关键词: RFID数据，复杂事件处理，模式匹配，自动机

Abstract: RFID data is generally uncertain. Complex event processing (CEP) treats the data as different types of events,queries sequence of events in which match specific patterns of sequence are defined by high-level application from the event stream. Event stream is divided into multiple alternative event stream and single alternatives event stream.NFA-MMG pattern matching method for multiple alternatives event stream was proposed. hhe method uses combination of directed acyclic graph and automatic machines to achieve complex event pattern matching on the uncertain data.NFA-Tree pattern matching method for single alternatives event stream with the use of matching tree and automatic machines on uncertain data was proposed. The NFA-Tree algorithm was improved by pruning the matching tree to improve the efficiency of query optimization, which filters the results of the match situation based on probability threshold. The complex event processing system prototype uncertain data was developmented to realize the above algorithm,and the experiment examines the validation and performance of the algorithms.

Key words: RFID data, Complex event processing, Pattern matching, Automatic machines