【Udacity笔记】What is Machine Learning?

Teaching computers to learn to perform tasks from past experiences(recorded data)

一、Decision Tree(决策树)

——Example:for recommend app

二、Naive Bayes Algorithm(朴素贝叶斯)

——Example:for detecting Spam e-mails(垃圾邮件)

三、Gradient descent(梯度下降)

——Example:Minimize the Error

四、Linear Regression(线性回归)

——Example:Price of a house

五、(对数几率回归)

Logistic Regression(逻辑回归)
Log loss function (对数损失函数)——代表错误数目

——Example:找出能最小化误差函数的最佳拟合线

【Udacity笔记】What is Machine Learning?

最小化误差函数——结合梯度下降算法
爬山——每一步都走梯度下降最快的(误差减小更多)

六、Support Vector Machine(支持向量机SVM)

  • Cutting data with style
  • Support——靠近边界的点被称为支撑(support)

【Udacity笔记】What is Machine Learning?

【Udacity笔记】What is Machine Learning?

七、Kernel Trick(核函数)

  • when a line is not enough
    【Udacity笔记】What is Machine Learning?
    【Udacity笔记】What is Machine Learning?
    【Udacity笔记】What is Machine Learning?

八、Neural Network(神经网络)

【Udacity笔记】What is Machine Learning?

九、类比忍者

【Udacity笔记】What is Machine Learning?
【Udacity笔记】What is Machine Learning?

十、K均值聚类(K-Means Clustering)

  • Example pizza parlors
  • 需要预知最后分成几组

【Udacity笔记】What is Machine Learning?

十一、层次聚类(Hierarchical Clustering)

  • 无需预知最后聚类数目

【Udacity笔记】What is Machine Learning?

阶段总结

【Udacity笔记】What is Machine Learning?