Complex Event Detection Based On Regular Tree Pattern Matching

With the advent of the big-data age, the demands for processing stream data from social networks, financial data managements, network monitoring and other real-time systems are increasing dramatically and complexly. This kind of big data is featured with huge volume, complex data organizations and semantics as well as low-value data. There for how to effectively analyze and process such low-value density data from massive and complex stream data is a big problem in the development of data processing technology.A considerable number of stream data processing can be abstracted to the processing of event flows which makes streaming data query techniques can be used to detect complex events, and streaming data query mode can be used to describe the complex event patterns. Therefore, to develop the pattern of events with stronger ability for description and event detection pattern matching algorithm with higher efficiency for detection have become one of the main means to solve the problems of stream data processing in Big Data age.Quite a large number of complex event processing languages use formal language to describe complex events which are made up of several successive events, use succession, disjunctive, Kleene closure and other operators to appoint to the relation between events and detect the emergence of complex events through formal language pattern matching. However, semi-structured stream data organization in the internet is also a source of complexity of complex events. This complexity of the stream data structures can also cause the complexity of the event.Focused on the growing need of semi-structured stream data processing in Big Data era, we proposed a pattern of events called regular tree pattern and a pattern matching algorithm for detecting complex event derived from semi-structured data stream, trying to identify the occurrence of complex events through regular tree pattern matching for XML data stream, in accordance with regular expression to describe the arriving time of stream data and brotherhood between XML nodes, and also possible to describe the children of parents, and other ancestors descendants XML structure constraints. This stream data query technology integrated with normal tree pattern matching for semi-structured data and regular pattern matching technology for consecutive event, has a strong ability to describe and detect complex events, which is able to meet the need of the detection of complex patterns and the efficiency of complex event processing. It shows that the pattern has good performance and strong ability to describe events which can adapt to the feature of large-volume and low-value data in large data processing.