Κριτική της μεθοδολογίας ανάλυσης των ατυχημάτων στην παγκόσμιο ποντοπόρο ναυτιλία: προτάσεις - βελτιώσεις

Abstract

The problem of testing a set of data for outliers is not new in statistics. Methods have been proposed for the general linear model. These methods are not appropriate for detecting outliers in time series data. However, it seems to be particularly important to be able to detect outliers in time series, especially if these data are to be used for forecasting purposes. In this paper certain tests for the presence of outliers are described. These tests are then applied, using simulation methods, to several ARIMA models, and results are compared. An illustration of the procedure is given with published U. K. economic data.