Session Detail Information
Add this session to your itinerary

Cluster :  Tutorials

Session Information  : Monday Nov 06, 08:00 - 09:30

Title:  Tutorial: Semidefinite and Second Order Cone Programming and Their Application to Shape Constrained Regression and Density Estimation

Abstract Details

Title: Semidefinite and Second Order Cone Programming and their Application to Shape Constrained Regression and Density Estimation
 Presenting Author: Farid Alizadeh,Associate Professor, Management Science and Information Systems and RUTCOR Rutgers, State University of New Jersey, 640 Bartholomew, Piscataway NJ 08854, United States, alizadeh@rutcor.rutgers.edu
 
Abstract: Many statistical analyses require the approximation of a functional relationship between explanatory and response variables, subject to certain "shape constraints". This implies a connection between shape-constrained regression and semidefinite programming. In this tutorial, we introduce semidefinite programming and second order cone programming, review specific statistical problems amenable to solution via semidefinite programming, and discuss extensions and limitations to this method.