AN UNCERTAIN COMPLEX EVENT PROCESSING METHOD BASED ON MNFA

Abstract

A plethora of data streams are generated every day due to the rapid development of the Internet. In the era of big data, real-time data stream processing becomes more and more popular in data mining. However, traditional techniques can no longer satisfy our needs, because of noise, sensor errors and other natural factors within data streams. In this paper, on the basis of complex event processing (CEP) technology, we propose a method combining NFA with match buffer (mNFA) using probability theory to address the issue. Our method can not only process massive uncertain event streams efficiently, but also support probabilistic event streams query processing. Furthermore, we also apply dynamic probability calculation algorithm and filtering event early to optimization.