JUCS - Journal of Universal Computer Science 21(6): 842-855, doi: 10.3217/jucs-021-06-0842
Discrete Neighborhood Representations and Modified Stacked Generalization Methods for Distributed Regression
expand article infoHéctor Allende-Cid, Héctor Allende§, Raúl Monge§, Claudio Moraga|
‡ Pontifícia Universidad Católica de Valparaíso, Valparaíso, Chile§ Universidad Técnica Federico Santa María, Valparaíso, Chile| European Centre for Soft Computing, Mieres, Spain
Open Access
Abstract
When distributed data sources have different contexts the problem of Distributed Re-gression becomes severe. It is the underlying law of probability that constitutes the context of a source. A new Distributed Regression System is presented, which makes use of a discrete rep-resentation of the probability density functions (pdfs). Neighborhoods of similar datasets are detected by comparing their approximated pdfs. This information supports an ensemble-basedapproach, and the improvement of a second level unit, as it is the case in stacked generalization. Two synthetic and six real data sets are used to compare the proposed method with otherstate-of-the-art models. The obtained results are positive for most datasets.
Keywords
distributed machine learning, context-aware regression, similarity representation