Sequence Tree: A Data Structure For Improving Learning Sequence Constraint
Complex event processing (CEP) system is aimed at finding the rule-satisfied condition from plenty of event flows,
generating a composite event, and then sending it to the interested components. Usually, rules used in the CEP System are
defined by relevant experts. But for various reasons, it is a really difficult task to define perfect rules. This situation stimulates
the works about CEP rules learning. Sequence between the primitive events in a CEP rule is indispensable part of CEP rules,
and the most former works on learning sequence constraint are extracting the positive trace in the history record directly.
Through lots of experiments we find out that different positive traces have exactly similar number of primitive events and their
sequential relations under the same size of the window. Base on that we propose a new data structure—sequence tree to
combine the same positive traces to avoid unnecessary computing for improving efficiency of the algorithm. In the end of the
paper we design a set of experiments, and the experimental results validate the thought that using the sequence tree to
combining the positive trace is feasibly and effectively.
Index Items- complex event processing, rule learning, sequence constraint, combining positive trace