什么是机器学习工作室(经典版)?What is Machine Learning Studio (classic)?

提示

鼓励当前正在使用或评估机器学习工作室(经典版)的客户尝试 Azure 机器学习设计器(预览版),它提供拖放 ML 模块以及 可伸缩性、版本控制和企业安全性。Customers currently using or evaluating Machine Learning Studio (classic) are encouraged to try Azure Machine Learning designer (preview), which provides drag-n-drop ML modules plus scalability, version control, and enterprise security.

Microsoft Azure 机器学习工作室(经典版)是一个协作型拖放式工具,可用于根据数据构建、测试和部署预测分析解决方案。Microsoft Azure Machine Learning Studio (classic) is a collaborative, drag-and-drop tool you can use to build, test, and deploy predictive analytics solutions on your data. Azure 机器学习工作室(经典)将模型发布为可让自定义应用或 BI 工具(如 Excel)方便使用的 Web 服务。Azure Machine Learning Studio (classic) publishes models as web services that can easily be consumed by custom apps or BI tools such as Excel.

机器学习工作室(经典版)中融合了数据科学、预测分析、云资源和数据。Machine Learning Studio (classic) is where data science, predictive analytics, cloud resources, and your data meet.

机器学习工作室(经典版)交互式工作区The Machine Learning Studio (classic) interactive workspace

若要开发预测分析模型,通常使用一个或多个源中的数据,然后通过各种数据操作和统计函数对该数据进行转换和分析,生成一组结果。To develop a predictive analysis model, you typically use data from one or more sources, transform, and analyze that data through various data manipulation and statistical functions, and generate a set of results. 开发此类模型是一个迭代过程。Developing a model like this is an iterative process. 在修改各种函数及其参数时,结果会不断趋于一致,直到已训练的有效模型令人满意。As you modify the various functions and their parameters, your results converge until you are satisfied that you have a trained, effective model.

Azure 机器学习工作室(经典)提供交互式的可视工作区,可在其中轻松构建、测试和迭代预测分析模型。Azure Machine Learning Studio (classic) gives you an interactive, visual workspace to easily build, test, and iterate on a predictive analysis model. 可以将数据集和分析模块拖放到交互式画布,将它们连接在一起构成试验,然后在机器学习工作室(经典版)中运行。You drag-and-drop datasets and analysis modules onto an interactive canvas, connecting them together to form an experiment, which you run in Machine Learning Studio (classic). 若要在模型设计上迭代,则需要编辑试验,可根据需要保存一个副本,并重新运行该试验。To iterate on your model design, you edit the experiment, save a copy if desired, and run it again. 准备就绪后,可以将 训练实验 转换为 预测试验,然后将其发布为 Web 服务”发布,以便其他人可以访问模型。When you're ready, you can convert your training experiment to a predictive experiment, and then publish it as a web service so that your model can be accessed by others.

不需要编程,只需以可视方式连接数据集和模块即可构建预测分析模型。There is no programming required, visually connect datasets and modules to construct your predictive analysis model.

Azure 机器学习工作室(经典)示意图:创建试验、读取多个源的数据、编写评分的数据、编写模型。

机器学习工作室(经典版)与 Azure 机器学习有什么区别?How does Machine Learning Studio (classic) differ from Azure Machine Learning?

Azure 机器学习提供 SDK 和 Azure 机器学习设计器(预览版),可以快速准备数据,以及对机器学习模型进行训练和部署 。Azure Machine Learning provides both SDKs -and- the Azure Machine Learning designer (preview), to quickly prep data, train and deploy machine learning models. 此设计器提供与工作室(经典版)类似的拖放体验。The designer provides a similar drag-and-drop experience to Studio (classic). 但是,不像工作室(经典版)的专用计算平台,此设计器使用你自己的计算资源,并且已完全集成到 Azure 机器学习中。However, unlike the proprietary compute platform of Studio (classic), the designer uses your own compute resources and is fully integrated into Azure Machine Learning.

