tlpipe.rfi.combinatorial_threshold.CombinatorialThreshold

class tlpipe.rfi.combinatorial_threshold.CombinatorialThreshold(time_freq_vis, time_freq_vis_mask=None, first_threshold=6.0, exp_factor=1.5, distribution='Rayleigh', max_threshold_length=1024)

Abstract base class for combinatorial thresholding methods.

The method will flag a combination of samples when a property of this combination exceeds some limit. The more connected samples are combined, the lower the sample threshold.

For more details, see Offringa et al., 2000, MNRAS, 405, 155, Post-correlation radio frequency interference classification methods.

For this implementation, the sequence of thresholds are determined by the following formula:

\[\alpha \times \frac{\rho^{i}}{w} \times (\sigma \times \beta) + \eta\]

in which \(\alpha\) is the first threshold set to 6.0, \(\rho = 1.5\), \(i\) is the current iteration, \(w\) the current window size, \(\sigma\) the standard deviation of the values, \(\beta\) the base sensitivity, set to 1.0, and \(\eta\) the median.

__init__(time_freq_vis, time_freq_vis_mask=None, first_threshold=6.0, exp_factor=1.5, distribution='Rayleigh', max_threshold_length=1024)

x.__init__(…) initializes x; see help(type(x)) for signature

Methods

execute([sensitivity, direction])	Execute the thresholding method.
execute_threshold(factor, direction)	Abstract method that needs to be implemented by sub-classes.
init_threshold_with_flase_rate(resolution, …)