Research On Evaluation Model Of University Scientific Research Team Based On Neural Network

The comprehensive strength of the scientific research team in universities is an important indicator to evaluate the academic research level and innovative ability of universities,and it also has an important impact on the international competitiveness of Chinese universities.At present,the existing methods of evaluation of scientific research teams at home and abroad often focus on the current achievements of the scientific research team of the university,and ignore the academic competitiveness that affects the development and progress of the scientific research team of the university and the environment that affects the academic competitiveness.Moreover,the comprehensive strength evaluation of university research teams is a multi-factor,non-linear and complex process.Due to problems such as difficulty in assigning index weights,it is difficult to correctly evaluate the comprehensive strength of university research teams,resulting in low accuracy.In order to improve the accuracy of the evaluation results of the comprehensive strength of the scientific research team of the university,based on the analysis of the factors affecting the comprehensive strength of the scientific research team of the university,the mathematical model of the evaluation of the scientific research team of the university is established by using the neural network theory,in order to improve the scientificity and accuracy of the evaluation process.To provide a more scientific and comprehensive basis for the evaluation of the comprehensive strength of scientific research teams in universities.The main results of this paper are as follows:1.This paper studies and analyzes the current status of university scientific research team evaluation.Through the analysis of the connotation and main problems of university scientific research evaluation,based on the principle of scientific rationality,this paper proposes a method that includes academic competitiveness such as academic influence,industryuniversity-research Ability,team growth ability,subject integration ability and other indicators,including academic environment such as team leadership,member structure,member competition,incentive mechanism and other indicators comprehensive evaluation system.2.This paper introduces the SOM neural network algorithm,and improves the learning rate and neighborhood function of the traditional SOM neural network algorithm.Based on its excellent learning classification ability,a comprehensive strength evaluation model for university scientific research teams based on the SOM neural network is established.,Combined with Matlab software for simulation experiments,and compared the classification effects before and after improvement.The results show that the learning speed of the improved SOM neural network has been improved,and the accuracy of classification evaluation is relatively high,but the method still has certain limitations.3.As the SOM neural network training is affected by the input sequence when the sample data is small,this will lead to the phenomenon of local optimization,and its evaluation and classification effects have certain limitations.Therefore,this paper introduces the particle swarm optimization algorithm,and introduces its algorithm process and structure to a certain extent.Through the combination of the particle swarm optimization algorithm and the improved SOM neural network algorithm,that is,the improved SOM neural network is used to train and test the data samples to obtain better Weights,and then initialize the obtained weights to the PSO,which again significantly improves the evaluation classification effect.