机器学习基石第一节

1、What is machine learning

  • define
    observation -》learning -》 skill

    data -》ML -》 skill

    skill = improve some performance measure

    ML = improve some performance measure with experience computed from data
    eg. stock data -》 ML -》 more investment gain

    data -》ML -》 improve performance measure

  • some use scenarios
    when human cannot program the system manually
    when human cannot ‘define the solution‘ easily
    when needing rapid decisions that humans cannot do
    when needing to be user-oriented in a massive scale

  • Key Essence of Machine Learning
    exists some ’underlying(mean潜在的) pattern‘ to be learned、
    but no programmable(easy)definition
    somehow there is data about the pattern

  • application of machine learning
    Twitter 分析餐馆好不好
    推荐电影 需要知晓用户对于一些电影的评价

  • components problem
    f: x -> y, x,y 为data的输入和输出,而机器学习是在模拟f,从数据集D(x,y),通过演算法A,算出一个假说g(hypothesis)
    {(xn,yn)} from f -> ML -> g
    D(training examples) -> A(learning program) -> g 约等于 f (final hypothesis)

    假说的集合 H (hypothesis set),g 属于 H
    learning model = A and H

    D —A on H—>(g:x->y)

  • Machine Learning and Data Mining
    机器学习和数据挖掘是密不可分的
    机器学习是实现人工智能的一种方式
    统计是实现机器学习的一种方法
    provable 可证明的 assumptions 假设,假定

你可能感兴趣的:(机器学习基石第一节)