NEW PROPOSALS IN MULTIVARIATE OUTLIERS IDENTIFICATION

ABSTRACT: Occurrences of outliers in multivariate time series are unpredictable events which may severely distort the analysis of the series. It may be noticed that a convenient way for representing multiple outliers consists in superimposing a deterministic disturbance to a Gaussian multivariate time series. Then outliers may be modelled as non – Gaussian time series components. The independent component analysis is a recently developed tool that is likely to be able to extract possible outlier patterns. In practice the independent component analysis may be used to analyze multivariate observable time series and separate regular and outlying unobservable components. In the factor models framework too, independent component analysis turns out to be a useful tool for outliers detection in multivariate time series