基于超网络的企业微博知识挖掘及整合方法研究

There implies rich knowledge in huge data of microblog. However, the effective mining and integrating of fragmented knowledge on microblog are the urgent problems on knowledge mining for enterprise microblog. Therefore, this project presents the methods of knowledge mining and integrating based on supernetwork：the integrated association mining for data of enterprises microblog through supernetwork method, which combined the text mining method and SNA(Social Network Analysis) method,and can acquire knowledge through multi-network relations. Themethod can also integrate fragmented knowledge,etc. The research mainly includes: 1)Modeling method for systematic integration of enterprises microblog based on supernetwork, to build the Microblog Knowledge Supernetwork(MKSN). 2)Knowledge identifying, mining, integrating Methods for microblog texts based on MKSN through technologies as ontology, VSM, text mining. 3) Methods of fans' clustring, opinion integrating, and fans' interest discovering based on MKSN. 4)Methods of monitoring on dissemination of hotspots,special facts, and discovering opinion leaders based on MKSN and microblog interactive and transferring activities. A case study will be conducted to test the proposed methods, and a prototype system will be designed for the project.

相较于海量微博整体网数据，企业微博知识挖掘存在着因数据规模有限导致的轻量数据如何有效挖掘问题及碎片知识整合问题。提出一种基于超网络的企业微博知识挖掘与整合方法，基本思路：以超网络为工具，融合基于语义的文本挖掘与基于SNA的网络关系挖掘方法，实现微博知识的多网关联整合型挖掘，并对碎片知识进行整合。具体内容：1.基于超网络的企业微博系统整合建模方法，通过多种知识网络集成，构建微博知识超网络模型（MKSN）；2.基于MKSN的微博知识文本识别、挖掘及碎片整合方法，利用本体、VSM、文本挖掘等实现微博文本知识挖掘与整合；3.基于MKSN的粉丝（客户）聚类、粉丝观点挖掘整合、粉丝兴趣发现方法；4.基于MKSN及"文本-粉丝"互动模型的企业舆情监测、粉丝关注热点发现、意见领袖发现方法。此外，开发一原型系统并以若干企业微博为实例验证上述方法。

相较于海量微博整体网数据，企业微博知识挖掘存在着因数据规模有限导致的轻量数据如何有效挖掘问题及碎片知识整合问题。提出一种基于超网络的企业微博知识挖掘与整合方法，基本思路：以超网络为工具，对企业微博全要素数据进行整合性关联挖掘，整合内容包括微博要素整合（如文本、粉丝、互动传播等）、挖掘方法整合（文本语义挖掘+网络关系挖掘）、挖掘结果整合（碎片知识整合）等。具体内容：1.基于超网络的企业微博系统整合建模方法，通过多种知识网络集成，构建微博知识超网络模型（MKSN）；2.基于MKSN的微博知识文本识别、挖掘及整合方法，利用本体、VSM、文本挖掘等实现微博文本知识挖掘与整合；3）基于MKSN的粉丝（客户）聚类、意见领袖发现、潜在客户发现方法；4）基于MKSN及“文本—粉丝”互动模型的粉丝兴趣发现、意见整合、传播热点监测方法。此外，开发一原型系统并以若干企业微博或产品微博为例对上述方法进行验证。