An implementation of CAPA (Collective And Point Anomaly) for the detection of anomalies in time series data. The package also contains Kepler lightcurve data and shows how CAPA can be applied to detect exoplanets.
Fast anomaly detection in R
This R package implements CAPA (Collective And Point Anomalies) introduced by Fisch, Eckley and Fearnhead (2018). The package is available on CRAN and contains lightcurve data from the Kepler telescope to illustrate the algorithm.
CAPA detects and distinguishes between collective and point anomalies. The algorithm's runtime scales linearly at best and quadratically at worst in the number of datapoints. It is coded in C and can process 10000 datapoints almost instantly.