Geometric Data Analytics, Inc.

Expertise: Dynamics

Detecting periodic patterns in time series data.

Dynamic data is everywhere - physiology, tracking and navigation, social networks, biological systems. Data that change in time present interesting challenges to the design of analytics that are simultaneously efficient and effective. Users of dynamic data also require real time performance, including the monitoring of unusual behavior (anomalies). Even more, users need to anticipate behavioral changes in order to be ready to respond to important events.

GDA specializes in identifying patterns in time series data. We have many original methods for finding recurring or periodic patterns, matching patterns in different data streams, and finding anomalies in data. We use community accepted features in all our work so that users can understand the meaning of what we find. These are used to transform dynamic data in a number of important ways so that Geometric and Topological Methods can be used to find features. Bayesian machine learning methods then allow us to train these transformations and the feature extraction process to find classifiers. Our methods are very adaptable and have been used in all the above mentioned areas as well as in cyber defense.