The term „Class Imbalance“ originates from the fields of Artificial Intelligence, Big Data and Smart Data, and Automation. It describes a situation in the development of AI systems or the analysis of large data sets where the individual groups or "classes" in the data set are very unevenly distributed.
This often happens, for example, if a dataset for an AI designed to detect online banking fraud contains 9900 normal transactions but only 100 fraudulent ones. The AI model then primarily „learns“ what normal transactions look like because they occur much more frequently. This can lead to fraudulent cases being overlooked because they are relatively rare.
Class imbalance is important because it can significantly impair the results of data analyses and the performance of artificial intelligence. Developers must specifically counteract this by, for example, collecting additional data for the rare classes or using special balancing methods. Only in this way can truly reliable and fair AI solutions be created.
Deutsch 
English (UK)