Auto Law: Web Solution for Registration and Support for Guidance On Rights of Victims of Acts of Violence

Proceedings of ‏The 3rd International Conference on Research in Science, Engineering and Technology

Year: 2021


[Fulltext PDF]

Auto Law: Web Solution for Registration and Support for Guidance On Rights of Victims of Acts of Violence

Fabricio Luiz de Souza Pereira, João Pedro Amaral de Oliveira, Luan Alves Godoy, Davi Pedro dos Santos, Victor Araújo Sguisato, Willian dos Santos Miranda, Luiz Melk de Carvalho, Flávio Henrique Batista de Souza



Unfortunately, society is marked by issues of intolerance and aggression. Many victims of aggression often do not know whom to appeal to, what are their rights and how to obtain support and help. Recent movements, such as the “Black Lives Matter”, emphasize the need to pay attention to this sort of situation. Thus, the proposed web solution of this work, called Autolaw tool was developed, based on technologies such as: Cloud Computing, Natural Language Processing (NLP) and Machine Learning. This tool was developed according to the following methodology: a cloud structure was developed to gather and process the testimony of the user/victim; a set of algorithms was implemented in the back-end of the tool platform to process the demand via NLP and a machine learning process was used to classify the type of legal protection that is needed focusing according to the Brazilian laws applicable to violence against women and racism; and finally, an accuracy assessment via AUC (Area Under the Curve) ranging from 0 to 1 was performed. As a result, a functional web structure was created and accuracy of the Machine Learning algorithms was assessed. In summary, the web tool, represented by the interface via a website, gathers the testimony of the victim testimony; evaluates, via NLP, the present terms; it recognizes the similarity with the terms of protection laws for the elderly, women and racial ethnicity using machine learning algorithms; and returns the protection laws to the user. Experiments were carried out to gather consultations (700 samples) and the accuracy of the tests were evaluated considering the analysis of the resampling process. As a result, an AUC between 0.81 and 0.92 for the recommendations were obtained.

Keywords: women rights; machine learning; natural language processing; black lives matter; racism.