Application Research Of Improved RBF Neural Network Algorithm In Enterprise Management Iin Dongfeng County

Enterprises are the basic elements of the development zone.The healthy development of the development zone depends on the green development of enterprises.The economic development,green development and scientific and technological development of enterprises have become the important cornerstones for the sustainable development of the development zone.In the past,the economic development of enterprises was the most important part in the evaluation of enterprises.Then,the economic gross value and tax value of enterprises were systematically analyzed.According to the single economic data,the degree of support and the work of leaving or staying in the park were decided.However,the scale and development of enterprises were not uniform.With the overall influence of social factors,a single help was used.Enterprise policies can not meet the needs of most enterprises,and enterprises with high energy consumption and pollution tend to be concentrated in development zones.Therefore,green enterprises are proposed in development zones all over the country.The enterprises that are most suitable for development in the zones are determined according to various variables of enterprises.The goal is to achieve the best economic development of enterprises under the premise of reducing energy consumption and improving science and technology.With extension in different areas of science and technology,data mining is widely applied to the enterprise developing step by step,using the data mining technology can improve the management of the development zone,the neural network is one of the most widely used classification data mining algorithm,has been in pattern recognition,intelligent robot,predictive evaluation,economic and other fields a wide range of fields to solve the practical problems,mainly used in the nonlinear data processing work,it has associative storage,self-learning,high-speed looking for optimization solution,and other functions,the typical algorithms are: BP neural network,RBF neural network and so on several kinds of algorithms.This research according to the diversity and difference of Dongfeng county development zone enterprises,using the BP neural network in data mining,the research area of the staff input,output,energy consumption of three main influence zone were studied respectively,and the test results analysis,in order to the enterprise green development of the development zone to provide technical support.Main research contents and results:(1)BP neural network is used to study the hierarchical classification of enterprise management.BP neural network was used to divide the grade domain of green development of enterprises in Dongfeng county development zone in 2015,2016 and 2017.The experimental results showed that the accuracy was 86.3%,80.90% and 86.36%,respectively.The BP neural network algorithm can carry on the preliminary division to the development zone enterprise,which indicates that the BP neural network algorithm has the feasibility to the development zone enterprise division.(2)An improved RBF neural network is proposed.Utilizing the complementarity of rough set and RBF neural network,extracting and determining the weight factors of data attributes,and determining the comprehensive weight of enterprise development management data of the grass-roots development zone through comprehensive determination of the weight of each attribute,the weight is 0.55,0.25 and 0.2.The weight is applied in combination with RBF neural network algorithm to avoid the occurrence of minimum value phenomenon in the calculation process of BP neural network.(3)The comparative experiment of rough set RBF and BP neural network is carried out.The classification of green development status of enterprises in Dongfeng county development zone in 2015 is made.The accuracy of rough set RBF neural network and BP neural network are 96.45% and 86.3% respectively.It shows that rough set RBF is more accurate than BP neural network in enterprise classification of suitable development zone.(4)Rough set RBF is applied to hierarchical management of enterprise management in development zone.The application of rough set RBF neural network algorithm in the grade domain division of enterprise green development in dongfeng county development zone is carried out.The rough set RBF neural network algorithm can efficiently and accurately predict and analyze the green development of enterprises in development zone,which is relatively consistent with the actual local enterprise management.Research shows that the application of neural network algorithm and other data mining algorithms can achieve the analysis of enterprise management data,scientific and reasonable use of human resources in development zones,and to improve the land use rate,strengthen the enterprise green development classification of efficient means,which will be the direction for countries to study the new pattern of development zones.