基于模糊文本分类的多知识领域专家推荐方法

在知识管理系统中,推荐满足知识需求的专家是隐性知识共享的有效手段.为了推荐多知识领域的专家,以及完整的识别用户的知识需求,提出了基于模糊文本分类的多知识领域专家推荐方法.首先建立模糊文本分类器,通过对专家上传的文档进行模糊文本分类,结合时间等因素建立专家的知识模型.通过分析浏览文档日志识别用户直接知识需求,在此基础上采用信息熵计算用户的潜在的知识需求.然后根据专家的知识模型与用户的知识需求模型的匹配度进行专家推荐.通过在知识管理系统中的成功应用验证了该方法的可行性与有效性.

Approach to expert recommendation with multiple knowledge areas based on fuzzy text categorization

Recommending an appropriate expert in knowledge management systems is an effective and efficient way to utilize tacit knowledge. To recommend the experts with multiple knowledge areas and identify the user-s knowledge needs completely, an approach to expert recommendation with multiple knowledge areas based on fuzzy text categorization was proposed. Firstly, fuzzy text classifier was constructed. An expert profile was built by classifying the newly registered knowledge artifacts fuzzily and with volume and time factors. Knowledge needs model was composed of explicit knowledge needs and implicit knowledge needs. Explicit knowledge needs was identified by analyzing the browsing logs. Implicit knowledge needs was measured by the entropy of the explicit knowledge needs. Expert was recommended based on the matching degree of expert profile and the user-s knowledge needs model. The proposed approach was developed and was used successfully in the knowledge management system. The approach was proved to be applicable.