On the Use of the Beta Distribution for a Hybrid Time Series Segmentation Algorithm

Áreas de investigación:

• Methodology - Machine learning

Año:

2016

Tipo de publicación:

Artículo en conferencia

Autores:

• Durán-Rosal, Antonio Manuel
• Dorado-Moreno, Manuel
• Gutiérrez, Pedro Antonio
• Hervás-Martínez, César

Volumen:

9868

Título del libro:

Proceedings of the 17th Conference of the Spanish Association for Artificial Intelligence (CAEPIA 2016)

Serie:

Lecture Notes in Computer Science (LNCS)

Páginas:

418-427

Organización:

Salamanca (Spain)

Mes:

14th-16th September

ISBN:

978-3-319-44635-6

ISSN:

0302-9743

BibTex:

@conference{CAEPIA2016Antonio, author = "Antonio Manuel Dur{\'a}n-Rosal and Manuel Dorado-Moreno and Pedro Antonio Guti{\'e}rrez and C{\'e}sar Herv{\'a}s-Mart{\'i}nez", abstract = "This paper presents a local search (LS) method based on the beta distribution for time series segmentation with the purpose of correctly representing extreme values of the underlying variable studied. The LS procedure is combined with an evolutionary algorithm (EA) which segments time series trying to obtain a given number of homogeneous groups of segments. The proposal is tested on a real problem of wave height estimation, where extreme high waves are frequently found. The results show that the LS is able to significantly improve the clustering quality of the solutions obtained by the EA. Moreover, the best segmentation clearly groups extreme waves in a separate cluster and characterizes them according to their centroid.", booktitle = "Proceedings of the 17th Conference of the Spanish Association for Artificial Intelligence (CAEPIA 2016)", doi = "10.1007/978-3-319-44636-3_39", isbn = "978-3-319-44635-6", issn = "0302-9743", month = "14th-16th September", organization = "Salamanca (Spain)", pages = "418-427", publisher = "Springer International Publishing", series = "Lecture Notes in Computer Science (LNCS)", title = "{O}n the {U}se of the {B}eta {D}istribution for a {H}ybrid {T}ime {S}eries {S}egmentation {A}lgorithm", url = "dx.doi.org/10.1007/978-3-319-44636-3_39", volume = "9868", year = "2016", }

Abstract:

This paper presents a local search (LS) method based on the beta distribution for time series segmentation with the purpose of correctly representing extreme values of the underlying variable studied. The LS procedure is combined with an evolutionary algorithm (EA) which segments time series trying to obtain a given number of homogeneous groups of segments. The proposal is tested on a real problem of wave height estimation, where extreme high waves are frequently found. The results show that the LS is able to significantly improve the clustering quality of the solutions obtained by the EA. Moreover, the best segmentation clearly groups extreme waves in a separate cluster and characterizes them according to their centroid.