摘要
科技成果评价的理论与方法问题是科学学理论中的重要学术性问题,也是科技管理实践中的关键技术性问题。通过借鉴商业世界中的一种资源配置方式——个性化推荐系统的思想,可以重新认识和更好地理解科技成果评价的本质、内涵和问题。通过比较科技成果评价和推荐系统两种不同的任务,可以发现二者在数据基础、实现步骤、目标任务和发展方向等方面都具有共同性。首先,二者在数据基础上是相似的,都是基于显式或隐式两种性质的反馈数据而实现的量化计算;其次,在实现步骤上是相同的,都需要经过不断优化的赋值和赋权过程才能确保输出结果的科学性;再次,在目标任务上是统一的,科技成果评价主要是“非个性化”的,而推荐系统主要是“个性化”的;最后,在演进方向上是一致的,由于个体智慧的有限性和群体智慧的优势,未来科技成果评价工作将和推荐系统一样,通过个人智慧和群体智慧的双螺旋交互而不断迭代优化。
Evaluation of scientific and technological achievements is an important academic issue of the science of science,as well as a key technical issues in scientific and technological management practice.In the era of information and the internet,how to effectively use technical means such as big data and artificial intelligence to develop research evaluation tools,improve the mechanism of evaluation of scientific and technological achievements,and promote the closer integration of science and technology with economic and social development is a matter of great concern to science and technology management departments at all levels.To this end,a noval thinking framework of the evaluation of scien-tific and technological achievements is proposed based on a resource allocation method commonly used in the business world recommendation system.This research compares the similarities and differences between the evaluation of scientific and technological achieve-ments and the recommendation system from four dimensions:data,methods,goals and basics.First,both of them have the similar data sources.In the recommendation system,the feedback behavior as the data basis can be divided into explicit feedback or implicit feedback.In the evaluation of scientific and technological achieve-ments,it also consist of two methods:peer review methods based on explicit evaluation data and scientometrics methods based on implicit evaluation data.Therefore,various problems of implicit feedback behavior in the recommendation sys-tem also exist in the evaluation of scientific and technological achievements.It is necessary to improve the reliability of feedback data by means of technical confidentiality,introduction of new indicators,setting up anti-cheating mechanisms,and increasingmanipulation costs.Second,both of them have same implementation steps.The evaluation of scientific and technological achievements,need to go through the process of assigning and weighting,just as recommendation system does.The quantification accuracy and transformation method in the process of assigning,as well as the implementation orientation and adjustment strategy in the process of weighting,are two basic steps in the quantification of the recommendation system or the evaluation of scientific and technological achievements.The consideration of these technical details is the premise and guarantee to ensure that the recommendation system or the evaluation of scientific and technological achievements can be carried out effectively.Third,both of them have unified goals and tasks.The recommendation system is mainly"personalized",while the evaluation of scientific and technological achievements is mainly"non-personalized".The evaluation of scientific and technological achievements selects the most recognized and recommended outstanding scientific and technological achievements based on and for the entire scientific community.However,with the continuous subdivision of disciplines and other factors,it is becoming more and more difficult for scientists to reach a consensus.Therefore,the pursuit of"semi-personalized recommendation"in the evaluation of scientific and technological achievements has become an expedient measure that can be practiced.Fourth,both of them have the uniform evolution direction.Essentially,they are both designed based on the basic assumption that the wisdom of the group is greater than the wisdom of the individual.Due to the limitation of individual intelligence,the collective intelligence of scientists often has a better performance in evaluation of scientific and technological achievements.Therefore,a better evaluation effect should be achieved through reasonable expert group selection,perfect process design and multiple rounds of iterative optimization.At the end of the article,some enlightenments about evaluation of scientific and technological achievements are also given.Science and technology managers should have a better understanding of the complementarity between peer review method and scientometrics method.They should also realize the complexity of the issues of assigning and weighting in the process quantitative evaluation,as well as the unity of achieving non-personalized goals based on personalized data.They can explore the new paths for the evaluation of scientific and technological achievements through the interactive optimization of personal and collective wisdom.
作者
胡志刚
HU Zhigang(Institute of Science of Science and S.&T.Management,Dalian University of Technology,Dalian 116024,China)
出处
《科学学与科学技术管理》
CSSCI
CSCD
北大核心
2023年第8期66-80,共15页
Science of Science and Management of S.& T.
基金
国家自然科学基金面上项目(71974030)
辽宁省“兴辽英才”项目(XLYC2007149)。
关键词
科技成果评价
推荐系统
显隐式反馈
分类评价
群体智慧
evaluation of scientific and technological achievements
recommendation system
explicit and implicit feedback
categorized evaluation
collective wisdom
