thresholds

Documentation

Index

• Constants
• func HighThresholdFunctional(normalized_magnitude float64, sample_size int) (float64, error)
• func HighThresholdPerformance(normalized_magnitude float64, sample_size int) (float64, error)

Constants

const LowThreshold = 0.05

The low threshold used in hypothesis testing. Note that this implies we could be rejecting the null hypothesis if the p-value is less than this value. It can happen that because we're generating the table of "high" thresholds with randomised permutations, that this value is actually higher than the "high" thresholds.

Functions

func HighThresholdFunctional

func HighThresholdFunctional(normalized_magnitude float64, sample_size int) (float64, error)

HighThresholdFunctional returns the high threshold for functional hypothesis testing given the normalized_magnitude and the sample_size.

The normalized magnitude is an estimate of failure rate between 0 and 1. We need more values to find smaller differences.

func HighThresholdPerformance

func HighThresholdPerformance(normalized_magnitude float64, sample_size int) (float64, error)

HighThresholdPerformance returns the high threshold for performance hypothesis testing given the normalized_magnitude and the sample_size.

The normalized magnitude is an estimate of the size of differences to look for, normalized by the interquartile range (IQR). We need more values to find smaller differences.