Azure AI 异常检测 Java SDK

Azure AI Anomaly Detector Java SDK

使用 Azure AI Anomaly Detector Java SDK 构建异常检测应用程序。

安装

<dependency>
  <groupId>com.azure</groupId>
  <artifactId>azure-ai-anomalydetector</artifactId>
  <version>3.0.0-beta.6</version>
</dependency>

客户端创建

同步与异步客户端

import com.azure.ai.anomalydetector.AnomalyDetectorClientBuilder;
import com.azure.ai.anomalydetector.MultivariateClient;
import com.azure.ai.anomalydetector.UnivariateClient;
import com.azure.core.credential.AzureKeyCredential;
String endpoint = System.getenv("AZURE_ANOMALY_DETECTOR_ENDPOINT");
String key = System.getenv("AZURE_ANOMALY_DETECTOR_API_KEY");
// 用于多个相关信号的多变量客户端
MultivariateClient multivariateClient = new AnomalyDetectorClientBuilder()
    .credential(new AzureKeyCredential(key))
    .endpoint(endpoint)
    .buildMultivariateClient();
// 用于单变量分析的单变量客户端
UnivariateClient univariateClient = new AnomalyDetectorClientBuilder()
    .credential(new AzureKeyCredential(key))
    .endpoint(endpoint)
    .buildUnivariateClient();

使用 DefaultAzureCredential

import com.azure.identity.DefaultAzureCredentialBuilder;
MultivariateClient client = new AnomalyDetectorClientBuilder()
    .credential(new DefaultAzureCredentialBuilder().build())
    .endpoint(endpoint)
    .buildMultivariateClient();

核心概念

单变量异常检测 (Univariate Anomaly Detection)

• 批量检测 (Batch Detection)：一次性分析整个时间序列

• 流式检测 (Streaming Detection)：对最新数据点进行实时检测

• 变化点检测 (Change Point Detection)：检测时间序列中的趋势变化

多变量异常检测 (Multivariate Anomaly Detection)

• 检测 300 多个相关信号中的异常

• 使用图注意力网络 (Graph Attention Network) 处理相互关联性

• 三步流程：训练 $\rightarrow$ 推理 $\rightarrow$ 结果

核心模式

单变量批量检测

import com.azure.ai.anomalydetector.models.*;
import java.time.OffsetDateTime;
import java.util.List;
List<TimeSeriesPoint> series = List.of(
    new TimeSeriesPoint(OffsetDateTime.parse("2023-01-01T00:00:00Z"), 1.0),
    new TimeSeriesPoint(OffsetDateTime.parse("2023-01-02T00:00:00Z"), 2.5),
    // ... 更多数据点（至少需要 12 个点）
);
UnivariateDetectionOptions options = new UnivariateDetectionOptions(series)
    .setGranularity(TimeGranularity.DAILY)
    .setSensitivity(95);
UnivariateEntireDetectionResult result = univariateClient.detectUnivariateEntireSeries(options);
// 检查异常
for (int i = 0; i < result.getIsAnomaly().size(); i++) {
    if (result.getIsAnomaly().get(i)) {
        System.out.printf("在索引 %d 处检测到异常，值为 %.2f%n",
            i, series.get(i).getValue());
    }
}

单变量最后一点检测（流式）

UnivariateLastDetectionResult lastResult = univariateClient.detectUnivariateLastPoint(options);
if (lastResult.isAnomaly()) {
    System.out.println("最新点是异常值！");
    System.out.printf("预期值: %.2f, 上限: %.2f, 下限: %.2f%n",
        lastResult.getExpectedValue(),
        lastResult.getUpperMargin(),

### 变更点检测

UnivariateChangePointDetectionResult changeResult =
univariateClient.detectUnivariateChangePoint(changeOptions);

for (int i = 0; i < changeResult.getIsChangePoint().size(); i++) {
if (changeResult.getIsChangePoint().get(i)) {
System.out.printf("Change point at index %d with confidence %.2f%n",
i, changeResult.getConfidenceScores().get(i));
}
}

### 多变量模型训练

// 使用 Blob 存储数据准备训练请求
ModelInfo modelInfo = new ModelInfo()
.setDataSource("https://storage.blob.core.windows.net/container/data.zip?sasToken")
.setStartTime(OffsetDateTime.parse("2023-01-01T00:00:00Z"))
.setEndTime(OffsetDateTime.parse("2023-06-01T00:00:00Z"))
.setSlidingWindow(200)
.setDisplayName("MyMultivariateModel");

// 训练模型（长时间运行操作）
AnomalyDetectionModel trainedModel = multivariateClient.trainMultivariateModel(modelInfo);

String modelId = trainedModel.getModelId();
System.out.println("Model ID: " + modelId);

// 检查训练状态
AnomalyDetectionModel model = multivariateClient.getMultivariateModel(modelId);
System.out.println("Status: " + model.getModelInfo().getStatus());

### 多变量批量推理

MultivariateDetectionResult detectionResult =
multivariateClient.detectMultivariateBatchAnomaly(modelId, detectionOptions);

String resultId = detectionResult.getResultId();

// 轮询结果
MultivariateDetectionResult result = multivariateClient.getBatchDetectionResult(resultId);
for (AnomalyState state : result.getResults()) {
if (state.getValue().isAnomaly()) {
System.out.printf("Anomaly at %s, severity: %.2f%n",
state.getTimestamp(),
state.getValue().getSeverity());
}
}

### 多变量最后一点检测

MultivariateLastDetectionResult lastResult =
multivariateClient.detectMultivariateLastAnomaly(modelId, lastOptions);

if (lastResult.getValue().isAnomaly()) {
System.out.println("Anomaly detected!");
// 检查贡献变量
for (AnomalyContributor contributor : lastResult.getValue().getInterpretation()) {
System.out.printf("Variable: %s, Contribution: %.2f%n",
contributor.getVariable(),
contributor.getContributionScore());
}
}

### 模型管理

Status: %s%n",
        m.getModelId(),
        m.getModelInfo().getStatus());
}

// 删除模型
multivariateClient.deleteMultivariateModel(modelId);

错误处理

import com.azure.core.exception.HttpResponseException;
try {
    univariateClient.detectUnivariateEntireSeries(options);
} catch (HttpResponseException e) {
    System.out.println("状态码: " + e.getResponse().getStatusCode());
    System.out.println("错误: " + e.getMessage());
}

环境变量

AZURE_ANOMALY_DETECTOR_ENDPOINT=https://<resource>.cognitiveservices.azure.com/
AZURE_ANOMALY_DETECTOR_API_KEY=<your-api-key>

最佳实践

1. 最小数据点：单变量检测至少需要 12 个点；数据越多，准确度越高。
2. 粒度对齐：确保 TimeGranularity 与实际数据的频率一致。
3. 灵敏度调优：较高的值（0-99）会检测到更多异常。
4. 多变量训练：根据模式复杂度，使用 200-1000 的滑动窗口。
5. 错误处理：始终处理 HttpResponseException 以应对 API 错误。

触发词

• "anomaly detection Java"

• "detect anomalies time series"

• "multivariate anomaly Java"

• "univariate anomaly detection"

• "streaming anomaly detection"

• "change point detection"

• "Azure AI Anomaly Detector"

使用场景

局限性

• 仅在任务明确符合上述范围时使用此技能。

• 不要将输出视为针对特定环境的验证、测试或专家评审的替代方案。

• 如果缺失必要的输入、权限、安全边界或成功标准，请停止并请求澄清。