The proposed topic of the doctoral dissertation concerns the field of Artificial Intelligence called Complex Event Recognition. This field includes techniques designed to infer complex events from input data streams consisting of simple events with time stamps. Simple event streams usually stem from the processing of sensor measurements, while complex event inference is achieved using pattern matching algorithms. The purpose of this dissertation is to develop and optimize complex event recognition techniques with the aim of developing a system that can be used directly in real world applications. The tools developed in this field have been used to monitor shipping and aviation, to optimize road transport, and to timely identify the effectiveness of a combination of drugs in a cancer cell simulation. To this end, it is considered that the management of uncertainty and the ability to handle Big Data are essential features of an event recognition system. Uncertainty in the recognition environment may arise from a temporary malfunction of a sensor resulting in incorrect or blank indications in the input data. In addition, in modern applications we encounter high-velocity input data streams from numerous sources. For example, in naval fleet management, uncertainty can arise from communication gaps resulting in incomplete information about the position and course of a ship, while it is critical to handle high-velocity data streams from transmitters of hundreds of thousands of ships, which can be achieved by a highly efficient system. Thus, this dissertation will strive for the development of an event recognition system which is robust to uncertainty and handles data streams.