面向移动对象的松散型传染模式挖掘方法

针对现有研究方案中对病毒或病菌的传染模式定义过于严格，可能丢失重要且正确的传染事件的问题，提出面向移动对象的松散型传染模式挖掘算法. 给出松散型传染事件的模式定义；提出基于滑动窗口的松散型传染模式挖掘算法(LIPMA)，按照传染事件发生的时间先后顺序，从初始传染源开始，利用滑动窗口机制，依次对每一个待检测对象进行分析处理，进而挖掘所有传染事件；提出基于R-tree索引的优化挖掘算法LIPMA+，该优化算法在每一轮的处理过程中，通过降低每一轮待检测对象的规模，实现挖掘效率的提升. 实验结果表明，所提出的传染模式挖掘算法能够对松散型传染事件进行高效、正确的挖掘，且能够挖掘更多潜在的传染事件；优化算法的挖掘效率显著提升，LIPMA+的平均挖掘时间仅占LIPMA的2%.

Loose infection pattern mining algorithms over moving objects

The definition of infection pattern of viruses or germ in the existing research schemes is so strict that important and correct infection events may be missed. Thus, a loose infection pattern mining algorithm oriented to moving objects was proposed. Loose infection model was defined and loose infection pattern mining algorithm (LIPMA), which used a sliding window mechanism, was proposed. According to the time sequence of the occurrence of infectious events, LIPMA uses the sliding window mechanism to analyze and process each object to be detected in turn, thereby mining all infectious events. On this basis, an optimized mining algorithm LIPMA+ based on R-tree was proposed. The optimized algorithm reduces the size of the object to be detected in each round of processing to improve the mining efficiency. Experimental results show that the proposed infection pattern mining algorithm can efficiently and accurately mine loose infectious events, and can mine more potential infectious events. The mining efficiency of the optimized algorithm was significantly improved, and the average mining time of LIPMA+ only accounted for 2% of that of LIPMA.