Correlation-estimated conditional average method and its application on solitary oscillation in PANTA.

The conditional average is a technique to extract a typical waveform from bursty or quasi-periodic phenomena regarded as a combination of deterministic trend and residual fluctuations (or probabilistic part). The article proposes a new conditional average technique that is an extension from what is called the template method that allows automatic selection, using the correlation analysis, of the identical events of the phenomena. The proposed method is applied on a quasi-periodic oscillation observed in a linear magnetized plasma, and successfully divides the phenomenon into the deterministic trend and residual fluctuations. Moreover, the statistical error analysis on the power of the residual fluctuations discloses the presence of mutual interactions between the deterministic trend and the residual fluctuations. [ABSTRACT FROM AUTHOR]