DTA

Corso Ordinario Secondo Livello

FAN, YIJIANG

etd-10132024-190353

Forecasting curves portions using motif discovery inspired method

Classe Scienze Sociali

SCIENZE ECONOMICHE E MANAGERIALI - SCIENZE ECONOMICHE E MANAGERIALI

Tutor Prof. DI MININ, ALBERTO
Relatore Prof.ssa CHIAROMONTE, FRANCESCA
Presidente Prof. IRALDO, FABIO
Membro Dott.ssa CANTARELLI, PAOLA
Membro Prof. BOTTAZZI, GIULIO
Membro Prof. ROVENTINI, ANDREA
Membro Prof. TURCHETTI, GIUSEPPE
Membro Prof. MONETA, ALESSIO
Membro Prof. TENUCCI, ANDREA
Membro Prof. CINQUINI, LINO
Membro Dott. GIACHINI, DANIELE
Membro Prof.ssa VAINIERI, MILENA
Membro Prof. BARONTINI, ROBERTO
Membro Prof. MINA, ANDREA

• Nessuna parola chiave trovata

25/11/2024;

parziale

In the present work, we propose a nonparametric forecasting method within the framework of Functional Data Analysis (FDA). Our method is based on the notion of functional motifs, which are patterns that recur multiple times in data comprising a single curve, or a set of curves. A functional k-nearest neighbour approach is utilized to identify portions of the data similar to that immediately preceding the forecast interval, and the forecast is based on the observed evolution of such portions. Along with the forecasting method, we introduce diagnostic techniques that allow us to tune the parameters of our algorithm. The performance of our approach is assessed through a simulation study, considering multiple scenarios and comparing our results with those of standard ARIMA forecasts. Finally, we apply our approach to a case study concerning food price inflation.

Nome file	Dimensione
Ci sono 1 file riservati su richiesta dell'autore. Contatta l'autore