Book chapter · Routledge eBooks · April 22, 2025

Logical-linguistic model of fact extraction from a text corpus

Orken J. Mamyrbayev, Nina Khairova, Waldemar Wójcik

Abstract

The extraction and representation of knowledge from semi-structured textual sources involve identifying essential elements of a subject area and their relationships. In crime-related information, facts are represented as structured triples (subject-attribute-value or subject-relation-object) that link participants and their interactions. These facts can be simple, involving subjects and actions (verbs), or more complex, incorporating additional attributes such as time, place, or instrument. This chapter demonstrates how, based on grammatical structures in Kazakh, Russian, and English, four types of structured facts can be identified: subj-fact, obj-fact, subj-obj fact, and complex fact, each reflecting different aspects of a crime-related event. This structured approach allows for the extraction of relevant concepts and relationships, facilitating their representation in a structured form for further analysis.

Advanced Text Analysis TechniquesNatural Language Processing TechniquesTopic Modelingfact extractionknowledge extractionsemi-structured text

Logical-linguistic model of fact extraction from a text corpus

Abstract

More by Orken Mamyrbayev