What is anomaly score in SAP IOT-PDM - Predictive Maintenance?

SAP Term: anomaly score

• Component: IOT-PDM

• Component Name: Predictive Maintenance

• Description: A value indicating the actual behavior of an asset as compared to the required behavior. The anomaly score is calculated on the basis of specific algorithms and indicates whether an asset's behavior is anomalous.

Smart SAP Assistant

• Key Concepts: 
  Anomaly score is a metric used in SAP’s IOT-PDM Predictive Maintenance component to measure the deviation of a given data point from the expected value. It is calculated by taking the difference between the actual value and the expected value, and then dividing it by the standard deviation of the expected values. The result is a score that indicates how far away from the expected value a given data point is. 
  
  How to use it: 
  Anomaly score can be used to identify potential problems with equipment or processes. By monitoring the anomaly scores of various data points, it is possible to detect when something is not functioning as expected and take corrective action. Additionally, anomaly scores can be used to compare different pieces of equipment or processes to determine which ones are performing better or worse than expected. 
  
  Tips & Tricks: 
  When using anomaly scores, it is important to remember that they are relative measures. This means that they should not be used as absolute indicators of performance, but rather as indicators of relative performance. Additionally, it is important to remember that anomaly scores are only useful when compared to other data points or processes. 
  
  Related Information: 
  Anomaly scores are closely related to other metrics such as root mean square error (RMSE) and mean absolute error (MAE). These metrics measure the difference between actual and predicted values in different ways, but all three can be used to identify potential problems with equipment or processes. Additionally, anomaly scores can be used in conjunction with other predictive maintenance techniques such as machine learning algorithms to improve accuracy and reliability.