Mathematics Colloquia and Seminars
The generalized Haar-Walsh transform on graph and further applications on matricesStudent-Run Applied & Math Seminar
|Speaker:||Yiqun Shao, UC Davis|
|Start time:||Thu, Oct 12 2017, 12:10PM|
My research is a continuation of previous PHD student Jeff Irion's work. The GHWT, generalized Haar-Walsh transform, is useful for data analysis on graph, for example, data compression or noise reduction. It involves recursive partitioning of the graph through fiedler vector of graph laplacian, building wavelet dictionary with the partition tree and searching best basis of the data set on the graph.
We will first give a short description of GWHT. Then we will introduce a new best-basis search algorithm with time-frequency analysis. Finally, we will show how to extend the applications of GHWT to matrices, such as term-document matrix in text analysis.
If you would like to help us estimate how much pizza to order you can RSVP at this link, even if you aren't sure you can make it: https://docs.google.com/spreadsheets/d/1wyOmPJvaqsSngBuIcjnN3vLAWeRFTyru47HFaT0wlMc/edit#gid=1812354212