Banca de DEFESA: WASHINGTON PINHEIRO DE SOUZA

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
DISCENTE : WASHINGTON PINHEIRO DE SOUZA
DATA : 30/08/2018
HORA: 14:00
LOCAL: Auditório 101 do Prédio 17 (DECOM)
TÍTULO:

..


PALAVRAS-CHAVES:

.


PÁGINAS: 83
GRANDE ÁREA: Outra
ÁREA: Multidisciplinar
RESUMO:

.In several segments that require a prior budget and / or a revenue forecast, estimates are common. This procedure is based on statistical models. The Binomial distribution is a discrete statistical model that determines the chances of success or failure of the occurrence of a certain number of events in a limited sample space. The normal distribution is the standard statistical model that follows the law of large numbers, that is, random occurrences that tend to infinity converge to an average and present a variance. Gathering the characteristics of the two distributions in a single tool is of great value to make estimates and, for that, the convolution is used. In order to perform the convolution between the two distributions, it is necessary that both are discrete or continuous. The Binomial distribution was approximated to a continuous function and subsequently convolved with the normal distribution, while the normal distribution was approximated from a discrete series and was subsequently convolved with the Binomial distribution. It is made the forecast of the quantity of defective pieces that must be collected by Magneti Marelli throughout Brazil through the proportionality relation of the discrete convolution. The estimated values include the actual values disclosed by the company. At the UFMG Veterinary Hospital, estimates were made of the number of animals that were able to perform castration surgery in 2016, through discrete convolution and using data from the year 2015. All monthly values collected at the veterinary hospital in the year of 2016 are within the forecast made using the model. The National Broad Consumer Price Index (IPCA) for the years 2011 to 2016 was estimated based on data from 1996 to 2010 released by IBGE, using continuous convolution. The observed values are within the range predicted by the method with proximity to the mean. Due to the application in the several sectors mentioned with satisfactory results, the proposed method proved effective and multidisciplinary, that is, it can be applied in several areas.


MEMBROS DA BANCA:
Interno - ARTHUR RODRIGO BOSCO DE MAGALHAES
Interno - FABIO ROCHA DA SILVA
Externo ao Programa - LUIS ARGEL POVEDA CALVINO
Presidente - THIAGO GOMES DE MATTOS
Notícia cadastrada em: 29/08/2018 11:09
SIGAA | Diretoria de Tecnologia da Informação - DTI - (31) 3319-7000 | Copyright © 2006-2024 - UFRN - vm-sig-app-06.ditic.sgi.cefetmg.br.inst6