A method for dividing and identifying industrial development stages based on characteristic statistical measurement

Abstract

The present invention discloses a method for dividing and identifying industrial development stages based on characteristic statistical measurement, including: constructing an industrial stage development characteristic indicator system; performing statistical measurement on the industrial stage development characteristic indicators: collecting, processing and statistically measuring indicator data, and identifying industrial development stages based on this division, wherein the statistical measurement first divides the industrial development stages through clustering measurement, and then identifies the industrial development stages through structural measurement, and obtains the final division and identification results of the industrial development stages based on the statistical measurement; the present invention constructs an industrial stage development characteristic indicator system, performs statistical measurement on the indicator system, and identifies the rail transit industry development stages based on this division, and obtains corresponding conclusions.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial planning, and in particular to a method for dividing and identifying industrial development stages based on characteristic statistical measurement. Background Art

[0002] In the process of development, industries often present different development characteristics at different stages, that is, the evolution of industrial development stages is also accompanied by phased changes in development characteristics. The phased changes in development characteristics are an important external manifestation of the evolution of industrial stages.

[0003] As an evolutionary system, the development process of industry is often manifested in the form of stage evolution. In the research related to the division and identification of industrial development stages, most scholars will give priority to quantitative analysis methods to make preliminary judgments on industrial development stages to ensure the accuracy and objectivity of the research results. At the same time, related research will also divide and identify industrial development stages from different research perspectives. The mainstream research direction abroad is the stage division and identification method based on industrial organization. In earlier studies, scholars generally divide and identify industrial development stages by observing the micro-behavior of enterprises, such as the absolute number of enterprise entities entering the industry and the absolute number of the previous period. Domestic scholars often cut in from the perspective of industrial scale, such as using the output growth rate, that is, the change in the annual average growth rate compared with the industry average growth rate as the basis for the division and identification of industrial development stages. At the same time, due to the complexity of industrial development and the data missing problem caused by the long time span, quantitative analysis will have limitations, so the assistance of qualitative analysis is also very necessary in the division and identification of industrial stages.

[0004] Quantitative methods include:

[0005] 1. Fitting curve method: The fitting curve method is an important quantitative method for identifying the division of industrial development stages, which is mainly divided into two forms: Logistic curve and Gompertz curve.

[0006] (1) Logistic curve model:

[0007] The Logistic curve is a common S-shaped curve, which was originally proposed by Pierre when studying population growth trends. According to the theory of industrial life cycle, the evolution of industrial development stages generally goes through four stages: formation stage, growth stage, maturity stage and decline stage. The general shape also presents an S-shaped curve, and a large amount of empirical evidence shows that there are similarities between industrial evolution and biological evolution. Therefore, this model has been widely used in the development stage research of various industries such as environmental protection and agriculture. The basic expression of the Logistic curve is: where y tis the index value of the tth period, generally referring to production or sales; a and b are undetermined parameters; K is the saturation value, indicating that with the growth of time, the growth rate of the industry scale changes from relatively slow in the formation period to gradually accelerate, and after entering the mature period, the growth rate of the industry scale slows down, and finally reaches the saturation value, and the industry enters the recession period. Therefore, by analyzing the current relevant data of the industry, we can get the current development stage of the industry.

[0008] (2) Gompertz curve model: This curve model is usually used to describe the cyclical process of the development of things from germination, growth to saturation, and is applied to the study of trend changes of phenomena. The basic expression is: y t =Ka bt (K>0), and its variable meaning is the same as the Logistic curve. In this curve model, the values ​​of the unknown parameters a and b will determine the final shape of the curve. By determining the estimated values ​​of a and b of the research industry, the development stage of the industry can be identified.

[0009] 2. Indicator judgment method: The indicator judgment method refers to selecting specific indicators based on the research industry, such as industry growth rate, profit margin, market share, price-earnings ratio, revenue growth rate, etc., calculating the current value of the indicator, and then comparing it with specific standards to preliminarily determine the stage the industry is in. For example, by comparing the current industry growth rate with the previous industry growth rate or the industry average growth rate, the comparison results can be used to determine whether the industry is in the growth and expansion stage or the recession stage.

