Approximate entropy (ApEn) is a technique used to quantify the unpredictability and regularity of fluctuations in time-series data. It reflects the likelihood that similar patterns will not be followed by additional similar observations. ApEn is useful for relatively short, noisy time-series data, as it is less affected by noise and has lower computational demands than other complexity measures. ApEn has been used to successfully distinguish patient groups in applications like endocrine hormone secretion and epilepsy detection from EEG data, with accuracies over 90% in some cases. It has advantages over entropy as it can be used on smaller sample sizes and applied in real-time applications.