专家多源信息融合算法研究

Research on Multi-source Information Fusion Algorithm for Experts

为了全面获取科研及技术领域专家的信息,更加科学规范地对专家影响力进行评估,同时为专家选择科研合作对象提供参考,提出了专家多源信息融合算法。该算法从科技论文和专利中抽取作者信息,构建专家合作关系子网;从专家Web页面抽取专家相关数据构建Web子网;从科技论文、专利中抽取关键词和摘要,构建关键词词典,并分析专家研究主题,构建主题子网。通过专家合作关系子网、Web子网和主题子网融合算法,实现了专家多源信息的有效融合。实验以机器学习领域为研究点构建机器学习领域专家信息网络进行实证研究。结果表明,专家信息网络融合算法在个人信息构建与领域信息构建方面均能够准确表示专家个人信息与领域专家信息,验证了融合算法的有效性。

In order to comprehensively obtain expert information of scientific reasearch and technology,evaluate the influence of experts in a more scientific and standardized manner,and at the same time provide a reference for experts to choose scientific research cooperation objects,a multi-source information fusion algorithm for experts is proposed.The algorithm extracts author information from scientific and technological papers and patents,and constructs a subnet of expert cooperation relationships;extracts expert-related data from expert web pages to construct a Web subnet;extracts keywords and abstracts from scientific and technological papers and patents,constructs a keyword dictionary,and analyzes experts research topics and build topic subnets.Through the fusion algorithm of expert partnership subnet,Web subnet and subject subnet,the effective integration of multi-source information of experts is realized.The experiment uses machine learning as the research point to construct an expert information network in the field of machine learning to conduct empirical research.The empirical results show that the expert information network fusion algorithm can accurately represent the expert’s personal information and domain expert information in both personal information construction and domain information construction,which verifies the effectiveness of the fusion algorithm.