Senior Honors Thesis and Final ARTU Presentation: "Algorithms for Reducing the Complexity of High-Dimensional Data Sets"
Anthony Grebe
Abstract: Given a set of points {x_n} lying on or near an unknown p-dimensional manifold in R^d, d >> p, we seek to construct a p-dimensional representation of the data. To obtain such a representation, we investigate procedures such as the diffusion map, Isomap, construction of triangulations, and wavelet decomposition. We also explore methods for estimating the number of dimensions p that our representations will need.
Host: Victor Wickerhauser