The Technology:

A researcher at the University of Tennessee has developed an award-winning method for predictive analysis. Real time monitoring of multi-channel signals is an essential tool for a variety of applications such as clinical diagnosis, research, fault detection, etc. Transient corruption or loss of signal can be disruptive, especially when continuous monitoring is required to rule out rare events or when the data is used as a basis for forecasting. This software, based on feed-forward neural networks, uses two techniques: iterative retraining and accumulated average. It has been demonstrated that these techniques significantly improve the accuracy of signal reconstruction over standard network models. The algorithm also received an award at the “Mind the Gap” challenge hosted by 2010 Computing in Cardiology/PhysioNet.

Benefits:

• Versatile code that can be customized for different computing platforms and applications

• Method fills gaps in multi-parameter datasets

• Robust estimation of parameters when the primary signal becomes unavailable or unreliable

• Method can be used to recognize intervals of signal corruption