Propagación de los errores de proyección de las series de tiempo ajustadas con modelos de espacio de estado

One of the most important problems in the analysis of the series of time is the projection of future values. But at any moment one is due to consider that cannot be predicted the future, is only possible to be analyzed the past, to be hoped that the model persists and on the basis of it to make the estimations of future values by means of a projection towards in front of the generating model of the observations. When the series represents some economic variable, the practical importance of the projection is evident. The objective in the present work is to study the errors that take place in the process of estimation and prediction, explaining the form in which they propagate, by means of a simple mathematical expression, and how they influence when a series is the accumulated ones of others. For statistical the modeled analysis and of the treated series of time, the approach of state space is applied, which, as it were demonstrated in diverse published works, is very satisfactory at the time of making projections and in addition very useful in the application to diverse real problems.