Research on the Optimization of Performance Evaluation Indicator System for Scientific Research Institutions
[Purpose/Significance]The performance evaluation of scientific research institutions is an effec-tive way to reflect the value of their research activities,outputs,and impacts,and its core content is to establish a scientific and reasonable evaluation indicator system.[Method/Process]This paper designed a set of systematic sub-domain and sub-regional performance evaluation indicator system to support the reform of scientific research evaluation.According to its characteristics of the performance evaluation,this paper constructed an optimization model based on random forest algorithm.By analyzing the contribution indicators in the existing system,it selected the common key indicators of performance evaluation.Meanwhile,it respectively analyzed the importance of indi-cators in different subjects and regions,optimized the structural weight ratio of common and individual indicators,and established a systematic system in different fields and regions.[Result/Conclusion]This indicator system can not only meet the evaluation objectives,but also dynamically and appropriately adjust the indicators according to the characteristics of scientific research institutions in many subjects and wide regions.
performance evaluationindicator system optimizationrandom forest algorithm
NETL
NSTL
万方数据
