“…The H‐exponent is a self‐similar parameter that determines the long‐term correlation in the recorded time series. [ 8,10,12,33 ] Generally, 0.5 < H < 1 indicates the persistency of the signal, H = 0.5 indicates an uncorrelated process, while 0 < H < 0.5 indicates the anti‐persistence of the data series. To examine how far the data set reflects long‐term correlation, the whole signal is divided into non‐overlapping bins and the H‐coefficients are obtained for each bin, shown in Figure 9a for V A ˜ 340 V. The R/S analysis provides H = 0.61 for the time lags from 12 to 300 while H = 0.83 for time lags from 300 to 1100, which indicates the existence of long‐term memory in the time series, and the trend persists in the next blocks of the data set.…”