论文部分内容阅读
针对现行科研项目立项评估方法中存在的不足,提出多维框架证据推理的科研项目立项评估方法。根据多个识别框架之间的逻辑关系,给出平行框架、递进框架和混合框架的概念;将同行评议表中的“熟悉程度”和“综合评价等级”、“熟悉程度”和“资助意见”视为两个平行的递进框架,分别将两个递进框架评价信息转化为证据体,用评价等级(资助意见)和信度表示专家的评审意见,从而即能方便描述专家评价中的不完全信息,又能充分整合评审中有价值的信息;利用“熟悉程度”对专家进行赋权,利用证据推理算子将专家评估信息分别在两个平行框架上集结,并将不同项目进行量化排序择优;最后结合实例来检验该方法的有效性。
Aiming at the shortcomings in the current project evaluation methods of scientific research projects, a multi-dimensional framework evidence reasoning project evaluation method is proposed. According to the logical relationship among multiple recognition frames, the concepts of parallel frame, progressive frame and mixed frame are given. The familiarity of “” and “” “And ” Funding Opinions “as two parallel progressive frameworks, which respectively transform the evaluation information of two progressive frameworks into evidence body, and use the evaluation level (funding opinion) and reliability to represent the expert opinion, Can easily describe the incomplete information in the expert evaluation, and can fully integrate the valuable information in the evaluation; use ”familiarity" to empower the experts and use the evidence reasoning operator to divide the expert evaluation information into two parallel frames On the assembly, and different projects to quantify the priority selection; Finally, with examples to test the effectiveness of the method.