Skip to content

Anomaly Detection

The Anomaly detection extension as part of AI for Everyone offering on MindSphere, offers an out-of-the-box model building and detect the variations in your time series data values, using advanced artificial intelligence and machine learning technologies.

Anomaly detection uses time series data from MindSphere asset as dataset and does easy configuration for model building and detection of anomalies.

The Anomaly detection report visualizes the anomaly spots, indicators, which could be further utilized for your targeted use cases like comparing with threshold value, reviewing historic records.

User Interface of the Anomaly Detection

The following figure schows the anomaly detection screen:

Anomaly-UI ① Use case details of the Anomaly detection job
② Data visualization area
③ Anomaly detection configuration settings
④ Anomaly executions

Build the Anomaly Detection model

To build the anomaly detection model, proceed with the following steps: 1. Open "AI for Everyone" in the "Analyze" tab. 2. Select the "Use Case", and click "Anomaly Detections". 3. In the "In-sample Period" configure setting, select the time period required for the training dataset for the model building. 4. Select the target variable and the input variable of your choice in the Influencers setting. 5. Configure the Perspective and Sensitivity settings: - Residual Sensitivity: Detects the residuals with significantly higher magnitude than those observed on the in-sample period as anomalous. - Residual Change: Detects the most extreme changes compared to the anomalous behavior model can see on the in-sample period. - Fluctuation: Anomalous behavior model detects anomalies if different fluctuations are observed than those which were present during the in-sample period. - Fluctuation Change: The fluctuation change perspective focuses on the fluctuation of the residuals output similarly to the fluctuation perspective. However, the difference is that the fluctuation change perspective seeks only for the change in the fluctuation. - Imbalance Sensitivity: The imbalance perspective helps to detect anomalies accompanied by deviation of the residuals output from zero for a longer period of time. - Imbalance Change: Anomalous behavior model with an imbalance change perspective is suited for detecting anomalies that occur when a change of imbalance in the residuals output is observed.

  1. Configure the Normal Behaivor setting.
  2. Configure the Data Imputation setting.
  3. Click "Build a model".

Anomalymodelbuilding

Detect the Anomalies

Once the model for anomaly detection is built, configure the "Out of sample period" setting, the time period to execute the anamoly detection. Click "Detect" to start detecting the anomaly spots for the selected dataset.

Anomaly-Detection

Once the Anomaly detection is completed, the anomaly spots are displayed as shown below:

Anomalydetection-results

① Use case details of the Anomaly detection
② Data visualization area
③ Anamoly detection indicators as per the perspectives
④ Anomaly detection configuration settings
⑤ Anomaly executions

Any questions left?

Ask the community


Except where otherwise noted, content on this site is licensed under the MindSphere Development License Agreement.


Last update: July 29, 2022