次の方法で共有


AnomalyDetectorClient class

Azure Anomaly Detector サービスと対話するためのクライアント クラス。

Extends

コンストラクター

AnomalyDetectorClient(string, TokenCredential | KeyCredential, PipelineOptions)

AnomalyDetectorClient のインスタンスを作成します。

使用例:

import { AnomalyDetectorClient, AzureKeyCredential } from "@azure/ai-anomaly-detector";

const client = new AnomalyDetectorClient(
   "<service endpoint>",
   new AzureKeyCredential("<api key>")
);

継承されたプロパティ

apiVersion
endpoint

継承されたメソッド

deleteMultivariateModel(string, AnomalyDetectorDeleteMultivariateModelOptionalParams)

modelId に従って、既存の多変量モデルを削除します

detectAnomaly(string, DetectionRequest, AnomalyDetectorDetectAnomalyOptionalParams)

多変量異常検出タスクを modelId のトレーニング済みのモデルと共に送信します。入力スキーマは、トレーニング要求と同じである必要があります。 そのため、要求は非同期的に完了し、検出の結果をクエリするための resultId を返します。この要求は、外部からアクセス可能な Azure ストレージの URI (可能であれば Shared Access Signature URI) を示すソース リンクである必要があります。 モデルの生成で使用されたすべての時系列を 1 つのファイルに zip 圧縮する必要があります。 各時系列は、最初の列がタイムスタンプで、2 番目の列が値になります。

detectChangePoint(DetectChangePointRequest, AnomalyDetectorDetectChangePointOptionalParams)

すべての系列ポイントの変更ポイント スコアを評価する

detectEntireSeries(DetectRequest, AnomalyDetectorDetectEntireSeriesOptionalParams)

この操作により、系列全体を含むモデルが生成され、各ポイントは同じモデルで検出されます。 このメソッドでは、特定のポイントの前後のポイントを使用して、異常であるかどうかを判断します。 検出全体を使用すると、時系列の全体的な状態をユーザーに提供できます。

detectLastPoint(DetectRequest, AnomalyDetectorDetectLastPointOptionalParams)

この操作では、最新のポイント以前のポイントを使用してモデルを生成します。 このメソッドでは、履歴ポイントのみを使用して、ターゲット ポイントが異常であるかどうかを判断します。 最新のポイント検出操作は、ビジネス メトリックのリアルタイム監視のシナリオと一致します。

exportModel(string, AnomalyDetectorExportModelOptionalParams)

modelId に基づいて、多変量異常検出モデルをエクスポートします

getDetectionResult(string, AnomalyDetectorGetDetectionResultOptionalParams)

DetectAnomalyAsync API によって返された resultId に基づいて、多変量異常検出の結果を取得します

getMultivariateModel(string, AnomalyDetectorGetMultivariateModelOptionalParams)

多変量モデルの詳細情報 (トレーニングの状態やモデルで使用された変数を含む) を取得します。

lastDetectAnomaly(string, LastDetectionRequest, AnomalyDetectorLastDetectAnomalyOptionalParams)

異常検出のための同期された API。

listMultivariateModel(AnomalyDetectorListMultivariateModelOptionalParams)

サブスクリプションのモデルを一覧表示します

sendOperationRequest(OperationArguments, OperationSpec, ServiceCallback<any>)

指定された OperationSpec を使用して設定された HTTP 要求を送信します。

sendRequest(RequestPrepareOptions | WebResourceLike)

指定された httpRequest を送信します。

trainMultivariateModel(AnomalyDetectorClientModelInfo, AnomalyDetectorTrainMultivariateModelOptionalParams)

多変量異常検出モデルを作成してトレーニングします。 この要求には、外部からアクセス可能な Azure ストレージの URI (可能であれば Shared Access Signature URI) を示すソース パラメーターを含める必要があります。 モデルの生成で使用されたすべての時系列を 1 つのファイルに zip 圧縮する必要があります。 各時系列は、最初の列がタイムスタンプで、2 番目の列が値である 1 つの CSV ファイルに含まれます。

コンストラクターの詳細

AnomalyDetectorClient(string, TokenCredential | KeyCredential, PipelineOptions)

AnomalyDetectorClient のインスタンスを作成します。

使用例:

import { AnomalyDetectorClient, AzureKeyCredential } from "@azure/ai-anomaly-detector";

const client = new AnomalyDetectorClient(
   "<service endpoint>",
   new AzureKeyCredential("<api key>")
);
new AnomalyDetectorClient(endpointUrl: string, credential: TokenCredential | KeyCredential, options?: PipelineOptions)

