AnomalyDetectorClient class

與 Azure 異常偵測程式 服務互動的用戶端類別。

Extends

建構函式

AnomalyDetectorClient(string, TokenCredential | KeyCredential, PipelineOptions)

建立 AnomalyDetectorClient 的實例。

使用方式範例:

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

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

繼承的屬性

apiVersion
endpoint

繼承的方法

deleteMultivariateModel(string, AnomalyDetectorDeleteMultivariateModelOptionalParams)

根據 modelId 刪除現有的多變數模型

detectAnomaly(string, DetectionRequest, AnomalyDetectorDetectAnomalyOptionalParams)

使用 modelId 的定型模型提交偵測多重變數異常工作,輸入架構應該與定型要求相同。 因此,要求會以非同步方式完成,並傳回 resultId 來查詢偵測結果。要求應該是來源連結,以指出外部存取的 Azure 儲存體 URI (最好是共用存取簽章 URI) 。 產生模型時所使用的所有時間序列都必須壓縮成一個單一檔案。 每個時間序列都會如下所示:第一個資料行是時間戳記,而第二個數據行是值。

detectChangePoint(DetectChangePointRequest, AnomalyDetectorDetectChangePointOptionalParams)

評估每個數列點的變更點分數

detectEntireSeries(DetectRequest, AnomalyDetectorDetectEntireSeriesOptionalParams)

這項作業會產生具有整個數列的模型,每個點都會以相同的模型來偵測。 使用此方法時,會使用特定點前後的點來判斷它是否為異常。 整個偵測可為使用者提供時間序列的整體狀態。

detectLastPoint(DetectRequest, AnomalyDetectorDetectLastPointOptionalParams)

此作業會使用最新點之前的點產生模型。 使用此方法時,只會使用歷程記錄點來判斷目標點是否為異常。 最新偵測作業符合即時監視商務計量的案例。

exportModel(string, AnomalyDetectorExportModelOptionalParams)

根據 modelId 匯出多重變數異常偵測模型

getDetectionResult(string, AnomalyDetectorGetDetectionResultOptionalParams)

根據 DetectAnomalyAsync api 所傳回的 resultId 取得多重變數異常偵測結果

getMultivariateModel(string, AnomalyDetectorGetMultivariateModelOptionalParams)

取得多重變數模型的詳細資訊,包括模型中使用的定型狀態和變數。

lastDetectAnomaly(string, LastDetectionRequest, AnomalyDetectorLastDetectAnomalyOptionalParams)

已同步處理異常偵測的 API。

listMultivariateModel(AnomalyDetectorListMultivariateModelOptionalParams)

列出訂用帳戶的模型

sendOperationRequest(OperationArguments, OperationSpec, ServiceCallback<any>)

傳送使用提供的 OperationSpec 填入的 HTTP 要求。

sendRequest(RequestPrepareOptions | WebResourceLike)

傳送提供的 HTTPRequest。

trainMultivariateModel(AnomalyDetectorClientModelInfo, AnomalyDetectorTrainMultivariateModelOptionalParams)

建立和定型多重變數異常偵測模型。 要求必須包含來源參數,以指出外部可存取的 Azure 儲存體 URI (最好是共用存取簽章 URI) 。 產生模型時所使用的所有時間序列都必須壓縮成一個單一檔案。 每個時間序列都會位於單一 CSV 檔案中,其中第一個資料行是時間戳記,而第二個數據行是值。

建構函式詳細資料

AnomalyDetectorClient(string, TokenCredential | KeyCredential, PipelineOptions)

建立 AnomalyDetectorClient 的實例。

使用方式範例:

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)

參數

endpointUrl

string

Azure 異常偵測程式服務端點的 URL

credential

TokenCredential | KeyCredential

用來驗證對服務的要求。

options
PipelineOptions

用來設定表格辨識器用戶端。

繼承的屬性詳細資料

apiVersion

apiVersion: string

屬性值

string

繼承自AnomalyDetector.apiVersion

endpoint

endpoint: string

屬性值

string

繼承自AnomalyDetector.endpoint

繼承的方法的詳細資料

deleteMultivariateModel(string, AnomalyDetectorDeleteMultivariateModelOptionalParams)

根據 modelId 刪除現有的多變數模型

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

參數

modelId

string

模型識別碼。

傳回

Promise<RestResponse>

繼承自AnomalyDetector.deleteMultivariateModel

detectAnomaly(string, DetectionRequest, AnomalyDetectorDetectAnomalyOptionalParams)

