Research On Evaluation Of Regional Innovation Ability Based On Decision Tree Genetic Algorithm And Neural Network

Regional innovation capability has increasingly become an important indicator for measuring a region's international competitiveness and economic development strength.The report of the 19 th National Congress clearly stated that "innovation is the primary driving force for the development and construction of an innovative country." At the same time,in order to implement the concept of "Internet +",the integration of Internet development and regional innovation capabilities has become a hotspot in recent years.The traditional evaluation method of innovation ability lacks the judgment of the importance of the evaluation index,and the determination of the index weight is easily interfered by subjective factors,which makes it lack of rationality and objective constraints.And the research methods mostly use a single neural network model,which has the problems of easy to fall into the local minimum and slow learning convergence.Under this background,on the basis of the traditional BP neural network(BP),this paper optimizes the structure and initial weight of the neural network through the decision tree and genetic algorithm respectively,so as to construct the BP based on the decision tree genetic algorithm Neural network(DTGA-BP)evaluation model aims to improve the accuracy of evaluation results and provide new ideas for regional innovation capability evaluation.This thesis firstly based on the research on the literature about innovation ability,through analyzing the regional innovation ability evaluation index,combined with the information gain feature selection method to determine the optimal feature combination,and finally constructed the evaluation index system.Experimental results show that using the optimal feature combination can increase the prediction accuracy of the BP neural network model,effectively reduce the impact of noisy data,and improve the stability of the prediction.Secondly,the combination weighting method is used to combine the analytic hierarchy process and the entropy weight method to determine the index weight,which avoids the arbitrariness problem of the traditional average weighting method when assigning weights,making the judgment of each index weight more reasonable and accurate.Then this thesis designs a regional innovation capability evaluation model of DTGA-BP neural network.The model solves the problem of the randomness of the initial weight assignment in the neural network by improving the selection,crossover and mutation operators of the traditional adaptive genetic algorithm.The decision tree algorithm determines the number of hidden layers of the neural network to reduce the model structure.the complexity.Experimental results show that the evaluation method based on the DTGA-BP neural network model proposed in this paper is more scientific than the traditional subjective evaluation method;in terms of prediction accuracy,the DTGA-BP neural network model is 41% higher than the single BP neural network model Compared with GA-BP neural network(GA-BP)model,it is improved by 20%,and the model has better stability and generalization ability.Finally,the prototype system of regional innovation capability evaluation is designed and implemented.The three aspects of system requirements analysis,database design,and core module design are elaborated.The core modules are mainly BP neural network model management module and regional innovation capability evaluation module.The evaluation provides a practical management tool.