Anomaly detection is the process of identifying data points that deviate significantly from the expected or normal behavior of a dataset. These data points are often referred to as anomalies, outliers, or novelties.
Anomaly detection is commonly used in various fields, such as fraud detection, intrusion detection, fault detection, and predictive maintenance, among others. The goal of anomaly detection is to identify unusual or suspicious events that could indicate a problem or opportunity.
There are different methods for anomaly detection, including statistical methods, machine learning techniques, and rule-based approaches. Statistical methods typically rely on defining a threshold or range of expected behavior and flagging data points that fall outside of this range as anomalies. Machine learning methods can learn the expected behavior of a dataset and detect deviations from it. Rule-based approaches use predefined rules to identify anomalies based on specific criteria.
Overall, anomaly detection is a useful technique for identifying unusual events in a dataset and can help improve decision-making and prevent potential problems.
In this context, an 'unusual' event is a buy or sell indication, which would typically occur only 1-2% of the time
Towards Data Science - "Anomaly Detection: A Survey": https://towardsdatascience.com/anomaly-detection-a-survey-7c7694035e6a
Machine Learning Mastery - "Anomaly Detection with Machine Learning in Python": https://machinelearningmastery.com/model-based-outlier-detection-and-removal-in-python/
Anomaly Detection - Wikipedia: https://en.wikipedia.org/wiki/Anomaly_detection
Anomaly Detection: A Comprehensive Review - Springer Link: https://link.springer.com/article/10.1007/s10994-019-05855-5
Anomaly Detection Techniques in Data Science - Analytics Vidhya: https://www.analyticsvidhya.com/blog/2019/02/outlier-detection-python-pyod/