
Project Type MQP Submission date 2009-04-27 Author Matthew Thomas Houde, MA URN E-project-042709-214204 Title Almost Independent Binary Random Variables Advisor Martin, William J, MA Availability unrestricted Abstract
A collection C of binary n-tuples is considered t-wise independent if the projection onto any t coordinates is uniformly distributed as c is chosen uniformly from C. Notice that for C to be t-wise independent, |C| must be greater than 2^t; for many applications this is too large. The aim of this project is to decrease the size of |C| while still allowing the projection onto any t coordinates to appear uniformly distributed in {0, 1}^t. In this paper I will be presenting two definitions for almost t-wise independence. Through coding theory tools I will show bounds imposed on the size of C based on those definitions. Through known constructions of almost independent binary random variables, I will then demonstrate how the definitions and bounds I established apply to these constructions.
Files Almost_Independent_Binary_Random_Variables.pdf There is one file which has been withheld at the advisor's request.
Browse by Author | Browse by Department | Search all available E-projects
Questions? Email project-questions@wpi.edu