持续性实时监测数据的无偏风险挖掘仿真

Unbiased Risk Mining Simulation of Real-time Monitoring Data in Persistence

持续性实时监测数据的挖掘和调度常用在战场监控、森林火警监测等领域。在对持续性实时监测数据调度和挖掘中,因为战场监控、森林火警等持续性实时监测数据的挖掘中存在物理干扰,会产生时间延迟等风险,需要对其进行无偏估计和风险转化,提高数据的挖掘和实时监测性能。提出一种基于聚集树的持续性实时监测数据的无偏风险挖掘算法,并进行仿真实现。构建持续性实时监测数据的采集和预处理模型,进行监测数据的聚集树关联性分析,求得数据无偏相位特征,得到数据挖掘的最短无偏时延估计值,求得持续性实时监测数据的无偏风险挖掘最优路径无偏估计值。实验结果表明,该算法能有效避免数据挖掘过程中产生的时间延迟风险,通过无偏估计和风险转化,提高数据的挖掘和实时监测性能。

The real time monitoring data in persistence is commonly used in the field of mining scheduling battlefield monitoring, forest fire monitoring etc.. In the mining of real-time monitoring data persistent scheduling processing, because of the existence of physical disturbance mining real-time monitoring data of battlefield monitoring, it will have a time delay risk, need to carry on the unbiased estimation and risk transformation, improve the performance of data mining and real time monitoring. An unbiased risk mining algorithm for real-time monitoring persistent data is proposed based on aggregation tree, and simulation is realized. Continuous real-time monitoring data acquisition and pre-processing model is constructed, analysis of monitoring data aggregation tree relevance is taken, the data unbiased phase characteristics are obtained, the data mining of the shortest unbiased estimation of time delay value is obtained, unbiased risk mining optimal unbiased estimator for continuous real-time monitoring data is computed. The experimental results show that, the algorithm can effectively avoid the data mining process of time delay risk, the unbiased estimation and risk transformation is obtained, the performance of data mining and real time monitoring is improved.