Improving the Global Fitting Method on Non-Linear Time Series Analysis

TL;DR

This paper introduces an improved global fitting method for non-linear time series analysis that enhances forecast accuracy without significant computational costs, building on existing global mapping techniques.

Contribution

The paper proposes a novel enhancement to the global fitting approach that significantly improves forecasting accuracy in non-linear time series analysis.

Findings

Increased forecast accuracy demonstrated

Method maintains low computational cost

Applicable to various non-linear time series

Abstract

In this paper, we are concerned with improving the forecast capabilities of the Global approach to Time Series. We assume that the normal techniques of Global mapping are applied, the noise reduction is performed, etc. Then, using the mathematical foundations behind such approaches, we propose a method that, without a great computational cost, greatly increase the accuracy of the corresponding forecasting.