Analysis And Prediction Of Private Car Ownership Based On Grey-Generalized Regression Neural Network

With the continuous enhancement of China’s comprehensive strength,people’s quality of life and material living standards have improved greatly.By the end of 2018,the number of private car ownership in China alone has exceeded 200 million.In the near future,the number of private car ownership will continue to grow rapidly and steadily.Although private cars are a convenient tool for people to travel,they have also created some social problems that need to be solved,such as increasing traffic congestion and air pollution.By forecasting private car ownership,we can not only see whether the country’s economic and social development is healthy and efficient,but also directly reflect the development of the entire domestic automobile market,so forecasting private car ownership in China has become a hot research problem.This paper uses gray theory as well as generalized regression neural network to forecast and analyze private car ownership.In Chapter 1,the background and significance of the research are discussed in detail,and the research content of this paper is proposed through an in-depth analysis of the current situation of domestic and foreign research.In Chapter 2,the characteristics of private car ownership forecasting and the analysis and selection of influencing factors are introduced,and the shortcomings of three commonly used models,namely,regression forecasting model,gray forecasting model and BP neural network forecasting model,in private car ownership forecasting are analyzed.In Chapter 3,the basic theory and network structure of generalized regression neural networks and the basic theory of gray correlation analysis are introduced,and the advantages of using generalized regression neural network forecasting models for private car ownership forecasting are pointed out.The three methods of finding the optimum for generalized regression neural networks,namely grid search,particle swarm optimization,and fruit fly optimization,are also introduced,and the methods of testing the prediction results are explained.In Chapter 4,the historical data of private car ownership and the influencing factors are selected,and the factors with strong correlation are selected using gray correlation analysis.The shortcomings of the generalized regression neural network in predicting private car ownership data are then analyzed and the data processing method of unequal spacing growth rate is proposed.A gray-generalized regression neural network prediction model is developed for private car ownership prediction,and the three different methods proposed in this paper are used to find the optimal smooth factors and compare the prediction accuracy of the three models.By comparing with the three commonly used models,the effectiveness of the gray-generalized regression neural network adopted in this paper for private car ownership forecasting is demonstrated,and the model with high accuracy is used to predict the data in the next 5 years.In Chapter 5,the forecasting methods used are summarized and future research is presented.