| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| DM_MarioLeite_MEI_2023 | 5.49 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Qualquer domínio técnico, seja o domínio jurídico, fiscal, político, social, é caracterizado por um discurso altamente especializado usando uma terminologia própria. Para além disso, o conhecimento encontra-se ainda pouco estruturado e disperso por vários recursos textuais. Estas características apresentam um desafio á representação digital do conhecimento, de modo a assegurar a recuperação inteligente de informação, adequada ao contexto.
Neste trabalho é proposto o desenvolvimento de um sistema de recuperação de informação tendo por base uma ontologia do domínio do direito fiscal desenvolvida numa dupla perspetiva, ou seja, capturando e representando a experiência dos especialistas do domínio, complementada com técnicas de processamento de linguagem natural.
Any technical field, be it legal, fiscal, political or social, is characterised by a highly specialised discourse using its own terminology. In addition, knowledge is still poorly structured and scattered across various textual resources. These characteristics present a challenge to the digital representation of knowledge, in order to ensure the intelligent retrieval of information, appropriate to the context. This work proposes the development of an information retrieval system based on an ontology of the tax law domain developed from a dual perspective, i.e. capturing and representing the experience of domain experts, complemented with natural language processing techniques. In addition, based on the ontology, machine learning principles were applied to the prediction and semantic annotation of Named Entities to add meta-information to legal documents. It is important that this ontology supports the incorporation of other existing vocabularies in order to create an intelligent information retrieval system through faceted searches with the possibility of consulting similar or relevant content for a given context.
Any technical field, be it legal, fiscal, political or social, is characterised by a highly specialised discourse using its own terminology. In addition, knowledge is still poorly structured and scattered across various textual resources. These characteristics present a challenge to the digital representation of knowledge, in order to ensure the intelligent retrieval of information, appropriate to the context. This work proposes the development of an information retrieval system based on an ontology of the tax law domain developed from a dual perspective, i.e. capturing and representing the experience of domain experts, complemented with natural language processing techniques. In addition, based on the ontology, machine learning principles were applied to the prediction and semantic annotation of Named Entities to add meta-information to legal documents. It is important that this ontology supports the incorporation of other existing vocabularies in order to create an intelligent information retrieval system through faceted searches with the possibility of consulting similar or relevant content for a given context.
Description
Keywords
Domínio Jurídico Inteligência artificial Processamento de Linguagem Natural Sistema de Recuperação de Informação