HotRank:热度敏感的非结构化数据检索排名算法

HotRank:heat-sensitive ranking algorithm for unstructured data search

为满足用户对非结构化数据检索的需求,分析用户对数据的操作行为,提出一种新型的数据热度敏感的非结构化数据检索排名算法HotRank。通过对数据操作情况(任务、访问次数、编辑时长等)进行日志记录,形成非结构化数据检索数据集。在此基础上,定义数据的任务相似度和数据热度计算方法实现该算法。结合实例仿真,对算法进行评估,并将仿真结果与其他算法进行比较,证明了该排名算法的准确率优于其他算法。

To satisfy users＇ retrieval needs for unstructured data, by analyzing data operations of users,this paper proposed a novel heat-sensitive ranking algorithm for unstructured data search which was named HotRank. By logging the operators of da- ta, such as task, access time, edit time etc. ,it created unstructured dataset. After that,it calculated similarity between task at- tribute of data and recent task, so as data heat, then established HotRank. Finally, it used unstructured data search to verify this algorithm and the result Of simulation. Compared with the other well-known algorithms, the results indicate that this algo- rithm is better than several other algorithms in precision.