A Hybrid Association Algorithm with Rule Tree of Multiple Fuzzy Sequences
Practical problems which are solved by temporal association may be fuzzy, i.e., attribute value fuzziness and time-interval fuzziness, e.g., mining knowledge rule, like if there are serious cracks in the boiler, then it will lead to serious leakage in a short term, needs to consider the fuzziness of items (serious cracks) and the fuzziness of time (short term). For this problem, the hybrid temporal association rules algorithm with fuzzy itemsets and fuzzy time-interval is proposed in this paper. Furthermore, Fuzzy factor is introduced to measure the impact of random uncertainty and fuzzy uncertainty. Finally, we use the proposed algorithm to mine knowledge rules for knowledge inference and security alerts on the data of industrial boilers. Application results show that compared to the temporal association just considering fuzzy itemsets or fuzzy time-interval, our algorithm, considering both of them, is much more effective.
multiple fuzzy sequence fuzzy temporal association fuzzy rule tree knowledge inference
Qin Liao Jiepeng Zeng Zhicong Qiu
School of Mathematical Science South China University of Technology Guangzhou, China
国际会议
重庆
英文
1564-1568
2011-08-20（万方平台首次上网日期，不代表论文的发表时间）