下面是一个快速比较:Here is a quick comparison:

机器学习工作室(经典版)Machine Learning Studio (classic) Azure 机器学习Azure Machine Learning
拖放界面Drag and drop interface Yes 是 - Azure 机器学习设计器(预览版)Yes - Azure Machine Learning designer (preview)
试验Experiment 可缩放(10 GB 训练数据限制)Scalable (10-GB training data limit) 使用计算目标进行缩放Scale with compute target
拖放界面的模块Modules for drag-and-drop interface 很多Many 常用模块的初始集Initial set of popular modules
训练计算目标Training compute targets 专用计算目标,仅限 CPU 支持Proprietary compute target, CPU support only 支持 Azure 机器学习计算(GPU 或 CPU)和笔记本 VM。Supports Azure Machine Learning compute (GPU or CPU) and Notebook VMs.
SDK 中支持的其他计算(Other computes supported in SDK)
推断计算目标Inferencing compute targets 专用 Web 服务格式,不可自定义Proprietary web service format, not customizable Azure Kubernetes 服务和 AML 计算Azure Kubernetes Service and AML Compute
SDK 中支持的其他计算(Other computes supported in SDK)
ML 管道ML Pipeline 不支持Not supported 支持管道Pipelines supported
MLOpsMLOps 基本模型管理和部署Basic model management and deployment 可配置的部署 - 模型和管道版本控制和跟踪Configurable deployment - model and pipeline versioning and tracking
模型格式Model format 专用格式,仅限工作室(经典)Proprietary format, Studio (classic) only 标准格式取决于训练作业类型Standard format depending on training job type
自动化模型训练和超参数优化Automated model training and hyperparameter tuning No 尚未在设计器中Not yet in the designer
在 SDK 和工作区登陆页中受支持(Supported in the SDK and workspace landing page)

尝试使用教程:使用设计器预测汽车价格Try out the designer with Tutorial: Predict automobile price with the designer

备注

在工作室(经典版)中创建的模型不能通过 Azure 机器学习来部署或管理。Models created in Studio (classic) can't be deployed or managed by Azure Machine Learning. 但是,在设计器中创建和部署的模型可以通过 Azure 机器学习工作区进行管理。However, models created and deployed in the designer can be managed through the Azure Machine Learning workspace.

下载机器学习工作室(经典版)概述示意图Download the Machine Learning Studio (classic) overview diagram

下载“Microsoft Azure 机器学习工作室(经典版)功能概述”示意图,并获取机器学习工作室(经典版)功能的高级视图 。Download the Microsoft Azure Machine Learning Studio (classic) Capabilities Overview diagram and get a high-level view of the capabilities of Machine Learning Studio (classic). 若要随时随地查看,可以打印卡片大小(11 x 17 英寸)的示意图。To keep it nearby, you can print the diagram in tabloid size (11 x 17 in.).

此处下载关系图:Microsoft Azure 机器学习工作室(经典版)功能概述 Microsoft Azure Machine Learning Studio (classic) Capabilities OverviewDownload the diagram here: Microsoft Azure Machine Learning Studio (classic) Capabilities Overview Microsoft Azure Machine Learning Studio (classic) Capabilities Overview

工作室(经典版)试验的组成部分Components of a Studio (classic) experiment

试验由数据集组成,数据集将数据提供给分析模块,将这些模块连接起来即可构成预测分析模型。An experiment consists of datasets that provide data to analytical modules, which you connect together to construct a predictive analysis model. 具体而言,有效的试验有以下特征:Specifically, a valid experiment has these characteristics:

  • 试验至少包含一个数据集和一个模块The experiment has at least one dataset and one module
  • 数据集只能连接到模块Datasets may be connected only to modules
  • 模块可以连接到数据集或其他模块Modules may be connected to either datasets or other modules
  • 模块的所有输入端口必须与数据流建立某种连接All input ports for modules must have some connection to the data flow
  • 必须设置每个模块的所有必需参数All required parameters for each module must be set

可以从头开始创建试验,或者使用现有的示例试验作为模板。You can create an experiment from scratch, or you can use an existing sample experiment as a template. 有关详细信息,请参阅复制示例试验以创建新的机器学习试验For more information, see Copy example experiments to create new machine learning experiments.