[0010] 3. Ordered cluster analysis method: Cluster analysis refers to the analysis process of counting and grouping physical or abstract objects into multiple classes composed of similar objects, so that individuals in the same category have the highest possible homogeneity, while the categories have the highest possible heterogeneity. Ordered clustering or optimal sample segmentation rule is a cluster analysis method that considers the order in the sample and follows the principle that the order cannot be disrupted when classifying the sample. At present, ordered clustering based on industrial development characteristics has been widely used in the research of industrial stage division and identification. Zhang Kunsheng and other scholars studied the cyclical fluctuations of the real estate industry based on ordered cluster analysis. Feng Lina and others took the development of China's network economy as the research object and used ordered clustering as the research method to identify the stage division of China's network economy.

[0011] Qualitative methods: Qualitative research on industrial development stages is generally supplemented by quantitative classification results. Usually, the preliminary stage judgment of quantitative analysis is combined with the characteristic differences of the industry at each stage described by the industry life cycle theory to determine the stage of industrial development. At the same time, the development experience of the leading countries and regions of the industry is analyzed, and the development status of the same industry in different countries and regions is compared to determine the stage of the research industry.

[0012] Quantitative methods are generally used as the preferred method for industrial development stage division and identification because of their objective accuracy. In the fitting curve method, the total indicators of industrial development are usually selected as the analysis object. This is because Western economics believes that total indicators are an important reflection of the status of industrial development and the core content of studying the relationship between industry and market. Therefore, the fitting curve method generally studies the law of changes in total indicators such as total growth or occurrence over time, and measures them to obtain the results of stage identification and division. However, total indicators are not the only standard for measuring industrial development. The process of industrial development is also accompanied by changes in many other development characteristics. Therefore, the fitting curve method with total indicators as the analysis object is difficult to fully reflect the richness of the connotation of industrial development. The indicator judgment method compares a specific single indicator with the standard value to derive the development stage of the industry, which is also slightly lacking in comprehensiveness. From the above analysis, it can be seen that the development characteristics of the industry are very rich and multi-dimensional. At the same time, the development characteristics of industries at different development stages often have stage differences, and the stage differences of development characteristics are often regarded as an important external manifestation of the industrial development process. Therefore, it is reasonable and comprehensive to divide and identify the development stages of the rail transit industry by statistically measuring the development characteristics. Summary of the invention

[0013] In order to solve the problems existing in the prior art, the purpose of the present invention is to provide a method for dividing and identifying industrial development stages based on characteristic statistical measurement. The present invention constructs a stage-by-stage development characteristic indicator system for the industry, performs statistical measurement on the indicator system, and identifies the industrial development stages based on this division and draws corresponding conclusions.

[0014] To achieve the above purpose, the technical solution adopted by the present invention is: a method for dividing and identifying industrial development stages based on feature statistical measurement, comprising the following steps:

[0015] Step 1: Construct an indicator system for the characteristic development of the industry at different stages: Select the characteristic indicators for the development of the industry at different stages by combining literature review and text analysis, and construct a corresponding indicator system;

[0016] Step 2, statistically measure the characteristic indicators of industrial stage development: based on the industrial development characteristic indicator system, collect, process and statistically measure the corresponding indicator data, and identify the industrial development stage based on this division, wherein the statistical measurement first divides the industrial development stage through clustering measurement, and then identifies the industrial development stage through structural measurement, and uses Fisher's optimal segmentation method to perform cluster analysis on the characteristics of the industrial development stage, and preliminarily divide the industrial development stage. At the same time, the entropy weight method is used to perform structural analysis on the industrial stage development characteristics, so as to locate the characteristics of each stage on the basis of the preliminary division of development stages, and identify the characteristics of the development characteristic structure of each stage; based on the statistical measurement, the final division and identification results of the industrial development stage are obtained.

[0017] As a further improvement of the present invention, the step 1 of selecting characteristic indicators of industry stage development through literature combing specifically includes:

[0018] By reviewing and sorting out relevant literature, combining the three perspectives of industry scale, industrial organization and industrial technology, and comprehensively considering the characteristics of the industry under study, a targeted industrial development characteristic indicator system is established.

[0019] As a further improvement of the present invention, the step 1 of selecting the characteristic indicators of the staged development of the industry through text analysis includes: further identifying the development characteristics of the industry by mining and analyzing the text of relevant news reports, specifically including the following steps:

[0020] (1) Text collection: Using the Chrome browser plug-in Web-Scraper, we crawled relevant news reports on industrial development according to the text collection principles and set corresponding crawling rules;

[0021] (2) Word segmentation: Use Python's Chinese word segmentation package Jieba to segment the text to be processed; at the same time, in order to obtain meaningful text information, stop words and filter words are processed on the segmented words;

[0022] (3) Word frequency statistics and word frequency cloud drawing: Use the wordcloud package in Python to draw the word frequency cloud of industrial development characteristics based on the processed text;

[0023] (4) Keyword extraction based on TF-IDF: The TF-IDF algorithm is used to identify and extract keywords for industrial development. The calculation formulas for word frequency TF and inverse document frequency IDF are as follows:

[0024]

[0025]

[0026] TF-IDF=TF×IDF

[0027] Among them, N{t i ∈T} is the feature word t in the text T i The number of occurrences, T n is a summary of feature words, |D| is the total number of documents in the training corpus, m{t i ∈D m} contains t i The number of documents.

