一种基于TF·IEF模型的在线新闻事件探测方法

空间电磁编队具有不消耗推进剂、无羽流污染、连续可逆和同步控制等优势,静态电磁编队在光学干涉成像等领域具有广阔应用前景。动力学平衡态的稳定性与控制是实现静态电磁编队的基础。论文针对双星电磁编队沿径向、切向和法向分布的三种平衡态构形,利用Kane方法建立了6-DOF耦合非线性动力学模型,分析了三种平衡态构形的开环稳定性、耦合特性与控制需求,最后设计LQR控制器,实现了平衡态构形的稳定控制,并进行了仿真验证。

On-line news event detection based on TF.IEF model

We propose a new online news event detection (ONED) method which is based on TF?IEF model to improve the performance of detection. This method is motivated by TF?IDF, and calculates the weight of terms directly which present the event to establish a new incremental event model. There are two stages involved in our method. In the first stage, similar reports collected in a certain period are clustered into micro-clusters by Single-Pass. In the second stage, the micro-clusters are matched with existing events, and then the event model is updated by recalculating the VSM of event. Experiment shows that the proposed method improves the efficiency and accuracy of ONED with lower complexity and less feature information loss.