1. 什么是机器学习?
google搜索、排序;苹果photo管理功能;邮件过滤等都是机器学习,机器学习是指不根据明确的程序来运行的程序,机器通过学习来获得技能,目前主要是通过神经网络来完成。这是计算机的一种新技能。
Examples:
-Database mining 从数据中发现新的知识
Large datasets from growth of automation/web.
E.g. , Web click data, medical records, biology, engeering
-Applications can't program by hand
E.g. , Autonomous helicopter, handwriting recognition, most of Natural Language Processing(NLP),Computer Vision.
-Self-customizing programs
- Understanding human learning
Definition
-Field of study that gives computers the ability to learn without being explicitly programmed. by Arthur Samuel(1959)
- 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.
E--T---->P---T---E better.
Machine learning algorithms: important-how to use these tools--找到最佳实践
- Supervised learning.
回归问题,指预测一个连续值的输出.
分类问题——离散值的输出。
- Unsupervised learning
聚类问题---例如新闻分组,从无序中发现有序
- Ohters: Reinforcement leanring, recommender systems.
使用正确的octave,很快的实现算法,好的编程可以很好的节省时间。
2. 几种算法
- 线性回归 Y=Ax+e
Training Set-->learning algorithm--->h(hypothesis-假设)
如何得到参数值? 最小二乘法——cost function-->目标函数 goal fuction
梯度下降算法:outline: start with some A ,keep changing A to reduce goal function. 学习效率(局部最优,不收敛问题--正反馈学习效率 );梯度更新。