[0028] As a further improvement of the present invention, in step 2, the Fisher optimal segmentation method specifically includes the following steps:

[0029] (1) Define the diameter of the class:

[0030] Suppose a class G contains samples {X (i), X (i+1) ,...X (j)}(j>i), denoted as G={i,i+1,...j}, the mean vector expression of this class is:

[0031]

[0032] D(i,j) is used to represent the diameter of this type, and the diameter calculation formula is:

[0033]

[0034] (2) Define the classification loss function:

[0035] b(n,k) is used to represent a method of dividing n ordered samples into k categories:

[0036] G1={i1,i1+1,...,i2-1},

[0037] G2={i2,i2+1,...,i3-1}, ...

[0038] G k ={i k ,i k +1,...,n}

[0039] The dividing point is 1=i1<i2<...<i k <n=i k +1-1 (i.e. i k =n+1);

[0040] The loss function of the above classification method is defined as:

[0041]

[0042] When n and k are fixed, the smaller L[b(n,k)] is, the smaller the sum of squares of the deviations of each category is, and the closer it is to the optimal solution; split. Seek a split method b(n,k) that minimizes the loss function L[b(n,k)], and denote this split method as p[n,k];

[0043] (3) Recursive formula:

[0044]

[0045]

[0046] The above two recursive formulas are the core, and their connotation is that if we want to find the optimal segmentation result of n samples into k classes, we should build on the optimal segmentation of j-1 (j=2,3,...,n) samples into k-1 classes;

[0047] (4) Method for finding the optimal solution:

[0048] If the number of classifications k (1 < k < n) is known, find the classification method p (n, k) so that it is minimized in terms of the loss function. The method is as follows:

[0049] First find the point j k , so that the recursive formula reaches a minimum, that is:

[0050] L[p(n,k)]=L[P(j k -1,k-1)]+D(j k ,n)

[0051] So we get the kth class G k ={i k ,i k-1 +1,...,j k-1}, then find j k-1 , so that it satisfies:

[0052] L[P(j k-1 ,k-1)]=L[P(j k -1,k-2]+D( jk-1 ,j k-1 )

[0053] Get the k-1th class G k-1 ={i k-1 ,i k-1 +1,...,j k -1}, we get all classes G1,G2,...,G K , which is the optimal solution, namely P(n,k)={G1,G2,...,G K}.

[0054] As a further improvement of the present invention, an empirical analysis of the Fisher optimal segmentation method is also included, which specifically includes the following steps:

[0055] (1) Standardization of indicator data:

[0056] The indicators are standardized to eliminate the influence of units and dimensions on the results. The z-scores normalization method is used to standardize the data. The formula is as follows:

[0057]

[0058] where xij is the original data, x′ ij is the standardized data, is the average of the jth index, s ij is the standard deviation;

[0059] (2) Determine the optimal number of categories:

[0060] Use Matlab software to solve the minimum loss function corresponding to each classification number, and make a trend chart of its relationship with the classification number;

[0061] (3) Determine the division results:

[0062] After determining the number of categories, the optimal split point is obtained through MATLAB software programming according to the optimal solution method, and all categories of the optimal solution are obtained.

[0063] As a further improvement of the present invention, in step 2, using the entropy weight method to perform structural analysis on the staged development characteristics of the industry includes:

[0064] Firstly, the entropy weight method is used to weight the established industry stage development characteristic indicator system, and the characteristic manifestation value of the indicator is calculated according to the comprehensive weight value. According to the divided stages, the characteristic manifestation values ​​of the years included in each stage are averaged to calculate the development characteristic manifestation value of each stage. Finally, based on the measurement results, the stage changes of the development characteristic structure are analyzed.

