The Research And Implementation Of Key Techniques In Science-technology Expert Selection System

With the development of technology,the research fields involved in science-technology projects are becoming more and more diversified,and cross-field is becoming more and more common.Therefore,project review is faced with severe challenges.Science-technology experts are the core of project review,their familiarity with the research and their academic level in related fields determine the quality of the review results directly,which also bring challenges to the accuracy of expert selection.For expressing the research topic of science-technology projects and experts accurately,the model of science-technology entry is put forward,which is a more accurate classification method than subjects and technical fields and can reflect the theoretical and technical research category in science-technology work.On the basis of science-technology entry,this thesis studies the key technologies in science-technology expert selection system.The research work of this thesis mainly includes:(1)This thesis studies the science-technology entry portrait of research papers.Firstly,the research topic of research papers is defined by extracting science-technology entries and calculating the weight of them.Then the influence model and innovation model are established to judge the quality of research papers.(2)On the basis of research paper's science-technology entry portrait,this thesis studies the science-technology entry portrait of experts.Firstly,expert's research degree model on science-technology entry is established to define the research category of them.Then expert's contribution model on science-technology entry is established to judge their ability.(3)This thesis presents a single task expert automatic selection method.Firstly,the research topic of the selection task is defined based on science-technology entry.On this basis,combining with expert's science-technology entry portrait,the matching degree calculation model of the selection task and the expert is established.Finally,the selection of a single task is completed according to matching degree value.(4)To meet the needs of large-scale science-technology project's reviews,this thesis studies automatic batch selection algorithm for science-technology experts.Three batch task expert automatic selection algorithms are proposed: Sequential Expert Selection(SES),Batch Task Expert Automatic Selection Algorithm Based on 0-1 Integer Programming(IPES),Batch Task Expert Automatic Selection Algorithm Based on 0-1 Integer Programming + Adjust Strategy(IPES-AS).SES selects the experts in the order of the task based on the greedy idea,but its matching degree value is not the global optimal solution.IPES establishes a 0-1 integer programming model to maximize the overall matching degree on the selection task set.Based on IPES,IPES-AS uses an expert adjustment strategy to distribute the matching value of experts evenly.