Geometric Clustering and its Applications
Piyush Kumar · 2013
Abstract : The AFOSR YIP Grant led to many activities and findings at FSU including the graduation of two PhD students who are now well placed. This report details the activities and findings of our project. The highlights of the achievements include: 1) Design of a new Support Vector Machine algorithm that is linearly convergent and yields an optimal number of support vectors. 2) A fast nearest neighbor algorithm and its optimized implementation in low dimensions. 3) A real time method to compute centers on maps given multiple addresses, when the distance is measured using shortest paths. 4) A 2-clustering algorithm for segments which is a pre-cursor to clustering curves that represent flight paths in air space. 5) A fast Euclidean MST algorithm and its implementation that computes clusterings with max spacing.