Event Extraction and Representation: A Case Study for the Portuguese Language

Abstract

Text information extraction is an important natural language processing (NLP) task, which aims to automatically identify, extract, and represent information from text. In this context, event extraction plays a relevant role, allowing actions, agents, objects, places, and time periods to be identified and represented. The extracted information can be represented by specialized ontologies, supporting knowledge-based reasoning and inference processes. In this work, we will describe, in detail, our proposal for event extraction from Portuguese documents. The proposed approach is based on a pipeline of specialized natural language processing tools; namely, a part-of-speech tagger, a named entities recognizer, a dependency parser, semantic role labeling, and a knowledge extraction module. The architecture is language-independent, but its modules are language-dependent and can be built using adequate AI (i.e., rule-based or machine learning) methodologies. The developed system was evaluated with a corpus of Portuguese texts and the obtained results are presented and analysed. The current limitations and future work are discussed in detail.In this work, we will present and describe a proposal for event extraction from Portuguese texts, based on a pipeline of specialized natural language processing tools; namely, a part-of-speech tagger, a named entities recognizer, a dependency parser, semantic role labeling, and a knowledge extraction module. This architecture was designed to be language-independent but its modules are language-dependent, in the sense that they depend on specialized rules or that their models need to be created using machine learning approaches, requiring previously annotated Portuguese corpora.The proposed system was evaluated with two Portuguese corpora, one being the publicly available corpus of PropBank [7], and the obtained results are presented and discussed. Due to the complexity of the task, there still exist many limitations and problems that need to be solved, but we believe this architecture can play an important tool in this domain and, in particular, in the context of the computational processing of the Portuguese Language. Moreover, this work is strongly related to the participation of the authors in the Portugal2020 Agatha project [8]. Basically, the aim of this project is to intelligently analyze open-source information for surveillance/crime control, following in the footsteps of similar open source information analysis, where author profiling [9], aggression identification [10] and hate-speech detection [11] over social media, as well as statute law retrieval and entailment for Japanese statutes [12] have already been done.The remainder of this paper is organized as follows: In Section 2, we present an overview of the related work. Section 3 describes our proposed architecture and Section 4 presents the Portuguese modules for its computational processing. Finally, Section 5 evaluates the proposal, Section 6 discusses different design options, and, in Section 7, we provide our conclusio...

Editorial for the Special Issue on “Natural Language Processing and Text Mining”

“…The paper "Event Extraction and Representation: A Case Study for the Portuguese Language" [6] describes an event extraction system from Portuguese documents which is based on a pipeline of specialized natural language processing tools: namely part-of-speech tagging, dependency parsing, named entity recognition, semantic role labeling, and a knowledge extraction module. The developed system is evaluated with a corpus of Portuguese texts and compared with the existing tool LinguaKit [7].…”