Outliers detection in multivariate time series using genetic algorithms

• Outlier detection essential step analyzing chemical data.
• We proposed a genetic algorithm for detecting additive outliers.
• Able to detect any number of potential outliers in multivariate time series.
• Simultaneously reduce possible masking and swamping effects.
• Good results in simulation and empirical data for patches of additive outliers.

A genetic algorithm to detect multiple additive outliers in multivariate time series is proposed. In contrast with many of the existing methods, it does not require to specify a vector ARMA model for the data and is able to detect any number of potential outliers simultaneously reducing possible masking and swamping effects. A generalized AIC-like criterion is used as objective function. The comparison and the performance of the proposed method are illustrated by simulation studies and real data analysis. Simulation results show that the proposed approach is able to handle patches of additive outliers.