Voting prediction using new spatiotemporal interpolation methods

  • Authors:
  • Jun Gao;Peter Revesz

  • Affiliations:
  • University of Nebraska-Lincoln, Lincoln, NE;University of Nebraska-Lincoln, Lincoln, NE

  • Venue:
  • dg.o '06 Proceedings of the 2006 international conference on Digital government research
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

Most spatial and spatiotemporal interpolation methods give back a surface function as the result. Instead of that we consider interpolation methods that yield a single value as the final result. Voting prediction is a natural example that requires this type of spatiotemporal interpolation, because the final result is the total percentage vote for a party or candidate. We propose a new spatiotemporal interpolation method for voting prediction and similar problems. The approach can also be used in election data verification for effective government. We test the new method using USA presidential election data from the states of California, Florida, and Ohio between 1972 and 2004. The experimental results show that our method can produce comparatively precise predictions (e.g., the difference between prediction and actual result is 1.09% for Florida in 2004).