Speaker: Unnati Nigam
Day, Date and Time: Wednesday, 25th September, 11.30 am
Online mode: join at https://meet.google.com/viw-jwsy-zdr
Title: Modelling of quasi-periodic data
Abstract: Quasi (pseudo/approximate) periodic signals often occur in natural settings, particularly when a periodic signal is recorded with noise. In this seminar, we will present a new dynamical equation system to construct a family of Quasi-Periodic Gaussian Processes (QPGP). We will describe a computationally inexpensive algorithm for the maximum likelihood estimation of parameters based on dynamic equations. This approach also simplifies the signal forecasting. We will illustrate via a simulation study that the proposed QPGP estimation strategy is faster as well and more accurate than existing constructions. Unlike these existing models, the proposed approach extends to multiple families of kernels, which we illustrate by modeling sunspot data, carbon dioxide emissions data and ECG data with both periodic Mat\'ern and MacKay's covariance kernel. This shows the exclusive advantage of the new QPGP family proposed in this work.