有关创建简单试验的示例,请参阅 Create a simple experiment in Azure Machine Learning Studio(在 Azure 机器学习工作室(经典版)中创建简单试验)。For an example of creating an experiment, see Create a simple experiment in Azure Machine Learning Studio (classic).

有关创建预测分析解决方案的更完整演练,请参阅 Develop a predictive solution with Azure Machine Learning Studio(使用 Azure 机器学习工作室(经典版)开发预测解决方案)。For a more complete walkthrough of creating a predictive analytics solution, see Develop a predictive solution with Azure Machine Learning Studio (classic).

数据集Datasets

数据集是指已上传到机器学习工作室(经典版),可在建模过程中使用的数据。A dataset is data that has been uploaded to Machine Learning Studio (classic) so that it can be used in the modeling process. 机器学习工作室(经典版)提供了许多示例数据集供试验,你可根据需要上传更多的数据集。A number of sample datasets are included with Machine Learning Studio (classic) for you to experiment with, and you can upload more datasets as you need them. 下面是随附数据集的一些例子:Here are some examples of included datasets:

  • 各种汽车的 MPG 数据 - 汽车的每加仑燃油英里数 (MPG) 值,按缸数、马力等参数列出。MPG data for various automobiles - Miles per gallon (MPG) values for automobiles identified by number of cylinders, horsepower, etc.
  • 乳腺症数据 - 乳腺癌诊断数据。Breast cancer data - Breast cancer diagnosis data.
  • 森林火灾数据 - 葡萄牙东北部森林火灾的规模。Forest fires data - Forest fire sizes in northeast Portugal.

构建试验时,可以从画布左侧提供的数据集列表中进行选择。As you build an experiment, you can choose from the list of datasets available to the left of the canvas.

有关机器学习工作室(经典版)随附的示例数据集列表,请参阅 Use the sample data sets in Azure Machine Learning Studio(使用 Azure 机器学习工作室(经典版)中的示例数据集)。For a list of sample datasets included in Machine Learning Studio (classic), see Use the sample data sets in Azure Machine Learning Studio (classic).

模块Modules

模块是可对数据执行的算法。A module is an algorithm that you can perform on your data. Azure 机器学习工作室(经典)有许多模块,包括数据引入函数、训练、评分和验证过程。Azure Machine Learning Studio (classic) has a number of modules ranging from data ingress functions to training, scoring, and validation processes. 下面是随附模块的一些例子:Here are some examples of included modules:

构建试验时,可以从画布左侧提供的模块列表中选择。As you build an experiment, you can choose from the list of modules available to the left of the canvas.

模块可能提供一组参数用于配置模块的内部算法。A module may have a set of parameters that you can use to configure the module's internal algorithms. 在画布上选择模块时,模块的参数会显示在画布右侧的“属性” 窗格中。When you select a module on the canvas, the module's parameters are displayed in the Properties pane to the right of the canvas. 可以在该窗格中修改参数来调整模型。You can modify the parameters in that pane to tune your model.

在浏览可用的机器学习算法大型库时如需帮助,请参阅 How to choose algorithms for Microsoft Azure Machine Learning Studio(如何选择 Microsoft Azure 机器学习工作室(经典版)的算法)。For some help navigating through the large library of machine learning algorithms available, see How to choose algorithms for Microsoft Azure Machine Learning Studio (classic).

部署预测分析 Web 服务Deploying a predictive analytics web service

准备好预测分析模型后,可以从机器学习工作室(经典版)将它部署为 Web 服务。Once your predictive analytics model is ready, you can deploy it as a web service right from Machine Learning Studio (classic). 有关此过程的信息,请参阅 Deploy an Azure Machine Learning web service(部署 Azure 机器学习 Web 服务)。For more information on this process, see Deploy an Azure Machine Learning web service.

后续步骤Next steps

可以使用分步快速入门基于样本的构建了解预测分析和机器学习的基础知识。You can learn the basics of predictive analytics and machine learning using a step-by-step quickstart and by building on samples.