パラメーター

endpointUrl

string

Azure Anomaly Detector サービス エンドポイントの URL

credential

TokenCredential | KeyCredential

サービスへの要求を認証するために使用されます。

options
PipelineOptions

Form Recognizer クライアントを構成するために使用されます。

継承されたプロパティの詳細

apiVersion

apiVersion: string

プロパティ値

string

AnomalyDetector.apiVersionから継承

endpoint

endpoint: string

プロパティ値

string

AnomalyDetector.endpointから継承

継承済みメソッドの詳細

deleteMultivariateModel(string, AnomalyDetectorDeleteMultivariateModelOptionalParams)

modelId に従って、既存の多変量モデルを削除します

function deleteMultivariateModel(modelId: string, options?: AnomalyDetectorDeleteMultivariateModelOptionalParams): Promise<RestResponse>

パラメーター

modelId

string

モデル識別子。

options
AnomalyDetectorDeleteMultivariateModelOptionalParams

options パラメーター。

戻り値

Promise<RestResponse>

AnomalyDetector.deleteMultivariateModelから継承

detectAnomaly(string, DetectionRequest, AnomalyDetectorDetectAnomalyOptionalParams)

多変量異常検出タスクを modelId のトレーニング済みのモデルと共に送信します。入力スキーマは、トレーニング要求と同じである必要があります。 そのため、要求は非同期的に完了し、検出の結果をクエリするための resultId を返します。この要求は、外部からアクセス可能な Azure ストレージの URI (可能であれば Shared Access Signature URI) を示すソース リンクである必要があります。 モデルの生成で使用されたすべての時系列を 1 つのファイルに zip 圧縮する必要があります。 各時系列は、最初の列がタイムスタンプで、2 番目の列が値になります。

function detectAnomaly(modelId: string, body: DetectionRequest, options?: AnomalyDetectorDetectAnomalyOptionalParams): Promise<AnomalyDetectorDetectAnomalyResponse>

パラメーター

modelId

string

モデル識別子。

body
DetectionRequest

異常要求を検出する

options
AnomalyDetectorDetectAnomalyOptionalParams

options パラメーター。

戻り値

AnomalyDetector.detectAnomalyから継承されます

detectChangePoint(DetectChangePointRequest, AnomalyDetectorDetectChangePointOptionalParams)

すべての系列ポイントの変更ポイント スコアを評価する

function detectChangePoint(body: DetectChangePointRequest, options?: AnomalyDetectorDetectChangePointOptionalParams): Promise<AnomalyDetectorDetectChangePointResponse>

パラメーター

body
DetectChangePointRequest

時系列のポイントと粒度が必要です。 必要に応じて、要求で高度なモデル パラメーターを設定することもできます。

options
AnomalyDetectorDetectChangePointOptionalParams

options パラメーター。

戻り値

AnomalyDetector.detectChangePointから継承

detectEntireSeries(DetectRequest, AnomalyDetectorDetectEntireSeriesOptionalParams)

この操作により、系列全体を含むモデルが生成され、各ポイントは同じモデルで検出されます。 このメソッドでは、特定のポイントの前後のポイントを使用して、異常であるかどうかを判断します。 検出全体を使用すると、時系列の全体的な状態をユーザーに提供できます。

function detectEntireSeries(body: DetectRequest, options?: AnomalyDetectorDetectEntireSeriesOptionalParams): Promise<AnomalyDetectorDetectEntireSeriesResponse>

パラメーター

body
DetectRequest

必要に応じて、時系列のポイントと期間。 要求では、高度なモデル パラメーターを設定することもできます。

options
AnomalyDetectorDetectEntireSeriesOptionalParams

options パラメーター。

戻り値

AnomalyDetector.detectEntireSeriesから継承

detectLastPoint(DetectRequest, AnomalyDetectorDetectLastPointOptionalParams)

この操作では、最新のポイント以前のポイントを使用してモデルを生成します。 このメソッドでは、履歴ポイントのみを使用して、ターゲット ポイントが異常であるかどうかを判断します。 最新のポイント検出操作は、ビジネス メトリックのリアルタイム監視のシナリオと一致します。

function detectLastPoint(body: DetectRequest, options?: AnomalyDetectorDetectLastPointOptionalParams): Promise<AnomalyDetectorDetectLastPointResponse>

パラメーター

body
DetectRequest

必要に応じて、時系列のポイントと期間。 要求では、高度なモデル パラメーターを設定することもできます。

options
AnomalyDetectorDetectLastPointOptionalParams

options パラメーター。

戻り値

AnomalyDetector.detectLastPointから継承

exportModel(string, AnomalyDetectorExportModelOptionalParams)

modelId に基づいて、多変量異常検出モデルをエクスポートします

function exportModel(modelId: string, options?: AnomalyDetectorExportModelOptionalParams): Promise<AnomalyDetectorExportModelResponse>

