pyhctsa.operations.symbolic.binary_stats¶

pyhctsa.operations.symbolic.binary_stats(y, binary_method='diff')¶

Compute statistics on a binary symbolisation of the input time series.

The time series is first symbolized as a binary string of 0s and 1s using a specified coarse-graining (symbolisation) method. Then, various statistics are computed to characterize the structure of the resulting binary sequence.

Parameters:¶

y : array-like¶

The input time series.

binary_method : str, optional¶

The binary symbolisation rule. One of:

• ’diff’: Encode as 1 if the time-series difference is positive, and 0 otherwise.

• ’mean’: Encode as 1 if the value is above the mean, 0 otherwise.

• ’median’: Encode as 1 if the value is above the median, 0 otherwise.

• ’iqr’: Encode as 1 if the value is within the interquartile range (above the 25th percentile and below the 75th), 0 otherwise.

Default is 'diff'.

Returns:¶

Statistics computed on the binary symbolisation. The standard deviations of the stretch lengths are NaN if there are no stretches of that symbol, and 0 if there is a single stretch.

Return type:¶

dict