Information retrieval in big data using cognitive approaches
Santanu Acharjee, Ripunjoy Choudhury
TL;DR
The paper addresses the challenge of information retrieval in big data by proposing a topology-based cognitive information retrieval (CIR) framework that leverages cognitive similarity and retrieval topologies to semantically align queries with documents. It formalizes cognitive distance and retrieval functions, proves stability under small query perturbations, and extends CIR to Boolean logic through AND, OR, and NOT connectives, achieving retrieval sets that behave like intersections, unions, and complements. Key contributions include a cognitive retrieval topology that captures all Boolean cognitive retrievals, context-based extensions, and stability results that support robust, large-scale search. The work lays a foundation for context-aware, scalable CIR in big data, with potential expansions to ranking, multimodal data, and real-time feedback systems.
Abstract
Due to the exponential growth of big data in this digital era, an advanced method for effective information retrieval becomes essential. The basic objective of this paper is to propose a topology-based method for cognitive information retrieval (CIR) in big data environments. By using concepts such as cognitive similarity distances, metric spaces, retrieval topologies, etc., this paper aims to propose the semantic alignment between user queries and document repositories. The paper also extends this approach to incorporate logical connectives in cognitive information retrieval.