Dr. Yu-Ting Lin (Department of Anesthesiology, Taipei Veterans General Hospital)
Title
Practical Application of Biomedical Signal Waveform Data and Interactions in Spectral Embedding of Graph Laplacian
Abstract
Biomedical signal data possess dynamical and complex waveform change, which inspired us to develop methods of manifold learning to resolve the structure inside the data. With the spectral decomposition of graph Laplacian, we can embed data in one lower dimensional space to study its global structure and local detail. Two datasets inside one embedding interact with each other. In practical use, we can exploit this property to extract more information by adding more concomitant datasets. It is worth pursuing an Universal Coordinate idea with the adequate available data.