パラメーター

modelId

string

モデル識別子。

options
AnomalyDetectorExportModelOptionalParams

options パラメーター。

戻り値

AnomalyDetector.exportModelから継承

getDetectionResult(string, AnomalyDetectorGetDetectionResultOptionalParams)

DetectAnomalyAsync API によって返された resultId に基づいて、多変量異常検出の結果を取得します

function getDetectionResult(resultId: string, options?: AnomalyDetectorGetDetectionResultOptionalParams): Promise<AnomalyDetectorGetDetectionResultResponse>

パラメーター

resultId

string

結果識別子。

options
AnomalyDetectorGetDetectionResultOptionalParams

options パラメーター。

戻り値

AnomalyDetector.getDetectionResultから継承されます

getMultivariateModel(string, AnomalyDetectorGetMultivariateModelOptionalParams)

多変量モデルの詳細情報 (トレーニングの状態やモデルで使用された変数を含む) を取得します。

function getMultivariateModel(modelId: string, options?: AnomalyDetectorGetMultivariateModelOptionalParams): Promise<AnomalyDetectorGetMultivariateModelResponse>

パラメーター

modelId

string

モデル識別子。

options
AnomalyDetectorGetMultivariateModelOptionalParams

options パラメーター。

戻り値

AnomalyDetector.getMultivariateModelから継承

lastDetectAnomaly(string, LastDetectionRequest, AnomalyDetectorLastDetectAnomalyOptionalParams)

異常検出のための同期された API。

function lastDetectAnomaly(modelId: string, body: LastDetectionRequest, options?: AnomalyDetectorLastDetectAnomalyOptionalParams): Promise<AnomalyDetectorLastDetectAnomalyResponse>

パラメーター

modelId

string

モデル識別子。

body
LastDetectionRequest

最後の検出の要求。

options
AnomalyDetectorLastDetectAnomalyOptionalParams

options パラメーター。

戻り値

AnomalyDetector.lastDetectAnomalyから継承

listMultivariateModel(AnomalyDetectorListMultivariateModelOptionalParams)

サブスクリプションのモデルを一覧表示します

function listMultivariateModel(options?: AnomalyDetectorListMultivariateModelOptionalParams): PagedAsyncIterableIterator<AnomalyDetectorClientModelSnapshot, AnomalyDetectorClientModelSnapshot[], PageSettings>

パラメーター

options
AnomalyDetectorListMultivariateModelOptionalParams

options パラメーター。

戻り値

AnomalyDetector.listMultivariateModelから継承

sendOperationRequest(OperationArguments, OperationSpec, ServiceCallback<any>)

指定された OperationSpec を使用して設定された HTTP 要求を送信します。

function sendOperationRequest(operationArguments: OperationArguments, operationSpec: OperationSpec, callback?: ServiceCallback<any>): Promise<RestResponse>

パラメーター

operationArguments
OperationArguments

HTTP 要求のテンプレート値が設定される引数。

operationSpec
OperationSpec

httpRequest の設定に使用する OperationSpec。

callback

ServiceCallback<any>

応答を受信したときに呼び出すコールバック。

戻り値

Promise<RestResponse>

AnomalyDetector.sendOperationRequestから継承

sendRequest(RequestPrepareOptions | WebResourceLike)

指定された httpRequest を送信します。

function sendRequest(options: RequestPrepareOptions | WebResourceLike): Promise<HttpOperationResponse>

パラメーター

戻り値

AnomalyDetector.sendRequestから継承

trainMultivariateModel(AnomalyDetectorClientModelInfo, AnomalyDetectorTrainMultivariateModelOptionalParams)

多変量異常検出モデルを作成してトレーニングします。 この要求には、外部からアクセス可能な Azure ストレージの URI (可能であれば Shared Access Signature URI) を示すソース パラメーターを含める必要があります。 モデルの生成で使用されたすべての時系列を 1 つのファイルに zip 圧縮する必要があります。 各時系列は、最初の列がタイムスタンプで、2 番目の列が値である 1 つの CSV ファイルに含まれます。

function trainMultivariateModel(body: AnomalyDetectorClientModelInfo, options?: AnomalyDetectorTrainMultivariateModelOptionalParams): Promise<AnomalyDetectorTrainMultivariateModelResponse>

パラメーター

body
AnomalyDetectorClientModelInfo

トレーニング要求

options
AnomalyDetectorTrainMultivariateModelOptionalParams

options パラメーター。

戻り値

AnomalyDetector.trainMultivariateModelから継承