P2P技术被广泛的用于网络节点之间的文件共享与搜索.采用P2P的搜索技术可以有效地提高了共享资源的深度和广度,但往往存在仅支持弱语义(甚至缺乏语义)共享的局限性.针对这一弱点,在研究和分析基于查询条件概率的统计语言模型的基础上,引入机器翻译领域中的统计翻译模型,改进统计语言模型的经典算法——一元语言模型,提出基于概率翻译方法的一元语言模型检索技术,并将改进的语义信息检索算法引入基于超级节点(super-peer)的P2P信息共享模型,建立支持语义的P2P信息共享模型,改善文件共享与检索性能.理论分析及原型系统验证了利用此模型来实现P2P网络语义文档共享的有效性.
Peer-to-peer(P2P) technology is widely applied in file-sharing systems.P2P networks shorten the data-updating cycle,and enhance the searching efficiency.However,current P2P file-sharing systems generally don′t support semantic search.Based on researches and analyses in statistical language model which is based on probability theory,a probability translation method based unigram language model is designed and established.Based on this method,we propose a P2P semantic sharing model,which involves semantic information retrieval model in super-peer based P2P file-sharing system.It enhances the performance of semantic retrieval.We developed a system and carried out experiments to test the availability of the P2P semantic sharing system.