论文部分内容阅读
为了解决自动文摘系统所面临的领域通用性和文摘质量的矛盾 ,提出了建造面向自动文摘的多Agent系统 ( MAS/ ABS)的方案 .建造这样的系统 ,有两个关键问题 :建立什么样的系统模型和选择什么样的协调算法 .给出了在 Internet环境下面向自动文摘的多 Agent系统模型 ,提出了 3种协调算法 .在仿真的基础上分析了系统性能 ,对 3种协调算法进行了比较研究 ,并得到了在一定负载下面向各个领域合适的文摘Agent数目 .
In order to solve the contradiction between field generality and digest quality of automatic abstract system, this paper proposes a scheme of constructing multi-agent system (MAS / ABS) for automatic summarization.There are two key questions in constructing such a system: what kind of System model and the choice of what kind of coordination algorithm.It gives the multi-agent system model for automatic summarization under the Internet environment, and proposes three kinds of coordination algorithms.On the basis of the simulation, analyzes the system performance, and carries out three kinds of coordination algorithms Comparative study, and get the appropriate number of abstracts Agent under a certain load for various fields.