Anomaly detection in exploratory multivariate time series analysis

Data de caducitat: [30/06/09]
Orientació: recerca

Departament del director de la tesi: LSI

Director de la tesi: Alfredo Vellido
Correu electrònic: avellido@lsi.upc.edu

Comentaris:
Se espera que el desarrollo de esta tesis se realice completamente en inglés. La tesis puede estar sujeta (aunque no necesariamente) a la disponibilidad de datos externos.

Descripció de la tesi
One of the most relevant problems, in terms of practical applications, associated to the analysis of multivariate time series is that of the visual identification and characterization of atypical data subsequences. The research leading to the proposed master thesis will focus on the development of quantitative and qualitative markers of the different possible types of anomaly. In practical terms, it should result in the implementation of visually intuitive anomaly early warning methods.