Year: 2004 | ![Adobe Acrobat Reader Adobe Acrobat Reader](https://webarchive.library.unt.edu/eot2008/20090113173041im_/http://lhncbc.nlm.nih.gov/cgsb_site/images/cgsb/misc/getacro.gif)
Download Free Adobe Acrobat Reader |
LHNCBC-2004-022 |
Methods for Diagnosing Physiological Conditions of Human Subjects from Multivariate Time Series Biosensor Data |
Kayaalp M |
Physiological Data Modeling Contest, the Twenty-First International Conference on Machine Learning (ICML), 2004 July. |
In this study, we developed a set of new methods to diagnose physiological conditions (PCs) of human subjects based on biometric multi-channel time series data that were obtained through nine biosensors attached on the upper arms of the human test subjects. The task is to learn the relationships between PCs and the biosensor data on the training dataset and identify the PCs of interest on the test dataset. The goal of this study is to test the predictive performances of two sets of Bayesian network topologies and the effectiveness of a new parameter learning method applied to the problem on the provided dataset. |
PDF |