Intelligent Classification Technology For Natural Science Fund Management System

With the development of science and technology,the era of academic big data is coming.At present,the research on academic big data is showing diversified and deep development,and as an important part of this field,the classification and prediction system based on academic big data is also grad-ually mature.However,the current exploration of classification and predic-tion based on academic big data is mainly focused on the research of"things"(that is,academic papers),while the analysis of"people"(that is,scientific researchers)is insufficient.In fact,it is of great significance to the analysis of scholars,not only can explore the context and direction of scientific research development,but also become an important reference for the realization of engineering management.One of the goals of the Natural Science Foundation's academic big data intelligent classification system is to provide a more intelligent solution for the subject management of Natural Science projects.Because the applica-tion group of Natural Fund is very large,and the disciplines that the standard-ized management scholars fill in manually are very important and difficult,it is necessary to design a system that can realize the automatic division of scholars'disciplines.In response to this problem,based on the data provided by the Natural Science Foundation project,we explored and implemented a more reliable solution for intelligent classification and prediction of the Fac-ulty.Our work has mainly contributed in the following areas:? The novel operator-based spectral clustering was used to excavate the community structure of the academic collaborator network with a good degree of modularity,and the relationship between the dis-ciplinary characteristics and the community structure was quantita-tively analyzed using the NMI index.? A new feature extraction method is proposed to solve the faculty pre-diction for the Natural Fund academic collaborator network.The new feature model jointly models the cooperative relationship in the net-work of partners from three dimensions,which improves the accu-racy of the school's forecast.? A text feature model based on title enhancement is proposed to re-alize the faculty prediction for texts of Natural Science Foundation projects.This scheme explores the classification and prediction of disciplines from the perspective of natural language processing,and uses a new feature model to achieve more accurate predictions.