Evaluation of few-shot learning models using performance estimates and ranking
Geometric evaluation of detected building roof boundaries in digital surface model...
Development of Large Language Models for Applications in the Legal Domain
Full text | |
Author(s): |
Santos, Moises R.
;
Mundim, Leandro R.
;
Carvalho, Andre C. P. L. F.
;
DeLaCal, EA
;
Flecha, JRV
;
Quintian, H
;
Corchado, E
Total Authors: 7
|
Document type: | Journal article |
Source: | HYBRID ARTIFICIAL INTELLIGENT SYSTEMS, HAIS 2020; v. 12344, p. 13-pg., 2020-01-01. |
Abstract | |
Meta-learning has been successfully applied to time series forecasting. For such, it uses meta-datasets created by previous machine learning applications. Each row in a meta-dataset represents a time series dataset. Each row, apart from the last, is meta-feature describing aspects of the related dataset. The last column is a target value, a meta-label. Here, the meta-label is the forecasting model with the best predictive performance for a specific error metric. In the previous studies applying meta-learning to time series forecasting, error metrics have been arbitrarily chosen. We believe that the error metric used can affect the results obtained by meta-learning. This study presents an experimental analysis of the predictive performance obtained by using different error metrics for the definition of the meta-label value. The experiments performed used 100 time series collected from the ICMC time series prediction open access repository, which has time series from a large variety of application domains. A traditional meta-learning framework for time series forecasting was used in this work. According to the experimental results, the mean absolute error can be the best metric for meta-label definition. (AU) | |
FAPESP's process: | 19/10012-2 - Meta-learning for time-series forecasting |
Grantee: | Moisés Rocha dos Santos |
Support Opportunities: | Scholarships in Brazil - Doctorate |