吴恩达machine learning 学习笔记

机器学习介绍

机器学习在当今应用非常广泛，凡是看上去比较智能的地方都有可能是机器学习应用的地方，比如声音转文本、图像检索等等。

Definition: Field of study that gives computers the ability to learn without being explicitly programmed. -----Arthur Samuel(1959)

Machine learning algorithms types:

• Supervised learning
• Unsupervised learning
• Reinforce learning

Supervised learning 介绍

Given input (dataset with label)，learning output；learning x to y mappings;

two major type of supervised learning algorithm

1. regression（回归）predict a number of infinitely many possible numbers
2. Classification 分类 predict categories -- small number of possible outputs

Unsupervised learning 介绍

Find something interseting in unlabeled data. e.g. the common structure or pattern in all data

some types of unspervised learning algorithm

1. Clustering algorithm 聚类算法
   1. which is a type of unsupervised learning algorithm, takes data without labels and tries to automatically group them into clusters. (Group similar data points together)
2. Anomaly detection 异常检测
   1. which is used to detect unusual events/points
3. Dimensionality reduction 降维
   1. this lets you take a big dataset and almost magically compress it to a much smaller dataset while losing as little information as possible.

线性回归 linear regression model

Terminology

training set: data used to train the model

Notation

\(x\) = "input" variable/feature

\(y\) = "output"/"target" variable/feature

\((x,y)\) = single training example

\(m\) = number of training examples