Periodic Pattern Mining of Embedded Multimedia Application Traces

Patricia López-Cueva (1, 2) , Aurélie Bertaux (3) , Alexandre Termier (4) , Jean-François Méhaut (2) , Miguel Santana (1)

1 ST-CROLLES - STMicroelectronics [Crolles]
2 LIG - Laboratoire d'Informatique de Grenoble
3 AMA - Analyse de données, Modélisation et Apprentissage automatique [Grenoble]
4 HADAS - Heterogeneous and Adaptive distributed DAta management Systems

Increasing complexity in both the software and the underlying hardware, and ever tighter time-to-market pressures are some of the key challenges faced when designing multimedia embedded systems. Optimizing debugging and validation phases can help to reduce development time significantly. A powerful tool used extensively when debugging embedded systems is the analysis of execution traces; however, huge trace volumes can make trace analysis unmanageable. To help in debugging, we focused on discovering periodic behaviors of multimedia applications through pattern mining of execution traces. Existing definitions of periodic patterns do not take into account unrestricted-sized gaps in the periodicity. In this paper, we specified a new definition of frequent periodic patterns that removes this limitation, and proposed an algorithm to mine these new patterns. Several experiments were carried out on real embedded application traces demonstrating that using these new patterns it is possible to identify abnormal behaviors.

Domaines: Apprentissage [cs.LG]; Intelligence artificielle [cs.AI]

Patricia López-Cueva, Aurélie Bertaux, Alexandre Termier, Jean-François Méhaut, Miguel Santana. Periodic Pattern Mining of Embedded Multimedia Application Traces. EMC 2012 - 7th International Conference on Embedded and Multimedia Computing, Sep 2012, Gwangju, South Korea. pp.29-37, ⟨10.1007/978-94-007-5076-0_4⟩. ⟨hal-00741206⟩