Simple Bayesian Model for Bitmap Compression

  • Authors:
  • A. Bookstein;St Klein;T. Raita

  • Affiliations:
  • University of Chicago, 1010 E. 59 St., Chicago, IL 60637, USA. a-bookstein@uchicago.edu;Department of Mathematics & Computer Science, Bar-Ilan University, Ramat-Gan 52900, Israel. tomi@cs.biu.ac.il;Computer Science Department, University of Turku, SF-20520 Turku, Finland. raita@cs.utu.fi

  • Venue:
  • Information Retrieval
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

Bitmaps are a useful, but storage voracious, component of manyinformation retrieval systems. Earlier efforts to compress bitmapswere based on models of bit generation, particularly Markov models.While these permitted considerable reduction in storage, the shortmemory of Markov models may limit their compression efficiency. Inthis paper we accept the state orientation of Markov models, butintroduce a Bayesian approach to assess the state; the analysis isbased on data accumulating in a growing window. The paper describesthe details of the probabilistic assumptions governing the Bayesiananalysis, as well as the protocol for controlling the window thatreceives the data. We find slight improvement over the bestperforming strictly Markov models.