Detecting typical time series episodes

OData support
Supervisor:
Dr. Pataki Béla
Department of Measurement and Information Systems

In my thesis, I consider the problem of detecting typical episodes in event sequences. In contrast with the most popular thread within this theme, I detect an episode, that is already known.

At the introduction of the work, I give a review of the importance of the current area, and I introduce the first thoughts about the formalism and the application, that I going to make. After that, I write a summary of the related work in the current area.

In the rest of the thesis I describe my framework and formalism, that are capable for defining the most act of an everyday life easily, with the require flexible in it. I explain the algorithms, made upon my formalism, which makes us capable to detect the episode in an event sequence. At the end of this part I make a review about my application, that helps to define an episode upon my formalism, and can find the pattern in a sequence, according some parameters.

At the end of the thesis, I make some tests, with the data, made by the daily activity simulator on the department. During the tests I examine the points of the resource-usage, depends on some parameters.

The very end of my thesis I present some extension ideas, in the current area.

Downloads

Please sign in to download the files of this thesis.