【机器学习】吴恩达机器学习Lecture1

目录

[1-1 Introduction](#1-1 Introduction)

[Machine Learning](#Machine Learning)

Examples:

[1-2 What is machine learning](#1-2 What is machine learning)

[Machine learning definition](#Machine learning definition)

Question:

[Machine learning algorithms:](#Machine learning algorithms:)

[1-3 Supervised learning](#1-3 Supervised learning)

[Housing price prediction](#Housing price prediction)

[Breast cancer (malignant, benign)](#Breast cancer (malignant, benign))

Question:

[1-4 Unsupervised Learning](#1-4 Unsupervised Learning)

Examples:

[Cocktail party problem](#Cocktail party problem):

Question:

---

1-1 Introduction

Machine Learning

• Grew out of work in AI

• New capability for computers

Examples:

- Database mining 数据挖掘

Large datasets from growth of automation/web.

E.g., Web click data, medical records, biology, engineering

- Applications can't program by hand. 无法手动编程实现的应用

E.g., Autonomous helicopter, handwriting recognition, most of

Natural Language Processing (NLP), Computer Vision.

- Self-customizing programs 自适应程序

E.g., Amazon, Netflix product recommendations

- Understanding human learning (brain, real AI). 理解人类学习机制

---

1-2 What is machine learning

Machine learning definition

• Arthur Samuel (1959). Machine Learning: Field of study that gives computers the ability to learn

without being explicitly programmed.

• Tom Mitchell (1998) Well-posed Learning Problem: A computer program is said to learnfrom experience E with respect to some task Tand some performance measure P, if itsperformance on T, as measured by P, improveswith experience E.

Question:

Suppose your email program watches which emails you do or do not mark as spam, and based on that learns how to better filter spam. What is the task T in this setting?

Machine learning algorithms:

-Supervised learning 监督学习

-Unsupervised learning 无监督学习

Others: Reinforcement learning, recommender systems. 强化学习，推荐系统

Also talk about: Practical advice for applying learning algorithms.

---

1-3 Supervised learning

Housing price prediction

Supervised Learning: "right answers" given

Regression回归: Predict continuous valued output (price)

Breast cancer (malignant, benign)

Classification： Discrete valued output (0 or 1) or (0, 1, 2, 3)

one feature, two features or more features

• Tumor Size

• Age

• Clump Thickness

• Uniformity of Cell Size

• Uniformity of Cell Shape

...

Question:

You're running a company, and you want to develop learning algorithms to address each of two problems.

Problem 1: You have a large inventory of identical items. You want to predict how many of these items will sell over the next 3 months.

Problem 2: You'd like software to examine individual customer accounts, and for each account decide if it has been hacked/compromised.

Should you treat these as classification or as regression problems?

---

1-4 Unsupervised Learning

Examples:

• Google news

• Individual genes

  

• Organize computing clusters

• Social network analysis

• Market segmentation

• Astronomical data analysis

  

Cocktail party problem：

You can solve this problem by just one line program:

W,s,v = svd((repmat(sum(x.*x,1),size(x,1),1).*x)*x');

Source: Sam Roweis, Yair Weiss \& Eero Simoncelli

建议：先用Octave建立原型，再用C++、JAVA或Python语言实现

Question:

Of the following examples, which would you address using an unsupervised learning algorithm? (Check all that apply.)