Theme：Time-series representations of polynomials leading to better understanding and accurate estimations

Reporter：Professor Asoke K. Nandi FREng

Time：15:30-16:30, 16:40-17:40, May 9^th, 2025

Venue：Building 2, room 5008, iHarbour

Abstract：

Data modelling has a long history, easily going back to 19th century when Gauss and Legendre invented the least-squares method for polynomials. Time-series representation followed much later. Data modelling is a fast-growing topic in 21st century. This presentation (in two parts) will offer novel, insightful, and recent results. For example, all polynomials of degree q can be represented by a time-series of the same order with a constant. From noisy data, the degree of a polynomial, the additive noise as well as the polynomial coefficient corresponding to the highest polynomial degree can be estimated more accurately than the current methods, without using a polynomial model. This new formulation offers better estimation.