项目名称: 领域知识驱动的按需服务发现关键技术研究
项目编号: No.61202031
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 计算机科学学科
项目作者: 王健
作者单位: 武汉大学
项目金额: 23万元
中文摘要: 如何根据用户的个性化需求实施按需服务发现始终是面向服务的软件工程所面临的一个重要问题。Web 2.0模式下用户交互更加容易,基于用户反馈的增强式服务描述为服务发现提供了一条新思路,但与此同时,如何应对与情境相关的用户需求,如何为异构的各类服务提供一种统一的处理方式,又带来了新的问题。围绕"如何挖掘和利用领域知识以支持按需服务发现"这一关键科学问题,本项目拟开展如下研究:1)以数据库分块技术、形式概念分析方法和互信息计算为基础,建立情境感知的领域角色挖掘方法;2)利用融合领域特性的支撑向量机技术对服务进行领域分类,依据词汇的领域贡献度实现领域本体的渐进式构造,在此基础上,实现特定领域中服务面向主题的聚类、逐步累积领域目标;3)以领域知识(本体、角色、目标、服务)为基础,建立一种多策略相结合的按需服务发现方法。本项目的研究成果能为面向服务的领域知识生成以及在按需服务中的应用提供支持。
中文关键词: 服务发现;领域知识挖掘;按需服务选择;服务推荐;
英文摘要: How to realize on-demand services discovery based on users' personalized requirements is a key issue in service-oriented software engineering. Web 2.0 facilitates users' interaction with the Web, and thus service descriptions will be supplemented based on users' feedback, which provides new thought on services discovery. Meanwhile, how to deal with the context related requirements and the heterogenous services dscribed in description languages, poses new challenges on service discovery. In this project, faced with the key issue that how to mine and utilize domain knowledge for on-demand service discovery, we plan to conduct the following research: 1) we will propose a context-aware role mining approach based on database tiling, FCA (Formal Concept Analysis) and mutual information calculation; 2) we will improve the SVM (Supporting Vector Machine) Technology by merging domain-dependent information for domain services classification, which can also contribute to incremental domain ontology construction. Then we further cluster domain services based on topic (latent factor), and extract domain goals step by step; 3) based on the mined domain knowledge including domain ontology,roles,goals and services, we plan to develop a multi-policy on-demand service discovery method. The implementation of the project will contr
英文关键词: Service discovery;Domain knowledge mining;On demand service selection;Service recommendation;