Multipel regressionsanalys av variabler som paverkar BNP
In this report a model was constructed in order to determine how a number of covariates influence the gross domestic product, GDP. The covariates were chosen depending on their expected influence on GDP, for example education and life expectancy. The data used in this report are collected from the World Bank. The model to describe GDP has been calculated using multiple line arregression. In order to reach a reliable final model the number of covariates has been gradually decreased to eliminate insignificant covariates. In order to minimize the error term and find a reliable model the Baysian Information Criterion has been used together with hypothesis testing. At a 95% confidence interval the final model could predict 111 of 139 countries GDP. The influences of the covariates in the final model is well in line with the expectations. For instance a positive relationship between GDP and education is observed.