Encaminhamento de tickets que exigem formação de grupos por classificação de texto

• Ano de defesa: 2019
• Autor(a) principal: Augusto, Jones Marques
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Rio de Janeiro; Brasil; Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa de Engenharia; Programa de Pós-Graduação em Engenharia de Sistemas e Computação; UFRJ
• Palavras-chave em Português: Text classification; Ticket routing; CNPQ::ENGENHARIAS
• Link de acesso: http://hdl.handle.net/11422/14046

[EN] With the growth of Information Technology (IT) the use of IT services is becoming more important for the world economy, the commercialization of these services has become strategic for many companies and organizations. These providers are challenged to keep the technology infractutures up and running as they need to ensure that their services are always available to keep their customers happy. With the easy access thanks to the Internet, users usually have several options to perform the desired task and can migrate if the service does not meet them. Since the world has become dependent on IT systems, the poor quality or unavailability of these services is not good for business. Thus, resolving the incidents that cause most of these problems is one of the important points of IT service management. Seeking to reduce the workload involved in incident management, a system that routes recorded incidents to areas that have the ability to address them is created and tested using actual incidents collected from a Brazilian IT service provider. To do so, text classification techniques are employed using two different approaches. The first addresses incidents that are solved by only one expert group, and the second approach works with incidents that need more than one expert group to solve. The proposed experiments show good results in workload reduction and classification accuracy.