55

Week7 支持向量机 .............................................................................................................................. 56

11 支持向量机（Support Vector Machine） .................................................................................... 56

11.1 优化目标（Optimization Objective） ........................................................................................ 56

11.2 支持向量机判定边界（SVM Decision Boundary） ................................................................. 58

11.3 核函数（Kernels） ..................................................................................................................... 59

11.4 逻辑回归与支持向量机.............................................................................................................. 63

Week8 聚类 ........................................................................................................................................... 64

12 聚类（Clustering） ........................................................................................................................ 64

12.1K-均值算法 .................................................................................................................................. 64

12.2 优化目标...................................................................................................................................... 66

12.3 随机初始化.................................................................................................................................. 66

12.4 选择聚类数.................................................................................................................................. 66

Week8 降维 ........................................................................................................................................... 67

13 降维（Dimensionality Reduction） .............................................................................................. 67

13.1 动机一：数据压缩（Data Compression） ............................................................................... 67

13.2 动机二：数据可视化（Data Visualization） ............................................................................ 68

13.3 主要成分分析（Principal Component Analysis） .................................................................... 68

13.4 主要成分分析算法...................................................................................................................... 69

剩余92页未读，继续阅读

评论收藏

内容反馈