PAPERS
To return to the "Published Papers" main page, click here.
THE ELEVENTH INTERNATIONAL CONFERENCE ON FORENSIC COMPUTER SCIENCE AND CYBER LAW - ICoFCS 2019
Print ISBN 978-85-65069-15-1, pages 7-14
DOI: 10.5769/C2019001 and http://dx.doi.org/10.5769/C2019001
Descoberta de termos que caracterizam peças jurídicas
By Davi Alves Bezerra, Pedro H. G. Inazawa, Roberta Zumblik, Teófilo E. de Campos, Nilton Correia da Silva, Fabrício A. Braz, Fabiano Hartmann Peixoto.
To download this paper, click here.
Print ISBN 978-85-65069-15-1, pages 7-14
DOI: 10.5769/C2019001 and http://dx.doi.org/10.5769/C2019001
Descoberta de termos que caracterizam peças jurídicas
By Davi Alves Bezerra, Pedro H. G. Inazawa, Roberta Zumblik, Teófilo E. de Campos, Nilton Correia da Silva, Fabrício A. Braz, Fabiano Hartmann Peixoto.
To download this paper, click here.
ABSTRACT
The Brazilian Court System is currently the biggest judiciary system in the world, and has millions of scanned processes. However, it suffers from low efficiency problems due to low investment in new technologies. The usage of AI (Artificial Intelligence) is a suitable solution, since it can automate tasks that could, in recent past, only be done by humans. To facilitate the analysis of data, we propose to automatically discover the relevance of keywords. For that, we evaluate two classifiers (Naive Bayes and K-Nearest Neighbors) using Chi-squared statistical analysis and Term Frequency-Inverse Document Frequency count (TF-IDF) for feature ranking. Our experiments were performed on a dataset of 5 classes of judiciary pieces obtained from Brazil's Supreme Court (STF). Finally, a feature vector of the best ranked words is compared to the legal jargon of each selected piece.
KEYWORDS
Brazilian Court System, Artificial Intelligence, Naive Bayes, K-Nearest Neighbors, TF-IDF, Chi-squared.
The Brazilian Court System is currently the biggest judiciary system in the world, and has millions of scanned processes. However, it suffers from low efficiency problems due to low investment in new technologies. The usage of AI (Artificial Intelligence) is a suitable solution, since it can automate tasks that could, in recent past, only be done by humans. To facilitate the analysis of data, we propose to automatically discover the relevance of keywords. For that, we evaluate two classifiers (Naive Bayes and K-Nearest Neighbors) using Chi-squared statistical analysis and Term Frequency-Inverse Document Frequency count (TF-IDF) for feature ranking. Our experiments were performed on a dataset of 5 classes of judiciary pieces obtained from Brazil's Supreme Court (STF). Finally, a feature vector of the best ranked words is compared to the legal jargon of each selected piece.
KEYWORDS
Brazilian Court System, Artificial Intelligence, Naive Bayes, K-Nearest Neighbors, TF-IDF, Chi-squared.
