SETAR-Tree: A Novel and Accurate Tree Algorithm for Global Time Series Forecasting
DOI10.48550/ARXIV.2211.08661V1arXiv2211.08661v1MaRDI QIDQ74449FDOQ74449
Authors: Rakshitha Godahewa, Geoffrey I. Webb, Daniel Schmidt, Christoph Bergmeir 
Publication date: 16 November 2022
Cited In (1)
Summary: This article explores various advanced forecasting techniques including Threshold Autoregressive Models (TAR), Generalized Fractional Models (GFMs), and Tree-Based Forecasting Methods. These methodologies are increasingly popular due to their capability to handle nonlinear relationships, capturing regime switching dynamics, and generalize traditional linear models. Applications of these approaches span across finance, economics, environmental sciences, and beyond. The study introduces SETAR-Tree and SETAR-Forest algorithms, combining TAR strengths with tree-based methods for improved forecasting performance in multiple sectors.
Summary_simple: In recent years, researchers have developed several advanced models to predict trends in various fields like finance and economics. These include Threshold Autoregressive Models (TAR), Generalized Fractional Models (GFM), and tree-based methods that can handle complex relationships between variables. One study has combined TAR with tree-based algorithms for better forecasting, potentially useful across multiple domains.
This page was built for publication: SETAR-Tree: A Novel and Accurate Tree Algorithm for Global Time Series Forecasting
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q74449)
