论文部分内容阅读
【目的】为全面获取专家资源,探究多源专家特征信息融合方法。【方法】从传感器工作过程出发,依次论述基于知识传感器、Web传感器和社会网络传感器的专家特征识别方法。鉴于三种方法获取的专家特征向量存在冲突,围绕资源均衡度设计基于多源信息融合的专家特征识别方法。【结果】与C-DBLP统计专家特征进行匹配,相似度达到38.97%,与同类型方法比较,结果在正常范围内。【局限】识别对象多来自高校及科研院所,用于特征识别的资源也多为学术资源,同时Web传感器采集网址集合还有待扩展。【结论】在语词关系控制情形下,该方法可用于科研团队构建、专家推荐、专家检索等方面。
【Objective】 To comprehensively acquire expert resources and explore multi-source expert feature information fusion method. 【Method】 Starting from the working process of the sensor, the method of expert identification based on knowledge sensor, Web sensor and social network sensor is discussed in turn. In view of the conflict of the expert eigenvectors obtained by the three methods, an expert feature identification method based on multi-source information fusion is designed around the resource balance. 【Result】 The results showed that the similarity with C-DBLP statistics experts was 38.97%. Compared with the same type of methods, the results were within the normal range. [Limitations] Mostly identify objects from universities and research institutes, the resources used for feature identification are mostly academic resources, and web sensor collection web site collection remains to be extended. 【Conclusion】 Under the condition of word-relation control, this method can be used in the construction of scientific research team, expert recommendation and expert search.