ANNOTATING UNSTRUCTURED TEXTS FOR ENHANCING SEMANTIC ANALYSIS PROCESSES

Abstract

Annotation is a powerful instrument for enhancing knowledge containing in texts. When developing a text analysis process, we often make notes for identifying and characterizing concepts and relationships, or highlighting aspects in the text that could go unnoticed by some of its readers. In addition, text annotation can enrich the semantics of texts, giving them more value through the introduction of comments, explanations, references, among many other things. Today, most text annotation processes are carried out helped by computational tools, whose functionalities make it possible to simplify the most elementary annotation tasks and substantially reduce the annotation time. The annotation of old, unstructured texts is very relevant for all those who want to study and acquire knowledge about their contents. Annotating these texts makes them more accessible to people who are not experts in the domain or in the era in which they were produced. In this work we develop a specific annotation system, supported by natural language processing and machine learning tools, to reveal the knowledge contained in the Book of Properties – “Tombo da Mitra” –, a codex containing the inventory of the Archbishop's Table of Braga’s properties (Portugal) in the 17th century. This codex contains a huge amount and a wide variety of elements, containing names, nicknames, settlements, professions, types of land and buildings, among many others. All these elements are very important for studying and learning of geography, culture, economy, architecture, religion and Portuguese language until the 17th century. Annotating the Book of Properties makes possible to maintain a tag database for indexing the most relevant information contained in the book and make its knowledge accessible to a wider range of people.