You are in the accessibility menu

Please use this identifier to cite or link to this item: http://acervodigital.unesp.br/handle/10400.2/6780
Title: 
Estatística multivariada na análise aos inquéritos anuais às empresas em Cabo Verde
Author(s): 
Mendes, José Manuel Alves
Abstract: 
  • Este estudo teve como objetivo agrupar as empresas em grupos homogéneos, determinar a dependência entre variáveis, conhecer o comportamento dos dados, analisar e interpretar a matriz de proximidade entre as variáveis de estratificação e/ou as variáveis importantes do inquérito, analisar e identificar os principais fatores relacionados com o emprego e a faturação de diferentes tipos de empresas e provar finalmente se tem sentido ou não aplicar a estatística multivariada na análise aos inquéritos anuais às empresas, nomeadamente a Análise de Clusters (AC). Para esse trabalho de dissertação, utilizou-se como dados de suporte, os dados definitivos do inquérito anual às empresas do ano económico de 2014 disponibilizado pelo Instituto Nacional de Estatística de Cabo Verde. Numa primeira fase procedeu-se a um breve enquadramento sobre as estatísticas empresariais em Cabo Verde, destacando os objetivos desse inquérito, o seu historial e os principais conceitos. Numa segunda fase, procedeu-se a uma abordagem teórica sobre a Amostragem Aleatória Estratificada (AAE), as Características Amostrais e a AC, onde se destacou os respetivos objetivos, os principais conceitos, as características, as fórmulas de cálculo, as condições de aplicabilidade, entre outros aspetos. Na fase final, procedeu-se, primeiramente, a uma análise exploratória dos dados, onde se calculou as principais características amostrais (univariadas e bivariadas) das variáveis emprego e faturação e, em seguida, se fez a aplicação da AC aos dados do inquérito, onde se iniciou com uma análise descritiva dos dados para se conhecer o comportamento dos dados e depois aplicar os dois métodos de AC, a saber: o método hierárquico e o método não hierárquico. Concluiu-se que as variáveis de estratificação utilizadas constituem fatores relacionados com o Emprego e a Faturação de diferentes tipos de empresas, mas que ao em vez de estratificar por escalões de número de pessoas ao serviço, seria melhor estratificar por categorias de empresas e também que os escalões de trabalhadores 1 - 5, 6 - 12 e 13 - 20 são os mais apropriados.
  • This study aimed to group the companies into homogeneous groups, determine the dependency between variables, to understand the behavior of the data, analyze and interpret the proximity matrix between the stratification variables and / or important variables of the survey, analyze and identify key factors related to the use and billing of different types of companies. And to determine if multivariate statistical analysis is a methodology that can help interpret the annual business surveys. For this dissertation work, used as supporting data the final data from the annual survey of companies in the financial year 2014, provided by the National Institute of Statistics of Cape Verde. This study did a brief background analysis and research on the business statistics of Cape Verde, highlighting the objectives of the survey, its history and the main concepts. In a second step, a theoretical approach of the Random Sampling Stratified, the Sampling Features and AC, is presented. Where it highlighted the objectives, key concepts, features, calculation formulas, the applicability conditions, among others aspects. In the final phase is presented, an exploratory analysis of the data, including the calculation of the main sample characteristics (univariate and bivariate) of the variables employment and turnover, and the application of AC to survey data, which began with a descriptive analysis of data to understand the behavior of the data and then apply both AC methods, namely the method hierarchical and non-hierarchical method. This study concluded that the stratification variables are one of the main factors related to the use and billing of different types of companies. However, to instead of stratifying by categories of number of employees, it would be better to stratify by categories of companies and the ranks of workers 1 - 5, 6 - 12 and 13 - 20.
Issue Date: 
3-Oct-2017
Citation: 
Mendes, José Manuel Alves - Estatística multivariada na análise aos inquéritos anuais às empresas em Cabo Verde [Em linha]. Lisboa: [s.n.], 2017. 118 p.
Keywords: 
  • Análise de clusters
  • Análise exploratória de dados
  • Estatística
  • Empresas
  • Análise estatística multivariada
  • Inquéritos
  • Cabo Verde
  • Exploratory data analysis
  • Cluster analysis
  • Correlation analysis
  • Sampling techniques and multivariate statistical analysis
URI: 
Rights: 
embargoedAccess
Type: 
outro
Source:
http://repositorioaberto.uab.pt/handle/10400.2/6780
Appears in Collections:Dissertações de Mestrado - Universidade Aberta de Portugal

There are no files associated with this item.
 

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.