Análise de correspondência aplicada: perfil dos países africanos segundo os objetivos do milênio

Abstract

This work uses a database related to the Millennium Development Goals, a program developed by the UN, which is divided into eight targets data. Each goal has a set of indicators covering areas such as health, education, environment and economy. The motivation to investigate only African countries came from the historical problems that Africa presents in the main goals set by the UN , as in health (examples are malnutrition and malaria), environment ( desertification ) and with political conflicts. The first six goals were chosen for analysis and two common problems were identified in the database: ( i ) missing data and ( ii ) variables with different measurement units and types. To solve these two problems, we adopted primarily a form of multiple imputation via regression model to a variable that had missing values, using as auxiliary variables its goal companions. After completion of this stage, it was necessary to find a way to standardize the data. The standardization adopted, each country presented a value between 0 and 1, which shows the steps taken as a proportion of the whole way to go for its indicator. After these adjustments, the Simple Correspondence Analysis technique is used to analyze the total variability present in the database, and graphing the association between variables and countries. The software used for multiple imputation and the Simple Correspondence Analysis was SAS.