A Unified Approach for Textual and Geometrical Information Retrieval

Abstract

Textual and geometrical algorithms have been considered as two separate fields. This was caused by the fact that textual data are discrete in principal and interpolation is not defined as there is no metric in general, while geometrical data are considered discrete samples of continuous phenomena, geometrical surface etc. In this paper we present a unified approach to textual and geometrical data is hashing technique is used. We summarize principles, knowledge gain and experimental results obtained recently. Hashing techniques enable to answer a query with O(1) complexity that is very important if large textual or geometrical data are to be processed. This approach has lead to a new construction of hash functions for textual and geometrical data as well.