Would it be possible to include a "Relative Amplitude" metric please?
A function to compute Relative Amplitude (RA) for a time series (e.g., activity or light). RA quantifies the day-night amplitude by comparing average level in the brightest contiguous X hours to the darkest contiguous Y hours. Typical default for activity for example - X = 10 hours and Y = 5 hours (M10 / L5 for activity; B10 / D5 for light). Higher RA = stronger day/night contrast (theoretical range 0–1).
Definition & formula:
Let M10_mean be the mean value of the time series over the most active contiguous X hours (default X = 10).
Let L5_mean be the mean value over the least active contiguous Y hours (default Y = 5).
Amplitude (AMP) = M10_mean - L5_mean.
Relative Amplitude (RA) = AMP / (M10_mean + L5_mean) = (M10_mean - L5_mean) / (M10_mean + L5_mean).
RA is bounded in theory between 0 and 1 when values are non-negative and M10_mean >= L5_mean.
Many thanks in advance,
Emily
Would it be possible to include a "Relative Amplitude" metric please?
A function to compute Relative Amplitude (RA) for a time series (e.g., activity or light). RA quantifies the day-night amplitude by comparing average level in the brightest contiguous X hours to the darkest contiguous Y hours. Typical default for activity for example - X = 10 hours and Y = 5 hours (M10 / L5 for activity; B10 / D5 for light). Higher RA = stronger day/night contrast (theoretical range 0–1).
Definition & formula:
Let M10_mean be the mean value of the time series over the most active contiguous X hours (default X = 10).
Let L5_mean be the mean value over the least active contiguous Y hours (default Y = 5).
Amplitude (AMP) = M10_mean - L5_mean.
Relative Amplitude (RA) = AMP / (M10_mean + L5_mean) = (M10_mean - L5_mean) / (M10_mean + L5_mean).
RA is bounded in theory between 0 and 1 when values are non-negative and M10_mean >= L5_mean.
Many thanks in advance,
Emily