# Evaluating Real-time Anomaly Detection Algorithms - the Numenta Anomaly Benchmark

Much of the world's data is streaming, time-series data, where anomalies give significant information in critical situations; examples abound in domains such as finance, IT, security, medical, and energy. Yet detecting anomalies in streaming data is a difficult task, requiring detectors to process data in real-time, not batches, and learn while simultaneously making predictions... (read more)

PDF Abstract