It turns out that many problems can be cast in terms of anomaly detection, and thus, a small number of methods can be applicable to a vast number of radically different applications. We have perfected methods for anomaly detection under very noisy (and multiple stationary) time series. Rather than using traditional filtering methods to reduce noise, we use a combination of machine learning methods to reduce noise and denoising techniques developed by Rafi Coifman, David Donoho and colleagues.
These methods are applicable to early epilepsy detection, early Congestive Heart Failure detection, earthquake detection, financial and various Homeland and Cyber security appications.
See publications for more details, or contact us with specific questions at firstname.lastname@example.org.