神经网络模型在城市科技竞争力评价中的应用研究—基于浙江省的实证分析

发布时间：2018-05-19 00:24

本文选题：城市科技竞争力 + BP神经网络模型　；参考：《杭州电子科技大学》2012年硕士论文

【摘要】：城市科技竞争力的研究，可以使城市客观的分析自身存在问题和竞争优势，明确发展方向、制定具有针对性的科技发展战略，从而提高国家和地区整体的科技水平，促进经济社会的快速稳定发展。浙江省城市科技的发展整体上走在国家前列，但不可否认城市间的科技的发展程度和重视水平良莠不齐，差别很大。浙江省城市科技竞争力的研究，对如何利用已有的优势科技资源提高城市的经济发展水平，如何在整体上缩小城市间的科技实力距离，实现浙江省城市科技竞争力的整体提高，增加其国内国际影响力，有着非常重要的意义。目前，在城市科技竞争力的研究领域，大部分学者采用的是传统的统计方法如因子分析法等对城市科技竞争力进行研究。缺乏运用神经网络方法对其进行研究的相关文献。在此前提下，本文主要探讨了数据挖掘技术神经网络模型在城市科技竞争力研究中的应用。研究思路如下：首先，在参阅大量的相关研究文献的基础上，参照浙江省每年公布的各地市科技进步统计监测报告，结合专家意见打分，确定了全面反映浙江省各个地级市城市科技竞争力的指标体系。根据所建立的指标体系，运用浙江省11个地级市2009年的数据，建立关于城市科技竞争力的BP神经网络模型和径向基函数（RBF）神经网络模型，并利用CHAID决策树模型与神经网络模型进行对比分析研究。最后得到一个较好的反映浙江省城市科技竞争力的神经网络模型。然后，根据所建立的神经网络模型，代入2010年的指标数据，和浙江省2010年城市科技竞争力得分、排名进行对比分析，，以检验神经网络模型对浙江省城市科技竞争力的预测效果。为了分析影响浙江省城市科技竞争力的影响因素，考虑到政策建议上的时效性，运用2010年浙江省11个地级市的相关数据重新建立神经网络模型。根据模型的变量重要性分析，选取影响浙江省城市科技竞争力最大的前20名的指标，为浙江省提升城市科技竞争力提供政策建议参考。最后利用聚类分析研究浙江省城市科技竞争力在地域上的分布，研究结果表明，东部长三角地区城市的科技竞争力要明显高于其它地区城市。浙江省要想实现城市科技竞争力的整体提高，缩小城市间科技竞争力的差距，必须走区域科技建设合作道路，做到城市间科技资源的优势互补，相互帮助，以达到整体提升浙江省城市科技竞争力的目的。接着通过对浙江省各个地级市科技实力与经济实力的对比分析，将浙江省11个地级市分为了三类，分析了浙江省各个地级市的科技现状和经济发展方向，为提高各个地级市的城市科技水平指明了努力的方向。本文主要的创新点归纳为如下三点：一是在前人研究的基础上，提出了自己的城市科技竞争力的概念，并对其内涵做出了阐释。二是在城市科技竞争力评价指标的选取上，结合以往研究文献和专家建议，参考浙江省科技厅发布的各地市科技进步统计监测报告，多渠道结合保障所建立指标体系的合理性。三是在研究方法上，做了运用神经网络模型在城市科技竞争力研究方面的先期探索，并利用决策树模型进行对比研究分析，为城市科技竞争力的实时控制和动态分析做出了一定的贡献。通过论文研究，全面的分析了浙江省城市科技竞争力的现状，为浙江省提高城市科技竞争力，建设创新性城市提供参考借鉴。
[Abstract]:The research on the competitiveness of urban science and technology can make the city objectively analyze its own existing problems and competitive advantages, make the development direction clear and formulate the strategy of scientific and technological development with the aim, thus improve the scientific and technological level of the country and the region as a whole and promote the rapid and stable development of the economy and society. The development of urban science and technology in Zhejiang province is on the whole in the country. But it is not possible to recognize the development degree of the science and technology between cities and the difference between the good and the bad. The research on the competitiveness of urban science and technology in Zhejiang Province, how to improve the economic development level of the city by using the existing advantages and technology resources, and how to reduce the technological strength distance in the city on the whole, and realize the city science and technology competition in Zhejiang province. The overall improvement of competitiveness is of great importance to increase its domestic and international influence.
At present, in the research field of urban science and technology competitiveness, most scholars use the traditional statistical methods, such as factor analysis, to study the competitiveness of urban science and technology. The lack of relevant literature on the use of neural network method to study it. The application of scientific and technological competitiveness is as follows:
First of all, on the basis of referring to a large number of relevant research documents, referring to the annual statistics and monitoring reports on scientific and technological progress published in Zhejiang province every year, and in combination with experts' opinions, the index system that comprehensively reflects the scientific and technological competitiveness of various cities in Zhejiang province is determined. According to the established index system, 11 cities of the Zhejiang province are 2009 The BP neural network model and radial basis function (RBF) neural network model of urban science and technology competitiveness are established, and the CHAID decision tree model and neural network model are used to compare and analyze the neural network model. Finally, a better neural network model which reflects the competitive power of urban science and technology in Zhejiang is obtained.
Then, according to the established neural network model, the index data of 2010 and the city science and technology competitiveness score of Zhejiang Province in 2010 are compared and analyzed in order to test the prediction effect of the neural network model on the competitiveness of urban science and technology in Zhejiang province. In order to analyze the influence factors of the competitiveness of the city's science and technology in Zhejiang Province, the government will consider the politics. According to the relevant data of 11 cities of Zhejiang Province in 2010, the neural network model is re established by using the related data of 11 cities in the province. According to the analysis of the variable importance of the model, the first 20 indexes which affect the most competitive power of urban science and technology in Zhejiang province are selected, and the policy suggestions for improving the competitive power of urban science and technology in Zhejiang are provided.
At last, using cluster analysis to study the distribution of urban science and technology competitiveness in Zhejiang province. The results show that the scientific and technological competitiveness of the cities in the eastern Yangtze River Delta is obviously higher than that of other cities. In order to realize the overall improvement of the competitiveness of the city's science and technology and narrow the gap between cities and cities, Zhejiang must take regional science and technology. In order to achieve the overall promotion of the scientific and technological competitiveness of the city of Zhejiang Province, the technological and economic strength of the cities of Zhejiang province are analyzed, and the 11 prefectural cities in Zhejiang province are divided into three categories, and the departments of various prefectures and cities in Zhejiang province are analyzed. The status quo of technology and the direction of economic development have pointed out the way to improve the level of urban science and technology in various prefecture level cities.
The main innovation points of this paper are summed up as follows: three points:
First, on the basis of previous studies, I put forward my own concept of urban scientific and technological competitiveness and explained its connotation.
The two is the selection of the evaluation index of the city's scientific and technological competitiveness, combined with the previous research literature and expert advice, referring to the statistics and monitoring reports of the city's scientific and technological progress issued by the Zhejiang provincial science and Technology Department, and the rationality of the index system established by the multi channel support system.
The three is on the research method, it has made the first exploration of using the neural network model in the research of city science and technology competitiveness, and makes comparative analysis by making use of the decision tree model to make a certain contribution to the real time control and dynamic analysis of the city's scientific and technological competitiveness.
Through the study of this paper, the present situation of city science and technology competitiveness in Zhejiang is analyzed comprehensively, which can provide reference for Zhejiang province to improve the competitiveness of urban science and technology and to build an innovative city.
【学位授予单位】：杭州电子科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP183;G322.7

【参考文献】