Impact of data duration on absolute error (data sets with walking noise).

Data sets with different signal duration were analysed using all the methods and the average absolute error is reported in the table. The absolute error is defined as the absolute value of the difference between calculated period and the expected value (24.08 for asym signal and 24 h for the others). Data sets were created by adding walking noise of 160% of the original signal amplitude to the hourly-sampled templates of different duration. 1) The base shape of the signal: cosine (cos), pulse (pul); double pulse (dpl); shoulder (shl) and moderate asymmetry (asym), (all) represents aggregated results from all the sets. See SI Table S4 for the full table.