In the recent years, there has been a growing interest in identifying anomalous structure within multivariate data sequences. We consider problem of detecting collective anomalies, corresponding to intervals where one, or more, sequences behaves anomalously. first develop test for single anomaly that power simultaneously detect anomalies are either rare, is affecting few sequences, common. then...

Abstract It is quite common that the structure of a time series changes abruptly. Identifying these change points and describing model in segments between an important task financial analysis. Change point detection identification abrupt generative parameters sequential data. In application areas such as finance, online rather than offline mostly required, due to their use predictive tasks, pos...

To improve the detection efficiency of a long-distance dim point target based on dynamic programming (DP), this paper proposes multi-frame algorithm merit function filtering DP ring (MFF-DPR). First, to reduce influence noise pixel state estimation results, second-order named MFF-DP is proposed. The current states pixels an image plane are estimated by maximizing addition functions previous two...