摘要
认为学术论文评价方法在引导基础研究人员行为和科研产出方面具有重要作用,故有必要制定成本合理、科学公正的评价体系。以文献[3]提出的论文推荐-传播模型为基准,通过控制评价信息中大同行和小同行的比例、评价对象的选择方法以及给推荐者赋权3个策略来优化推荐-传播模型的学术评价功能。运用仿真方法验证优化策略的有效性后,提出利用平台用户的推荐-传播行为,自动计算学术论文影响指数、质量指数和价值指数的方法,从而构建一种新的学术论文评价方法。
Considering the critical role of academic paper＇ evaluation method in leading behaviors and scientific outputs of basic discipline researchers, it is of much necessary to design evaluation system with reasonable cost and scien- tific and fair methods. Taking the paper recommendation-propagation model proposed in Ref. [ 3 ] as the benchmark, three strategies are proposed to optimize the academic evaluation function of the recommendation-propagation model, including controlling the rate of big peers and small peers in evaluation information, choosing the appropriate method of evaluating objects, and assigning weight to eachpaper recommender. Based on validating the effects of the optimization strategies with simulation experiments, this paper proposes an academic paper＇ s evaluation method, which automatically computes influ- ence index, quality index and value index to measure the academic contribution of papers through users＇ recommendation -propagation behaviors.
出处
《图书情报工作》
CSSCI
北大核心
2014年第4期85-93,共9页
Library and Information Service
关键词
评价功能
协同过滤
同行评议
优化策略
evaluation function collaborative filtering peer review optimization strategy
