【学习笔记】吴恩达机器学习 | 第一章 | 机器学习绪论

简要声明

---

1. 课程学习相关网址
   1. Bilibili
   2. 网易云课堂
   3. 学习讲义
2. 由于课程学习内容为英文，文本会采用英文进行内容记录，采用中文进行简要解释。
3. 本学习笔记单纯是为了能对学到的内容有更深入的理解，如果有错误的地方，恳请包容和指正。
4. 非常感谢Andrew Ng吴恩达教授的无私奉献！！！

专有名词

---

Machine Learning	机器学习	Supervised learning	监督学习
Unsupervised learning	非监督学习	Reinforcement learning	强化学习
Regression	回归问题	Classification	分类问题
label	标签	feature	特征
cluster	簇	clustering algorithm	聚类算法

Welcome to machine learning

---

Machine Learning

1. Grew out of work in AI → 使机器自己学习如何去做有趣的事情
2. New capability for computers

Examples

1. Database mining 数据挖掘 (E.g., Web click data, medical records, biology, engineering)
2. Applications can’t program by hand 无法手动编写程序 (E.g., Autonomous helicopter, handwriting recognition, Natural Language Processing (NLP), Computer Vision.)
3. Self-customizing programs 私人订制程序 (E.g., Amazon, Netflix product recommendations)
4. Understanding human learning 理解人类的学习过程

What is machine learning

---

Machine Learning definition

1. Arthur Samuel (1959). Machine Learning: “Field of study that gives computers the ability to learn without being explicitly programmed.”
   在没有明确设置的情况下，使计算机具有学习能力的研究领域。
2. Tom Mitchell (1998) Well-posed Learning Problem: “A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P, improves with experience E.”
   字幕翻译：计算机程序从经验E中学习解决某一任务T进行某一性能度量P，通过P测定在T上的表现因经验E而提高。
   DeepL：如果一个计算机程序在某个任务T和某个性能指标P方面的性能随着经验E的增加而提高，那么它就被称为从经验E中学习。

Machine learning algorithms

1. Supervised learning & Unsupervised learning
2. Reinforcement learning, recommender systems.

Supervised Learning

---

Housing price prediction

1. 已知一个数据集，已知房子的平方英尺数，预测房子的价格。可以通过一条直线拟合数据，或者通过二次函数拟合数据。如何选择和如何决定用什么拟合数据会有更好的结果。
2. Supervised Learning“right answers” given →数据集包含正确答案，算法基于正确答案预测更多的正确答案
3. Regression: Predict continuous valued output (price) →回归问题：预测连续值输出

Breast cancer (malignant, benign)

1. 已知一个数据集，已知肿瘤的大小，预测肿瘤是良性还是恶性的概率。
2. Classification: Discrete valued output (0 or 1) →分类问题：预测一个离散值输出
3. 机器学习会有很多feature特征

Unsupervised Learning

---

Unsupervised Learning

1. 非监督学习没有任何labels标签
2. 非监督学习会把数据分成不同的clusters簇
3. 非监督学习是clustering algorithm聚类算法

Applications

1. Organize computing clusters 组织大型计算机集群高效协作
2. Social network analysis 社交网络分析
3. Market segmentation 市场分割
4. Segmentation data analysis 天文数据分析

吴恩达教授语录

---

• “Machine learning is one of the most exciting recent technologies.”
• “I like to make an analogy to learning to become a carpenter.”
• “I hope to make you one of the best people in knowing how to design and build serious machine learning and AI systems.”