An Automat for the Semantic Processing of Structured Information

TL;DR

This paper presents an automated system for semantic processing and indexing of structured information, inspired by human cognitive strategies, improving document retrieval accuracy and efficiency.

Contribution

It introduces an automat that simulates human indexing using a cognitive model, integrating ontology and graph algorithms for enhanced information retrieval.

Findings

Improved indexing exhaustivity

Enhanced precision and retrieval accuracy

High efficiency in document processing

Abstract

Using the database of the PuertoTerm project, an indexing system based on the cognitive model of Brigitte Enders was built. By analyzing the cognitive strategies of three abstractors, we built an automat that serves to simulate human indexing processes. The automat allows the texts integrated in the system to be assessed, evaluated and grouped by means of the bipartite spectral graph partitioning algorithm, which also permits visualization of the terms and the documents. The system features an ontology and a database to enhance its operativity. As a result of the application, we achieved better rates of exhaustivity in the indexing of documents, as well as greater precision and retrieval of information, with high levels of efficiency.