Abstract
Nowadays, telecommunication networks occupy a predominant place in our world. Indeed, they allow to share worldwide a huge amount of information. Networks are however complex systems, both in size and technological diversity. Therefore, it makes their management and repair more difficult. In order to limit the negative impact of such failures, some tools have to be developed to detect a failure as soons as it occurs, analyse its root causes to solve it efficiently, or even predict this failure to prevent it rather than cure it. In this thesis, we mainly focus on these last two problems. To do so, we use files, called alarm logs, storing all the alarms issued by the system. However, these files are generally noisy and verbose: an operator managing a network needs tools able to extract and handle in an interpretable manner the causal relationships within a log. First, we build online a structure, called DIG-DAG, that stores all the potential causal relationships involving the events of a log. We then propose a query system to exploit this structure. Finally, we apply this approach in the context of root cause analysis. Second, we discuss a generative approach for times series prediction. In particular, we compare two well-known models for this task: recurrent neural nets on the one hand, hidden Markov models on the other hand. Indeed, in their respective communities, these two models are state of the art. Here, we compare analytically their expressivity by encompassing them into a probabilistic model, called GUM.
