深层网中基于入口查询的表单填充策略

针对深层网中数据量大导致无法被传统搜索引擎索引的问题，在提取网页中，改进启发式规则识别表单查询入口，在表单标签与内容匹配时，改进基于语义的相似度匹配算法进行表单内容填充。实验结果表明，提取表单标签的准确率达到94.23%，匹配成功率达到88.83%，填充成功率达到95.43%。

Form Filling Strategy Based on Entrance Query in Deep Web

A Rating-Supervised Protocol(RSP)-based reputation mechanism is presented. By depicting autonomic elements trust degree on service and on rating differently and associating them dynamically, the non-rating, malicious rating can be distinguished and punished. Meanwhile, autonomic elements are stimulated to participate actively and rate honestly. The distributed implementation is provided. Experimental results show this RSP-based reputation mechanism can realize the differential service based on rating trustworthiness, and efficiently resist malicious attack.