基于滑动窗口模型的非典型突发事件数据去噪算法研究

s the continuous growth of the telecom service and customer demands, top-quality and efficient service increases in importance. One of its challenging issues is to deal with the atypical incidents. While the traditional mining algorithms are focus on the high-frequent item sets, a de-noising algorithm related to the atypical incidents still remains unsettled. This paper proposed a de-noising model based on the sliding window. In this model, FP-tree and multi-association rules are introduced to fix the up-threshold of the sliding window. In addition, the certain rules are defined to fix the down-threshold. Then the thresholds are adjusted by the feedback of the algorithm's output, that means expanding or narrowing the sliding window until the optimum state. Experimental results demonstrate that the proposed algorithm can apply an appropriate data set to the knowledge discovery of the atypical incidents.