Amazon Forecast 不再提供給新客戶。Amazon Forecast 的現有客戶可以繼續正常使用服務。[進一步了解」](https://aws.amazon.com/blogs/machine-learning/transition-your-amazon-forecast-usage-to-amazon-sagemaker-canvas/) 本文為英文版的機器翻譯版本，如內容有任何歧義或不一致之處，概以英文版為準。 # 使用 自動化 CloudFormation 在本教學課程中，您會使用 AWS CloudFormation 自動化堆疊來啟動 Amazon Forecast 管道，並使用示範資料集產生預測。 AWS 預測 CloudFormation 堆疊： + [使用Machine Learning解決方案範本部署改善預測準確性](https://docs.aws.amazon.com/solutions/latest/improving-forecast-accuracy-with-machine-learning/automated-deployment.html) CloudFormation 。 + 將 [NYC Taxi 資料集](https://registry.opendata.aws/nyc-tlc-trip-records-pds/)部署至預測資料 Amazon S3 儲存貯體。 + 在預測中自動啟動示範 NYC 計程車預測管道。 CloudFormation 範本會預先載入目標時間序列、相關時間序列和項目中繼資料示範資料集。主控台中的相關欄位會預先填入其各自的 S3 位置。 使用示範資料集完成本教學課程後，您可以使用相同的自動化堆疊，使用您自己的資料集產生預測。 下圖顯示本教學課程中所使用的元件。 ![\[AWS data pipeline for Amazon Forecast, showing data preparation, ingestion, forecasting, and evaluation stages.\]](http://docs.aws.amazon.com/zh_tw/forecast/latest/dg/images/cloudformationautomation-architecture.png) ## 先決條件 開始教學課程之前，請確定您已登入 AWS 帳戶 並安裝 CloudFormation 範本： 1. 登入您的 AWS 帳戶。如果您還沒有 ，請[建立 AWS 帳戶](https://aws.amazon.com/premiumsupport/knowledge-center/create-and-activate-aws-account/)。 1. 安裝 AWS CloudFormation 範本。選擇離您最近的區域： + 東京：[ap-northeast-1](https://console.aws.amazon.com/cloudformation/home?region=ap-northeast-1#/stacks/new?stackName=forecast-stack&templateURL=https:%2F%2Fs3.amazonaws.com%2Fsolutions-reference%2Fimproving-forecast-accuracy-with-machine-learning%2Flatest%2Fimproving-forecast-accuracy-with-machine-learning-demo.template) + 首爾：[ap-northeast-2](https://console.aws.amazon.com/cloudformation/home?region=ap-northeast-2#/stacks/new?stackName=forecast-stack&templateURL=https:%2F%2Fs3.amazonaws.com%2Fsolutions-reference%2Fimproving-forecast-accuracy-with-machine-learning%2Flatest%2Fimproving-forecast-accuracy-with-machine-learning-demo.template) + 孟買：[ap-south-1](https://console.aws.amazon.com/cloudformation/home?region=ap-south-1#/stacks/new?stackName=forecast-stack&templateURL=https:%2F%2Fs3.amazonaws.com%2Fsolutions-reference%2Fimproving-forecast-accuracy-with-machine-learning%2Flatest%2Fimproving-forecast-accuracy-with-machine-learning-demo.template) + 新加坡：[ap-southeast-1](https://console.aws.amazon.com/cloudformation/home?region=ap-southeast-1#/stacks/new?stackName=forecast-stack&templateURL=https:%2F%2Fs3.amazonaws.com%2Fsolutions-reference%2Fimproving-forecast-accuracy-with-machine-learning%2Flatest%2Fimproving-forecast-accuracy-with-machine-learning-demo.template) + 雪梨：[ap-southeast-2](https://console.aws.amazon.com/cloudformation/home?region=ap-southeast-2#/stacks/new?stackName=forecast-stack&templateURL=https:%2F%2Fs3.amazonaws.com%2Fsolutions-reference%2Fimproving-forecast-accuracy-with-machine-learning%2Flatest%2Fimproving-forecast-accuracy-with-machine-learning-demo.template) + 