Submitted by Automa??o e Estat?stica (sst@bczm.ufrn.br) on 2015-12-15T17:51:47Z No. of bitstreams: 1 BrunoDeLimaBarbosa_DISSERT.pdf: 1669877 bytes, checksum: 0d2845573a1cf257fdfb5c37b2f0380d (MD5) / Approved for entry into archive by Arlan Eloi Leite Silva (eloihistoriador@yahoo.com.br) on 2015-12-29T21:00:58Z (GMT) No. of bitstreams: 1 BrunoDeLimaBarbosa_DISSERT.pdf: 1669877 bytes, checksum: 0d2845573a1cf257fdfb5c37b2f0380d (MD5) / Made available in DSpace on 2015-12-29T21:00:58Z (GMT). No. of bitstreams: 1 BrunoDeLimaBarbosa_DISSERT.pdf: 1669877 bytes, checksum: 0d2845573a1cf257fdfb5c37b2f0380d (MD5) Previous issue date: 2014-09-04 / Neste trabalho, apresentamos nosso entendimento acerca do artigo de Aksoy (2003), executando uma aplica??o do modelo por ele apresentado, com o intuito de gerar dados de fluxos intermitentes, a partir de bancos de dados brasileiros. Em seguida, constru?mos um modelo oculto de Markov como uma nova proposta de abordagem para o problema. Nele utilizamos distribui??es gama para simular os aumentos e decr?scimos dos fluxos dos rios. / In this work, we present our understanding about the article of Aksoy [1], which uses Markov chains to model the flow of intermittent rivers. Then, we executed an application of his model in order to generate data for intermittent streamflows, based on a data set of Brazilian streams. After that, we build a hidden Markov model as a proposed new approach to the problem of simulation of such flows. We used the Gamma distribution to simulate the increases and decreases in river flows, along with a two-state Markov chain. The motivation for us to use a hidden Markov model comes from the possibility of obtaining the same information that the Aksoy?s model provides, but using a single tool capable of treating the problem as a whole, and not through multiple independent processes