使用 modelId 的定型模型提交偵測多重變數異常工作,輸入架構應該與定型要求相同。 因此,要求會以非同步方式完成,並傳回 resultId 來查詢偵測結果。要求應該是來源連結,以指出外部存取的 Azure 儲存體 URI (最好是共用存取簽章 URI) 。 產生模型時所使用的所有時間序列都必須壓縮成一個單一檔案。 每個時間序列都會如下所示:第一個資料行是時間戳記,而第二個數據行是值。

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

參數

modelId

string

模型識別碼。

body
DetectionRequest

偵測異常要求

傳回

繼承自AnomalyDetector.detectAnomaly

detectChangePoint(DetectChangePointRequest, AnomalyDetectorDetectChangePointOptionalParams)

評估每個數列點的變更點分數

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

參數

body
DetectChangePointRequest

需要時間序列點和細微性。 如有需要,也可以在要求中設定進階模型參數。

傳回

繼承自AnomalyDetector.detectChangePoint

detectEntireSeries(DetectRequest, AnomalyDetectorDetectEntireSeriesOptionalParams)

這項作業會產生具有整個數列的模型,每個點都會以相同的模型來偵測。 使用此方法時,會使用特定點前後的點來判斷它是否為異常。 整個偵測可為使用者提供時間序列的整體狀態。

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

參數

body
DetectRequest

如有需要,時間序列點和期間。 您也可以在要求中設定進階模型參數。

傳回

繼承自AnomalyDetector.detectEntireSeries

detectLastPoint(DetectRequest, AnomalyDetectorDetectLastPointOptionalParams)

此作業會使用最新點之前的點產生模型。 使用此方法時,只會使用歷程記錄點來判斷目標點是否為異常。 最新偵測作業符合即時監視商務計量的案例。

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

參數

body
DetectRequest

如有需要,時間序列點和期間。 您也可以在要求中設定進階模型參數。

傳回

繼承自AnomalyDetector.detectLastPoint

exportModel(string, AnomalyDetectorExportModelOptionalParams)

根據 modelId 匯出多重變數異常偵測模型

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

參數

modelId

string

模型識別碼。

傳回

繼承自AnomalyDetector.exportModel

getDetectionResult(string, AnomalyDetectorGetDetectionResultOptionalParams)

根據 DetectAnomalyAsync api 所傳回的 resultId 取得多重變數異常偵測結果

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

參數

resultId

string

結果識別碼。

傳回

繼承自AnomalyDetector.getDetectionResult

getMultivariateModel(string, AnomalyDetectorGetMultivariateModelOptionalParams)

取得多重變數模型的詳細資訊,包括模型中使用的定型狀態和變數。

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

參數

modelId

string

模型識別碼。

傳回

繼承自AnomalyDetector.getMultivariateModel

lastDetectAnomaly(string, LastDetectionRequest, AnomalyDetectorLastDetectAnomalyOptionalParams)

已同步處理異常偵測的 API。

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

參數

modelId

string

模型識別碼。

body
LastDetectionRequest

上次偵測的要求。

傳回

繼承自AnomalyDetector.lastDetectAnomaly

listMultivariateModel(AnomalyDetectorListMultivariateModelOptionalParams)

列出訂用帳戶的模型

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

參數

傳回

繼承自AnomalyDetector.listMultivariateModel

sendOperationRequest(OperationArguments, OperationSpec, ServiceCallback<any>)

傳送使用提供的 OperationSpec 填入的 HTTP 要求。

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

參數

operationArguments
OperationArguments

HTTP 要求樣板化值的引數將會從 中填入。

operationSpec
OperationSpec

用來填入 HTTPRequest 的 OperationSpec。

callback

ServiceCallback<any>

收到回應時要呼叫的回呼。

傳回

Promise<RestResponse>

繼承自AnomalyDetector.sendOperationRequest

sendRequest(RequestPrepareOptions | WebResourceLike)

傳送提供的 HTTPRequest。

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

參數

傳回

繼承自AnomalyDetector.sendRequest

trainMultivariateModel(AnomalyDetectorClientModelInfo, AnomalyDetectorTrainMultivariateModelOptionalParams)

建立和定型多重變數異常偵測模型。 要求必須包含來源參數,以指出外部可存取的 Azure 儲存體 URI (最好是共用存取簽章 URI) 。 產生模型時所使用的所有時間序列都必須壓縮成一個單一檔案。 每個時間序列都會位於單一 CSV 檔案中,其中第一個資料行是時間戳記,而第二個數據行是值。

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

參數

body
AnomalyDetectorClientModelInfo

訓練要求

傳回

繼承自AnomalyDetector.trainMultivariateModel