我们可以指定在 Vowpal Wabbit 中使用哪种算法(例如决策树、支持向量机、集成、神经网络)吗?或者,Automl select 是算法本身吗?
Can we specify which algorithm to use (e.g., decision tree, SVM, ensemble, NNs) in Vowpal Wabbit? Or, does Automl select the algorithm itself?
我正在尝试阅读 Vowpal Wabbit 的 documentation,它没有指定如何 select 特定的学习算法(不是损失),如 SVM、NN、决策树等。如何select 是一种特定的学习算法吗?
还是 select 算法本身取决于问题类型(regression/classification 像 automl 类型或低代码 ML 库?
有一些博客显示通过 -nn
命令使用神经网络,但这不是文档的一部分——这是因为它不关注特定算法,如上所述吗?如果是这样,Vowpal Wabbit 本质上是什么?
Vowpal Wabbit基于在线学习(SGD-like更新,但如果确实需要批量优化也有--bfgs
)和(机器学习)reductions. See some of the tutorials or papers to understand the idea of reductions. Many VW papers are also about Contextual Bandit, which is implemented as a reduction to a cost-sensitive one-against-all (OAA) classification (which is further reduced to regression). See a simple intro into reductions or a simple example how binary classification is reduced into regression.
据我所知,VowpalWabbit 不支持决策树或集成,但请参阅 --boosting
and --bootstrap
. It does not support SVM, but see --loss_function hinge
(hinge loss is one of the two key concepts of SVM) and --ksvm
. It does not support NN, but --nn
(and related options) 提供非常有限的模拟单个隐藏层的支持(feed-forward 具有 tanh 激活函数),可以添加到缩减堆栈中。
我正在尝试阅读 Vowpal Wabbit 的 documentation,它没有指定如何 select 特定的学习算法(不是损失),如 SVM、NN、决策树等。如何select 是一种特定的学习算法吗?
还是 select 算法本身取决于问题类型(regression/classification 像 automl 类型或低代码 ML 库?
有一些博客显示通过 -nn
命令使用神经网络,但这不是文档的一部分——这是因为它不关注特定算法,如上所述吗?如果是这样,Vowpal Wabbit 本质上是什么?
Vowpal Wabbit基于在线学习(SGD-like更新,但如果确实需要批量优化也有--bfgs
)和(机器学习)reductions. See some of the tutorials or papers to understand the idea of reductions. Many VW papers are also about Contextual Bandit, which is implemented as a reduction to a cost-sensitive one-against-all (OAA) classification (which is further reduced to regression). See a simple intro into reductions or a simple example how binary classification is reduced into regression.
据我所知,VowpalWabbit 不支持决策树或集成,但请参阅 --boosting
and --bootstrap
. It does not support SVM, but see --loss_function hinge
(hinge loss is one of the two key concepts of SVM) and --ksvm
. It does not support NN, but --nn
(and related options) 提供非常有限的模拟单个隐藏层的支持(feed-forward 具有 tanh 激活函数),可以添加到缩减堆栈中。