Anomaly Detection and Removal Using Non-Stationary Gaussian Processes
This addresses the problem of cleaning faulty time-series data for applications like sensor monitoring and EEG analysis, representing an incremental improvement with a new kernel for existing Gaussian process methods.
The paper tackles the problem of fault removal in time-series data by proposing a novel Gaussian process approach that models both the physical phenomenon and faults separately, using a new Markov Region Link kernel to handle non-stationarity while ensuring continuity. The method is applied to remove drift and bias errors in sensor data and recover EOG artifacts in EEG signals, though no concrete performance numbers are provided.
This paper proposes a novel Gaussian process approach to fault removal in time-series data. Fault removal does not delete the faulty signal data but, instead, massages the fault from the data. We assume that only one fault occurs at any one time and model the signal by two separate non-parametric Gaussian process models for both the physical phenomenon and the fault. In order to facilitate fault removal we introduce the Markov Region Link kernel for handling non-stationary Gaussian processes. This kernel is piece-wise stationary but guarantees that functions generated by it and their derivatives (when required) are everywhere continuous. We apply this kernel to the removal of drift and bias errors in faulty sensor data and also to the recovery of EOG artifact corrupted EEG signals.