Application of Artificial Intelligence in Jurisprudence Search Engine

Resumo

Search for jurisprudence is an arduous task that requires hours of commitment on the part of the legal
professional. Currently, there are several software that help in the search for the ideal document that the lawyer
needs to corroborate his processes. However, such systems use search for key words, that is, from a phrase or set
of words typed by the user, the software performs research in each document of the databases of jurisprudence.
This is a slow process and, by using keywords, the documents resulting from the search are mostly irrelevant with
the intention of the lawyer. Thus, this project aims to implement a web software to perform search in databases of
jurisprudence using artificial intelligence in its search engine. The methodology applied consists of obtaining the
database with all published jurisprudence and training an artificial intelligence algorithm with all the content of
the documents. After training the model, it will be inserted into a web software that will receive the lawyer's
research. Currently, an artificial intelligence model was generated with 420 documents and the results were
satisfactory. In the tests, the model reached 95% accuracy.