首页 \ 问答 \ 关于adaboost算法(About adaboost algorithm)

关于adaboost算法(About adaboost algorithm)

我正在进行交通流预测,在那里我可以预测一个地方交通繁忙或交通繁忙。 我将每个流量分类为1-5,1是最轻的流量,5是最重的流量。

我来到这个网站http://www.waset.org/journals/waset/v25/v25-36.pdf,AdaBoost算法,我真的很难学习这个算法。 特别是在S是集合的部分(( xiyi ), i=(1,2,…,m) )。 其中Y={-1,+1} 。 什么是xy和常数LL的价值是多少?

有人能解释一下这个算法吗? :)


I'm working on a traffic flow prediction where I can predict that a place has heavy or light traffic. I have classified each traffic as 1-5, 1 being the lightest traffic and 5 being the heaviest traffic.

I came across to this website http://www.waset.org/journals/waset/v25/v25-36.pdf, AdaBoost algorithm, and I'm really having a difficulty learning this algorithm. Specially in the part where S is the set ((xi, yi), i=(1,2,…,m)). where Y={-1,+1}. What are x, y and the constant L? what is the value of L?

Can someone explain me this algorithm? :)


原文:https://stackoverflow.com/questions/11825493
更新时间:2020-11-03 09:11

最满意答案

S={(x1,y1),...,(xm,ym)} :每个(x,y)对是用于训练(或测试)分类器的样本:

  • x =描述此特定样本的特征,例如列出amount of cars on the road day of the week等的值
  • y =特定x的标签,在您的情况下可以是1, 2, 3, 4 or 5

Table 1显示了他们使用的x功能,即: DAYTIMEINTDETLINKPOSGREDISVOLOCC 。 表格的最后一列显示标签( y ),它们设置为1-1 (即, yesno )。 表中的每一行都是1个样本。

L是AdaBoost训练弱学习者的轮数(在Random Forests中用作弱分类器)。 如果你将L设置为1则AdaBoost将运行1轮,并且仅训练1个弱分类器,这将产生不良结果。 使用L不同值执行多个实验以找到最佳值(即,当AdaBoost收敛或开始过度拟合时)。


S={(x1,y1),...,(xm,ym)}: Every (x,y) pair is a sample used for training (or testing) your classifier:

  • x = The features which describe this particular sample, for example values which list the amount of cars on the road, day of the week, etc
  • y = The label for a particular x, which in your case can be 1, 2, 3, 4 or 5

Table 1 in the paper shows the x features they used , namely: DAY, TIME, INT, DET, LINK, POS, GRE, DIS, VOL and OCC. The last column of the table shows the label (y), which they set to either 1 or -1 (i.e., yes or no). Every row in the table is 1 sample.

L is the amount of rounds in which AdaBoost trains a weak learner (in the paper Random Forests is used as the weak classifier). If you set L to 1 then AdaBoost will run 1 round and only 1 weak classifier will be trained, which will have bad results. Perform multiple experiments with different values for L to find the optimal value (i.e., when AdaBoost is converged or when it starts to overfit).

2012-08-06

相关文章

更多

最新问答

更多
  • 如何保护Solr只允许SELECT请求给用户并禁止其他任何东西?(How to secure Solr to allow SELECT request to only users and disallow anything else?)
  • XPath表达式无效/错误TFHpple SWIFT 1.2(XPath Expression not working/incorrect TFHpple SWIFT 1.2)
  • css3 3D变换不能平滑地制作动画(css3 3D transform doesn't animate smoothly)
  • 运行时错误'91'和Outlook.Application = <对象变量或没有设置块变量>?(Run-time error '91' & Outlook.Application = Object variable or With block variable not set?)
  • 慢的webservice问题(Slow webservice problem)
  • textview的不正确对齐方式(Improper alignment of a textview)
  • 在第一步“Hello World”中出现Java错误(Cannot run “Hello World” program in Eclipse)
  • 为什么十六进制地址是14个字符?(why the hex address is 14 character?)
  • 如何在Python中的不同类中使用变量?(How to use variable in different classes in Python?)
  • asp:GridView HYPERLINKFIELD - datanavigateurlformatstring中的asp代码(asp:GridView HYPERLINKFIELD - asp code inside datanavigateurlformatstring)
  • 关于adaboost算法(About adaboost algorithm)
  • 在Matlab上内置图像(Built in Images on Matlab)
  • java swing:输入键事件时焦点丢失(java swing: Focus lost on enter key event)
  • C#中的通用约束,T是相同的TSomethingElse,对吧?(Generic constraints in C#, T is the same TSomethingElse, right?)
  • 从mybatis中的光标获取数据(Fetching data from cursor in mybatis)
  • 在运行时从XML构建对象的最佳方法(Best way for building objects out of XMLs at runtime)
  • 在整个窗口中拉伸sf :: Sprite(Stretch sf::Sprite across entire window)
  • Selenium Node API Web驱动程序等待超时处理程序(Selenium Node API web driver wait timeout handler)
  • 函数使用并且是map的一部分(循环依赖?)(Function uses and is part of map (circular dependency?))
  • 是否可以在C ++ 14中使用可选模板参数创建类型元组?(Is possible to make a tuple of types with optional template parameters in C++14?)
  • PHP从窗帘后面打印/ f(PHP prints /f from behind the curtains)
  • JFrame的contentPane的LayoutManager(LayoutManager of JFrame's contentPane)
  • 用于Instagram Feed的图像滚动(Image Roll Overs for Instagram Feed)
  • 如何显示拉伸字体(双倍宽度/高度)?(How to display stretched font (double width/height)?)
  • 文件操作API在fileapi.h和stdio.h中的WP8差异?(WP8 differences between file manipulation APIs in fileapi.h and stdio.h?)
  • 保存失败后Rails没有回滚事务()(Rails not rolling back transaction after failed save())
  • jqgrid中的分页问题与数组数据(Pagination problem in jqgrid with array data)
  • 重定向时,通过其他页面上的URL调用javascript函数(Call javascript function through url on otherpage while Redirecting)
  • 如何使用InvokeCommandAction调用我的方法并传入参数?(How do I go about using InvokeCommandAction to call a method of mine and pass in parameters?)
  • Jquerymobile按钮仅显示页面的第一个外观(Jquerymobile buttons are shown only first apperance of the page)