Research On Evaluation Of Smart City Development Level Based On Support Vector Regression Optimized By Genetic Algorithm Model

The wisdom city is the important direction of urban development in the future,and it is helpful to scientifically evaluate the development level of Chinese wisdom city,which will contribute to the formulation of wisdom city planning and construction policy.The purpose of this paper is to construct a correct and objective intelligent city evaluation index system and to evaluate the construction level of China's intelligent city.First of all,through the establishment of first-level indicators and two indicators to build a complete objective and intelligent city evaluation index system,the selection of a target,the use of fuzzy sets of ideas,through the membership function,with the frequency of the expression of membership,That is,the indicators with high degree of membership are selected into the index system.The selection of the second-level indicators is divided into two steps: First,under each first-level indicator,by referring to the relevant statistical data such as "China Statistical Yearbook" In the second step,due to the wide range of the secondary indexes selected and the redundant attributes,the Affinity Propagation clustering algorithm is used to cluster the second-level indexes under each first-level index.Two indicators are classified as a class,to be a complete intelligent city evaluation index system.Then,a multi-attribute decision-making flow based on neighborhood rough set theory is established.On the basis of analyzing equivalence between attribute reduction and threshold reduction,a variable neighborhood rough set theory based on variable precision is proposed to attribute the intelligent city index system Reduction,to be streamlined,objective and intelligent city evaluation index system.Finally,according to the evaluation index system of intelligent city in China,the construction level of Chinese intelligent city is evaluated.The construction level of intelligent city is evaluated by machine learning algorithm.China is now building or put forward the concept of the wisdom of the city more than 100 seats,due to the limitations of the sample itself,so the algorithm should be selected while the performance of the algorithm itself,as well as the algorithm is optimized.Based on the above considerations,this paper will use the regression algorithm for the construction of the wisdom of the city to evaluate the level.The model includes genetic algorithm optimization support vector regression model,genetic algorithm optimization BP neural network model,stochastic forest model,gradient promotion regression tree model.The training data and prediction data come from China's intelligent urban evaluation index system in 2015.The training set and prediction set are separately divided according to the ratio of 8 to 2.The MSE is selected as the evaluation index and the MSE of the prediction set can be found that SVR is the best by genetic algorithm,so this method is used to evaluate the construction level of China's smart city in 2016,and the evaluation index of each intelligent city is obtained.