Forecasting the composition of Spain's unemployed population by genetic algorithms

Abstract

A genetic algorithm is developed to forecast the relative presence of different profiles in Spain’s unemployed population. A selection operator is defined that assumes that the higher the unemployment rate of a profile, the higher the probability that such a profile is present in future populations. A transition matrix takes other factors into account which may influence changes in the profiles. The algorithm is applied to the original quarterly populations of Spain's unemployed in 2014. Then, it is applied to obtain the forecast of the quarterly populations of unemployed in Spain in 2015 and 2016. This methodological proposal is shown to provide the type of forecast that is very useful in policy-making decisions to reduce the higher unemployment rates caused by the economic crisis in the euro area.