The purpose of my thesis is to develop a program for identifying typical or extraordinary events in time series.
I describe the usefulness of the subject in real life with the focus on health issues.
I got acquainted with existing pattern searching algorithms to prepare my thesis. With no claim of being exhaustive, I describe procedures for pattern search and anomaly testing.
In the dissertation I present an algorithm with sliding time window, which I also implement. I write down the initial decisions, with I have specified the problem and after that, gave a user guide to the program.
In addition, I provide the user with a solution to various event pattern and anomaly searches. I implemented a generator program, that can generate test data for search engines.
The event pattern and anomaly recognition program is a solution to the tasks presented. The program graphically presents the results got with the parameters set before.
Important considerations were the easy handling of the program and the display of clear, comprehensible results to the user.
My solution is a stand-alone Java application that is capable of analyzing events. I was able to implement, among other algorithms, the explored event-folding window algorithm and use it to search for patterns in time series.