法蘭克福：[eu-cental-1](https://console.aws.amazon.com/cloudformation/home?region=eu-central-1#/stacks/new?stackName=forecast-stack&templateURL=https:%2F%2Fs3.amazonaws.com%2Fsolutions-reference%2Fimproving-forecast-accuracy-with-machine-learning%2Flatest%2Fimproving-forecast-accuracy-with-machine-learning-demo.template) + 愛爾蘭：[eu-west-1](https://console.aws.amazon.com/cloudformation/home?region=eu-west-1#/stacks/new?stackName=forecast-stack&templateURL=https:%2F%2Fs3.amazonaws.com%2Fsolutions-reference%2Fimproving-forecast-accuracy-with-machine-learning%2Flatest%2Fimproving-forecast-accuracy-with-machine-learning-demo.template) + 維吉尼亞北部：[us-east-1](https://console.aws.amazon.com/cloudformation/home?region=us-east-1#/stacks/new?stackName=forecast-stack&templateURL=https:%2F%2Fs3.amazonaws.com%2Fsolutions-reference%2Fimproving-forecast-accuracy-with-machine-learning%2Flatest%2Fimproving-forecast-accuracy-with-machine-learning-demo.template) + 俄亥俄州：[us-east-2](https://console.aws.amazon.com/cloudformation/home?region=us-east-2#/stacks/new?stackName=forecast-stack&templateURL=https:%2F%2Fs3.amazonaws.com%2Fsolutions-reference%2Fimproving-forecast-accuracy-with-machine-learning%2Flatest%2Fimproving-forecast-accuracy-with-machine-learning-demo.template) + 奧勒岡：[us-west-2](https://console.aws.amazon.com/cloudformation/home?region=us-west-2#/stacks/new?stackName=forecast-stack&templateURL=https:%2F%2Fs3.amazonaws.com%2Fsolutions-reference%2Fimproving-forecast-accuracy-with-machine-learning%2Flatest%2Fimproving-forecast-accuracy-with-machine-learning-demo.template) 這會使用 [NYC Taxi 資料集](https://registry.opendata.aws/nyc-tlc-trip-records-pds/)部署示範堆疊。 ## 部署用於預測自動化的 CloudFormation 範本 使用 NYC Taxi 資料集部署 CloudFormation 範本 **步驟 1**：接受預設值，然後選擇**下一步**。 ![\[Create stack interface showing template options and Amazon S3 URL input field.\]](http://docs.aws.amazon.com/zh_tw/forecast/latest/dg/images/cloudformationautomation-step1.png) **步驟 2**：提供通知的電子郵件地址，然後選擇**下一步**。 ![\[Datasets configuration form with URL fields for time series data and email input for forecast results.\]](http://docs.aws.amazon.com/zh_tw/forecast/latest/dg/images/cloudformationautomation-step2.png) **步驟 3**：接受預設值，然後選擇**下一步**。 **步驟 4**：針對功能，選取兩個核取方塊， CloudFormation 以允許 建立 AWS Identity and Access Management (IAM) 資源和巢狀堆疊。選擇**建立堆疊**。 ![\[Capabilities section with checkboxes for IAM resources and CloudFormation capability acknowledgments.\]](http://docs.aws.amazon.com/zh_tw/forecast/latest/dg/images/cloudformationautomation-step4.png) 您已在預測中部署 CloudFormation 範本。 ## 清除 部署此 CloudFormation 範本之後，您可以清除新建立的資源、使用自己的資料集部署 CloudFormation 堆疊，以及探索其他部署選項。 + **清除**：刪除示範堆疊會保留「使用Machine Learning改善預測準確性」堆疊。刪除「使用Machine Learning改善預測準確性」堆疊會保留所有 S3、Athena、QuickSight 和預測資料。 + **使用您自己的資料集**：若要使用您自己的時間序列資料部署此 CloudFormation 範本，請在**步驟 2** 的資料集組態區段中輸入資料集的 S3 位置。 + **其他部署選項**：如需更多部署選項，請參閱[自動部署](https://docs.aws.amazon.com/solutions/latest/improving-forecast-accuracy-with-machine-learning/automated-deployment.html)。如果資料已可用，您可以在沒有示範資料的情況下部署堆疊。