Case based reasoning approach to automatic comparison of models

In recent years much work has been done to find regularities in data streams. Normally streams are complete and well ordered. Major defects, such as incomplete or misleading events, occur in practice at least after some finite time. In this paper we describe a method developed for recognizing previously known patterns from a data stream. The solution is based on case based reasoning (CBR) with an extension to handle imperfect information. We also introduce a new idea to use two different case-bases instead of only one. The performance has been evaluated using simulated data and the obtained results are considered encouraging.