System failure prediction through rare-events elastic-net logistic regression

Navarro González, José Manuel; Parada Gélvez, Hugo Alexer y Dueñas López, Juan Carlos (2014). System failure prediction through rare-events elastic-net logistic regression. En: "2nd International Conference on Artificial Intelligence, Modelling and Simulation (AIMS 2014)", 18/11/2014 - 20/11/2014, Madrid, Spain. pp. 120-125. https://doi.org/10.1109/AIMS.2014.19.

Descripción

Título: System failure prediction through rare-events elastic-net logistic regression
Autor/es:
  • Navarro González, José Manuel
  • Parada Gélvez, Hugo Alexer
  • Dueñas López, Juan Carlos
Tipo de Documento: Ponencia en Congreso o Jornada (Artículo)
Título del Evento: 2nd International Conference on Artificial Intelligence, Modelling and Simulation (AIMS 2014)
Fechas del Evento: 18/11/2014 - 20/11/2014
Lugar del Evento: Madrid, Spain
Título del Libro: 2nd International Conference on Artificial Intelligence, Modelling and Simulation (AIMS 2014)
Fecha: 2014
Materias:
Palabras Clave Informales: Online Failure Prediction; Machine Learning; System Management; Automatic Feature Selection; Logistic Regression; Multivariable Prediction
Escuela: E.T.S.I. Telecomunicación (UPM)
Departamento: Ingeniería de Sistemas Telemáticos [hasta 2014]
Licencias Creative Commons: Reconocimiento - Sin obra derivada - No comercial

Texto completo

[img]
Vista Previa
PDF (Document Portable Format) - Se necesita un visor de ficheros PDF, como GSview, Xpdf o Adobe Acrobat Reader
Descargar (792kB) | Vista Previa

Resumen

Predicting failures in a distributed system based on previous events through logistic regression is a standard approach in literature. This technique is not reliable, though, in two situations: in the prediction of rare events, which do not appear in enough proportion for the algorithm to capture, and in environments where there are too many variables, as logistic regression tends to overfit on this situations; while manually selecting a subset of variables to create the model is error- prone. On this paper, we solve an industrial research case that presented this situation with a combination of elastic net logistic regression, a method that allows us to automatically select useful variables, a process of cross-validation on top of it and the application of a rare events prediction technique to reduce computation time. This process provides two layers of cross- validation that automatically obtain the optimal model complexity and the optimal mode l parameters values, while ensuring even rare events will be correctly predicted with a low amount of training instances. We tested this method against real industrial data, obtaining a total of 60 out of 80 possible models with a 90% average model accuracy.

Más información

ID de Registro: 36455
Identificador DC: http://oa.upm.es/36455/
Identificador OAI: oai:oa.upm.es:36455
Identificador DOI: 10.1109/AIMS.2014.19
Depositado por: Memoria Investigacion
Depositado el: 19 Jul 2015 09:10
Ultima Modificación: 19 Jul 2015 09:10
  • Open Access
  • Open Access
  • Sherpa-Romeo
    Compruebe si la revista anglosajona en la que ha publicado un artículo permite también su publicación en abierto.
  • Dulcinea
    Compruebe si la revista española en la que ha publicado un artículo permite también su publicación en abierto.
  • Recolecta
  • e-ciencia
  • Observatorio I+D+i UPM
  • OpenCourseWare UPM