基于关联性的目标动向热点分析

为了从海量传感器数据中及时发现重要目标的动向，提出了一种目标动向信息表征及关联方法，即从多源异类传感器获取的信息中抽取出目标关联要素进行动向表征，利用语义决策树实现动向要素聚类，通过知识规则进行关联匹配扩展，从而发现目标动向的热点，并进一步统计分析目标活动规律与发展趋势。实验表明所提出的基于关联性的目标动向热点分析算法准确率高，具有实用价值。

Hot spot analysis based on correlation of target activity information

A target activity expression and correlation method is presented in order to extract important tar-get activity from varying and massive sensor data. Activity factors are extracted from various information to denote target activity,then,semanticdecision tree is used to classify activity factors,correlative relationships between sensor data are established using knowledge and rules. By extraction of target activity information, it is possible to analyze hot spot or predict target activity trends by correspondent historical target activities. Experiment shows that the proposed target activity information correlation algorithm has high precision, which can be used in real data set.