“训练异常情况检测模型”模块Train Anomaly Detection Model module

本文介绍如何在 Azure 机器学习设计器中使用“训练异常情况检测模型”模块来创建已训练的异常情况检测模型。This article describes how to use the Train Anomaly Detection Model module in Azure Machine Learning designer to create a trained anomaly detection model.

该模块将异常情况检测模型和未标记数据集的一组参数作为输入。The module takes as input a set of parameters for an anomaly detection model and an unlabeled dataset. 它将返回已训练的异常情况检测模型,以及训练数据的一组标签。It returns a trained anomaly detection model, together with a set of labels for the training data.

有关设计器中提供的异常情况检测算法的详细信息,请参阅基于 PCA 的异常情况检测For more information about the anomaly detection algorithms provided in the designer, see PCA-Based Anomaly Detection.

如何配置“训练异常情况检测模型”模块How to configure Train Anomaly Detection Model

  1. 在设计器中向管道添加“训练异常情况检测模型”模块 。Add the Train Anomaly Detection Model module to your pipeline in the designer. 可以在“异常情况检测”类别中找到此模块 。You can find this module in the Anomaly Detection category.

  2. 连接设计用于异常情况检测的模块之一,例如基于 PCA 的异常情况检测Connect one of the modules designed for anomaly detection, such as PCA-Based Anomaly Detection.

    不支持其他类型的模型。Other types of models are not supported. 运行管道时,会出现“所有模型都必须具有相同的学习器类型”错误。When you run the pipeline, you'll get the error "All models must have the same learner type."

  3. 通过选择标签列并设置特定于算法的其他参数来配置异常情况检测模块。Configure the anomaly detection module by choosing the label column and setting other parameters specific to the algorithm.

  4. 将训练数据集附加到“训练异常情况检测模型” 的右侧输入。Attach a training dataset to the right-side input of Train Anomaly Detection Model .

  5. 提交管道。Submit the pipeline.

结果Results

在训练完成后:After training is complete:

  • 若要查看模型的参数,请右键单击模块,并选择“可视化”。 To view the model's parameters, right-click the module and select Visualize .

  • 若要创建预测,请对新输入数据使用评分模型模块。To create predictions, use the Score Model module with new input data.

  • 若要保存已训练模型的快照,请选择该模块。To save a snapshot of the trained model, select the module. 然后,在右侧面板中选择“输出 + 日志” 选项卡下的“注册数据集” 图标。Then select the Register dataset icon under the Outputs+logs tab in the right panel.

后续步骤Next steps

请参阅 Azure 机器学习的可用模块集See the set of modules available to Azure Machine Learning.

有关特定于设计器模块的错误列表,请参阅设计器(预览版)的异常和错误代码See Exceptions and error codes for the designer (preview) for a list of errors specific to the designer modules. ''