Dynamic Evaluation and Optimization of Investment Environment in Node Cities on the Maritime Silk Road

Understanding and evaluating urban investment environment is essential for effectively improving the efficiency of resource allocation between cities and promoting overall development of the regional economy. This paper takes 15 node cities on maritime Silk Road covered by the “Belt and Road” as the research object, establishes a dynamic evaluation index system for investment environment, and uses projection pursuit cluster to analyze and evaluate the investment environment of the cities. It is found that the investment environment potential of a city is directly related to the level of social development, economic development, and the degree of opening to the outside world. It is recommended that node cities should seize the important opportunity of the construction of the Maritime Silk Road, introduce world-wide human, financial and material resources to promote regional resources allocation and flow, and continuously improve and upgrade the investment environment quality.


INTRODUCTION
The "Belt and Road" cooperation initiative was proposed by General Secretary Xi Jinping in 2013, which aims to strengthen interconnection with countries along the route under historical symbol of the ancient Silk Road, and build a community of interests, destiny and responsibility characterized by political mutual trust, economic integration, and cultural tolerance. The white paper "The Vision and Actions for Promoting the Joint Construction of the Silk Road Economic Belt and the 21st Century Maritime Silk Road" proposes "to jointly build a smooth, safe and efficient transportation channel under the support of central cities along the route with key ports on the sea as nodes." It can be seen that the development and construction of node cities in the "Belt and Road" strategy carries great significance. The development and construction of a city is closely related to the investment quality.
Introduction of high-quality resources can effectively improve the local investment structure, thereby driving economic development.
In fact, the unevenness and heterogeneity in development level of cities along the Maritime Silk Road has caused gaps in investment environment and investment attraction effectiveness among cities. So, what are the development laws for different node cities in time and space? What are the factors that affect and restrict the differences in investment environment? This paper takes 15 maritime Silk Road node cities covered by "Belt and Road" as the research objects. By constructing a dynamic evaluation index system for the investment environment, investment environment of maritime Silk Road node cities is objectively evaluated by projection pursuit cluster analysis. The innovation of this paper is that for the first time, projection pursuit cluster analysis is used to dynamically evaluate the investment environment of important node cities on the Maritime Silk Road, and on this basis, suggestions for optimizing the investment environment of node cities are proposed.

Evaluation index system
Investment environment refers to the general term for various social, economic, natural, and cultural factors surrounding the main body of construction investment that can affect the investment process and results. Evaluation of urban investment environment should depend on corresponding indexes. This paper conducts research from the six dimensions of innovation environment, human resources, economic environment, government behavior, social development, and opening to the outside world, and uses innovation performance, economic development level, industrial structure, financial service environment, labor quality, labor cost, input capacity, cultural service level, medical service environment, transportation level, informatization level, environmental status, import and export, investment as the first-level indexes to further establish corresponding second-level indexes. The evaluation index system is shown in Table 1. Fixed asset investment (100 million yuan) x4 Per capita disposable income of urban residents (yuan) x5

Data processing
This paper takes projection pursuit method to process the indexes of each city for analysis. Projection Pursuit is a statistical method used to analyze and process high-dimensional data. By minimizing a certain projection index, it finds the optimal projection direction that can reflect the structure or characteristics of the original high-dimensional data, and projects the high-dimensional data to low-dimensional space for analysis. In the case of big data dimension, its structure or characteristics are Industrial structure The proportion of secondary industry in GDP (%) x6 The proportion of tertiary industry in GDP (%) x7 usually presented in multiple projection directions. In this way, projection tracking can use dimensionality reduction methods to find the projection direction that reflects the data structure, and exclude interference effect produced by data in the projection direction irrelevant with the structure. Therefore, by projection pursuit, it is possible to effectively discover the structure and characteristics of high-dimensional values, obtain the total feature value of the evaluation target, and clearly reflect the value of each element.

Data source
This Yearbook" of each city. Individual missing data was supplemented by averaging method.

2013-2017 urban investment environment evaluation index analysis
According to the steps of projection pursuit evaluation model, MATLAB genetic algorithm toolbox is used to process the original data about the investment environment of 15 maritime Silk Road node cities. The evaluation results of the investment environment evaluation index of each city in 2013-2017 are shown in Table 2.
Seen from investment environment evaluation index and development trend of each city, it can be known from Table 2

Vector analysis of the optimal projection direction
According to the projection pursuit cluster model, the optimal projection direction vector is calculated using the MATLAB genetic algorithm toolbox, with results shown in Table 3  According to the optimal projection direction vector from 2013 to 2017, the evaluation value of each index dimension is obtained, as shown in Figure 1 and