Some variation of COBRA in sequential learning setup

TL;DR

This paper presents new variations of the COBRA regression method tailored for multivariate time series forecasting, demonstrating superior performance across diverse datasets using Bayesian optimization and grid search.

Contribution

It introduces novel COBRA-based approaches with specific data preprocessing and hyper-parameter tuning techniques for improved multivariate time series prediction.

Findings

Proposed methods outperform state-of-the-art models

Bayesian optimization enhances hyper-parameter tuning

Effective across cryptocurrency, stock, and load forecasting datasets

Abstract

This research paper introduces innovative approaches for multivariate time series forecasting based on different variations of the combined regression strategy. We use specific data preprocessing techniques which makes a radical change in the behaviour of prediction. We compare the performance of the model based on two types of hyper-parameter tuning Bayesian optimisation (BO) and Usual Grid search. Our proposed methodologies outperform all state-of-the-art comparative models. We illustrate the methodologies through eight time series datasets from three categories: cryptocurrency, stock index, and short-term load forecasting.