Detection method of complex event based on the similarity of matching results

The research of complex event is a fresh branch of the field of data mining. The detection of complex event can help people obtain useful information from masses of data, therefore, it has received much attention and research. Nowadays, There has been some basic methods and technologies for detecting complex events, however, more and more different kinds of datasets are produced by DAE(Data Acquisition Equipment), which presents a challenge for limited memory and traditional detecting techniques. Based on that, this essay proposes a kind of universal event model to express data gathered by different equipments and the structure of Event Tree to compress event instances in memory, which can make limited memory store more event instances. By analyzing the similarity of matching results, new detection method was proposed based on shared subsequences of matching results. This detection method can make the best of shared subsequences between matching result sequences to construct new result sequences, which can reduce output delay and improve the efficiency of detecting complex event.