Bottom_kurtosis

Bottom_kurtosis (unitless) is a bottom feature that may be extracted during bottom classification. A Bottom_kurtosis analysis variable can be selected under Bottom on the Export page of the EV File Properties dialog box. It can be included with onscreen analyses and exported. The kurtosis is descriptor for the outliers of the sample distribution for the bottom echo.

Bottom_kurtosis represents the mean of the kurtosis of the first bottom echoes in the Feature extraction interval for a bottom classification.

where:

j	=	Bottom point in a bottom points variable. Bottom point number j is assigned sequentially in time. The pings of the echogram are partitioned according to the number of pings in a specified Feature extraction interval.
i	=	Ping i in the Feature extraction interval j.
k	=	Sample k of the first bottom echo in Ping i in the Feature extraction interval j.
n	=	Number of pings in Feature extraction interval j.
m	=	Number of samples in the first bottom echo of ping i.
Ek	=	the linear Sv of sample k(m2/m3) -  set to zero for any sample where Ek< mSv or Ek> MSv, where: mSvis the minimum integration threshold (dB re 1 m-1) MSvis the maximum integration threshold (dB re 1 m-1)
EEchoMean	=	the Mean Sv of the samples, in the first bottom echo, in linear units (m2/m3) = 10Sv/10
ESDi, j	=	Standard deviation of the Sv (linear units) of sample k in ping i of Feature extraction interval j

See also

About bottom classification
Bottom classification algorithms
EV File Properties dialog box