Automatically weighting ensemble-based time series classification algorithms

• Grant number: 22/12498-2
• Support Opportunities: Scholarships abroad - Research Internship - Scientific Initiation
• Effective date (Start): November 21, 2022
• Effective date (End): March 20, 2023
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Diego Furtado Silva
• Grantee: Anderson Henrique Giacomini
• Supervisor: Anthony Bagnall
• Host Institution: Centro de Ciências Exatas e de Tecnologia (CCET). Universidade Federal de São Carlos (UFSCAR). São Carlos , SP, Brazil
• Research place: University of East Anglia (UEA), England
• Associated to the scholarship: 22/00302-6 - Meta-learning for the selection of algorithms and features for time series classification, BP.IC

Abstract

Time series data is becoming ubiquitous in our everyday tasks. The increasing collection of data observed over time has also increased the number of techniques for classifying time series. In this scenario, when a researcher or practitioner needs a classification solution, there is a vast range of options of algorithms to apply. When only one algorithm is chosen, its performance may not achieve satisfactory results for solving the problem. However, ensemble-based methods try to reduce the impact of a poor choice of an algorithm by avoiding a wrong choice of algorithm dominating the classification. For instance, HIVE-COTE 2.0 is a meta-ensemble that uses weighted ensembles from different categories of classification algorithms where the contribution of each category to the final decision depends on their accuracy estimated through cross-validation in the training set. This weight estimation step is a relevant bottleneck for using HIVE-COTE 2.0 for large volumes of data. In this context, to circumvent this bottleneck, we propose using simple meta-learning-based models to automatically and efficiently estimate weights for the algorithms comprised by HIVE-COTE 2.0. (AU)