Information Retrieval using probability and belief theory

摘要

This paper presents two approaches for Information Retrieval (IR) from a collection of documents: Bayesian theory of probability and Dempster-Shafer theory of belief functions. Each method has been supported with essential derivations to prove their suitability for IR. The conclusions of derivations have been applied in illustrative examples. Finally, comparison of both the methods suggests the suitability of each for specific domains.