Recent trends on HP-splines: from signals processing to Neural-networks for parameter estimation

Speaker: Rosanna Campagna, University of Campania "Luigi Vanvitelli"

Abstract: Hyperbolic-Polynomial Penalized splines (HP-splines), recently defined [1] to deal with data regression, are suitably designed for data that exhibit exponential trends, such as those encountered in signal processing. Multicomponent signals play a key role in many application fields, and they are well represented in the time-frequency plane where they are mainly characterized by ridge curves, which carry information about the instantaneous frequency of each signal component. The frequency parameter of HP-spline has been successfully used to predict the presence of time-frequency interference regions [2]. Recent studies concern a neural network approach for optimal and computationally efficient parameter selection.

[1] R. Campagna, C. Conti, Penalized hyperbolic-polynomial splines, Applied Mathematics Letters, 118, 2021.
[2] V. Bruni, R. Campagna, D. Vitulano, Multicomponent signals interference detection exploiting HP-splines frequency parameter, Applied Numerical Mathematics, vol. 209 (2025)

Biography: Associate Professor of Numerical Analysis, at University of Campania "Luigi Vanvitelli", Italy, her research interests are focused on: Approximation Theory, Kernel methods, inverse problems, and computational approaches for the identification and estimation of parameters of mathematical models for the applications.

Host: Prof. Michael Multerer