Negative examples for sequential importance sampling of binary contingency tables

Authors:
Ivona Bezáková;Alistair Sinclair;Daniel Štefankovič;Eric Vigoda
Affiliations:
Department of Computer Science, University of Chicago, Chicago, IL;Computer Science Division, University of California, Berkeley, CA;Department of Computer Science, University of Rochester, Rochester, NY;College of Computing, Georgia Institute of Technology, Atlanta, GA
Venue:
ESA'06 Proceedings of the 14th conference on Annual European Symposium - Volume 14
Year:
2006

Citing 3
Cited 2

A polynomial-time approximation algorithm for the permanent of a matrix with nonnegative entries

Journal of the ACM (JACM)
Sampling binary contingency tables with a greedy start

SODA '06 Proceedings of the seventeenth annual ACM-SIAM symposium on Discrete algorithm
Accelerating simulated annealing for the permanent and combinatorial counting problems

SODA '06 Proceedings of the seventeenth annual ACM-SIAM symposium on Discrete algorithm

Efficient circuits for quantum walks

Quantum Information & Computation
Uniformly generating origin destination tables

Information Theory, Combinatorics, and Search Theory

Quantified Score

Hi-index	0.00

Visualization

Abstract

The sequential importance sampling (SIS) algorithm has gained considerable popularity for its empirical success. One of its noted applications is to the binary contingency tables problem, an important problem in statistics, where the goal is to estimate the number of 0/1 matrices with prescribed row and column sums. We give a family of examples in which the SIS procedure, if run for any subexponential number of trials, will underestimate the number of tables by an exponential factor. This result holds for any of the usual design choices in the SIS algorithm, namely the ordering of the columns and rows. These are apparently the first theoretical results on the efficiency of the SIS algorithm for binary contingency tables. Finally, we present experimental evidence that the SIS algorithm is efficient for row and column sums that are regular. Our work is a first step in determining rigorously the class of inputs for which SIS is effective.