BDBComp
Parceria:
SBC
Identifying Marijuana Users Using Text Mining Techniques

Euclides C. Pinto NetoDiogenes CarvalhoGeorge CabralFilipe RolimRafael FerreiraGustavo Callou

Nowadays, illegal drug dealing has became a huge social problem inall countries. Such drugs are highly addictive and the results of using them are,for example, bad life quality and impacts on public health, as many diseasescan be developed. Even family wellness can be impacted by having an addicteduser at home. Among these illegal drugs, the most popular is the marijuana. Inthis paper we propose an approach to combat the marijuana sale by identifyingones who are related to drug dealers, i.e., ones who use the drug. To accomplishthat, we use an strategy, named term frequency-inverse document frequency (TFIDF),and a twitter based dataset. As classification metrics, we consider theaccuracy and Matthews correlation coefficient (MCC), in which the achievedresults are 78.97% and 0.56, respectively.

http://www.lbd.dcc.ufmg.br/colecoes/eniac/2016/061.pdf

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato: bdbcomp@lbd.dcc.ufmg.br
     Mantida por:
LBD