前言
“机器学习基石”是 Coursera 上一门关于机器学习的课程,由国立台湾大学的老师林轩田讲授。该课程一共有16节课,主要介绍了机器学习领域的基础理论知识。
授课大纲
课程的大纲如下,以下的每个小项目对应到约一小时的课程:
When Can Machines Learn? [何时可以使用机器学习]
– The Learning Problem [机器学习问题]
– Learning to Answer Yes/No [二元分类]
– Types of Learning [各式机器学习问题]
– Feasibility of Learning [机器学习的可行性]
Why Can Machines Learn? [为什么机器可以学习]
– Training versus Testing [训练与测试]
– Theory of Generalization [举一反三的一般化理论]
– The VC Dimension [VC 维度]
– Noise and Error [噪声时错误]
How Can Machines Learn? [机器可以怎么样学习]
– Linear Regression [线性回归]
– Linear Soft Classification [软性的线性分类]
– Linear Classification beyond Yes/No [二元分类以外的分类问题]
– Nonlinear Transformation [非线性转换]
How Can Machines Learn Better? [机器可以怎么样学得更好]
– Hazard of Overfitting [过度训练的危险]
– Preventing Overfitting I: Regularization [避免过度训练一:控制调适]
–