ActiSleep - Sleep Duration Estimation from Accelerometer Data

Estimates daily sleep duration from wrist or hip accelerometer data using a Pruned Dynamic Programming (PDP) changepoint algorithm. PDP partitions a minute-level activity series into segments, then identifies transitions between sleep and wakefulness by detecting extended periods of low physical activity. Five cost-function models (Poisson, Normal, Negative binomial, Variance, Exponential) accommodate different activity-count distributions. Optional non-wear detection and sleep diary integration are supported. Reference: Baek, Jonggyu, Banker, Margaret, Jansen, Erica C., She, Xichen, Peterson, Karen E., Pitchford, E. Andrew, Song, Peter X. K. (2021) An Efficient Segmentation Algorithm to Estimate Sleep Duration from Actigraphy Data <doi:10.1007/s12561-021-09309-3>.