System Driven Anomaly Detection¶
The "Anomaly Detection" category offers automated journey for identifying the variations in your time series dataset as part of Predict offering using advanced machine learning technologies.
System driven anomaly detection engine is created based on the anomaly profile observed from all the datapoints of the dataset. This engine is not monitoring the anomalous behaviors of only one target datapoint but analyzing the anomaly profile through multiple datapoints with different contributions to the anomalies. In industrial scenarios, it could contribute to monitoring continuously the anomalous behaviors from multiple sensors which could give early alerts to avoid more serious malfunctions in the near future from more systematic and comprehensive perspectives.
Anomaly Detection could leverage data from asset and performs easy configuration for engine building with anomaly profile and detection logic.
Note
- "engine": In this documentation, the term "engine" (e.g., "engine of prediction", "prediction engine", "anomaly detection engine") is used for domain experts from shopfloor. It is similar to what data scientists call a "model". This term is used because it better describes a powerful system that works continuously to analyze your assets and generate insights.
User Interface¶
The following images show the Anomaly Detection screen:
① Displays the dataset information
② Displays the anomaly detection approach. Use System Driven
③ Displays the advanced configuration parameters
④ Displays schedule configuration
Perform System Driven Anomaly Detection¶
To execute system driven anomaly detection, follow these steps:
- Select an asset from the list on the home page.
- Add a new dataset or choose an existing one, then click on the "Anomaly Detections" link.
- Click the "Create anomaly detection" button.
- Choose "System Driven" in the anomaly detection approach section.
- Configure the "Data Imputation" setting by selecting the type from the drop-down menu.
- Adjust the contamination setting by deselecting "auto" and providing a value in the range of 0.1 to 0.5.
- Schedule system driven anomaly detection by providing occurrence details. Select the preferred frequency for executing the anomaly detection engine. The schedule can be defined as per the below intervals:
- Minutes
- Hourly
- Daily
- Weekly
- Minutes
- Select the date in the "Schedule Start Date" and "Schedule End Date".
- Click the "Submit" button.
The generated anomaly detection engine can be used to visualizes the anomaly spots, indicators.
Anomaly Detection Results¶
Once the anomaly detection is scheduled for production, the anomalies are generated as per the schedule. The generated anomalies can be used to analyze the data for the all variables of the asset. To view the anomaly detection, click "Production Detections" and select the preferred anomaly detections from the "Detail result for anomaly execution" drop-down.
Feedback¶
Users can provide feedback if the generated results do not align with the actual system status. For example, if anomalies are detected by the engine but are false detection, users can provide feedback by selecting the time range in the graph and submitting their feedback in the pop-up window.