[0065] As a further improvement of the present invention, the entropy weight method is used to weight the established industry stage development characteristic indicator system, which specifically includes the following steps:

[0066] (1) Positive indicators;

[0067] (2) Data standardization: Select the maximum and minimum value standardization method, the formula is as follows:

[0068] Positive indicators:

[0069] Negative indicators:

[0070] Among them, xij' is the standardized value of the indicator data, xij is the original value of the indicator data, min,j is the minimum value of the jth indicator in the original data of the indicator, x max,j The maximum value of the jth indicator of the original data of the indicator;

[0071] (3) Calculate the information entropy E of the jth indicator j :

[0072]

[0073] where pij is the proportion of the i-th and j-th evaluation units in the total amount of the indicator, p ij , the calculation formula is as follows:

[0074]

[0075] (4) Calculate the indicator weight w ij :

[0076]

[0077] According to the above calculation steps, the weight results of the staged development characteristic indicator system are obtained.

[0078] The beneficial effects of the present invention are:

[0079] Starting from the richness of the connotation of industrial development, the present invention constructs a multi-dimensional industrial stage development characteristic indicator system, and uses the Fisher optimal segmentation method and the entropy weight method in the ordered clustering method to perform statistical measurement on the indicator system. Based on the statistical measurement work, the evolution process of industrial development stages is studied more comprehensively. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 A word frequency statistics chart of texts related to the rail transit industry in an embodiment of the present invention;

[0081] Figure 2 A schematic diagram of a word frequency cloud of characteristic texts of the rail transit industry development drawn in an embodiment of the present invention;

[0082] Figure 3 It is a schematic diagram of the staged development characteristic index system of the rail transit industry constructed in an embodiment of the present invention;

[0083] Figure 4 This is a graph showing changes in the minimum loss function in an embodiment of the present invention;

[0084] Figure 5 This is a graph showing the percentage of characteristic manifestation values ​​of stage I of the development of the rail transit industry in an embodiment of the present invention;

[0085] Figure 6 This is a graph showing the percentage of characteristic manifestation values ​​of the rail transit industry development stage II in an embodiment of the present invention;

[0086] Figure 7 This is a graph showing the percentage of characteristic manifestation values ​​of stage III of the rail transit industry development in an embodiment of the present invention;

[0087] Figure 8 It is a diagram showing the stage changes of the characteristic structure of the development of the rail transit industry in an embodiment of the present invention. DETAILED DESCRIPTION

[0088] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0089] Example

[0090] A method for dividing and identifying industrial development stages based on characteristic statistical measurement. This embodiment takes the development of my country's rail transit industry from 2005 to 2020 as the research object, constructs a staged development characteristic index system for the rail transit industry, performs statistical measurement on the index system, identifies the development stage of the rail transit industry based on this division, and draws corresponding conclusions, specifically including:

[0091] 1. Construction of the characteristic indicator system of the staged development of the rail transit industry:

[0092] 1.1 Basis for construction: The selection of characteristic indicators of industrial stage development is the basis for the research on industrial development stages. This example will combine the two methods of literature combing and text analysis to select characteristic indicators of rail transit industry stage development and construct a corresponding indicator system:

[0093] 1.1.1 Literature review method:

[0094] By reviewing and sorting out relevant literature, it is found that most of the research on the stage of industrial development is carried out from the stage changes of industrial scale, industrial organization and industrial technology. According to the theory of industrial life cycle, the evolution process of industrial development stage generally goes through four stages: formation, growth, maturity and decline, accompanied by stage changes in development characteristics. From the perspective of external characteristics and internal characteristics, the external characteristics of industrial development are often manifested as growth and expansion in the time dimension and transfer and diffusion in the spatial dimension. In terms of internal characteristics, it is mainly manifested in the changes in the proportion of internal factor inputs in the industry, which are concentrated in the three aspects of industrial scale, industrial organization and industrial technology. At the same time, a large number of theoretical empirical evidences also show that the stage development of the industry is mainly reflected through changes in these three aspects.

[0095] Industrial scale is generally the preferred perspective for domestic scholars to study the stage of industrial development. The development characteristics selected from the perspective of industrial scale are generally at the macro level, and the current stage of the industry is grasped by analyzing indicators such as industrial output value and its growth rate, product output and its changes, and sales revenue within the industry. As an early mainstream research perspective, the development characteristics selected from the perspective of industrial organization generally focus on the micro level, observing the changes in the behavior and quantity of enterprises in the industry to judge and identify the stage of the industry, and taking the changes in indicators such as the behavior of leading enterprises, industrial barriers, the number of enterprises and their marginal number, and industrial utilization rate as important bases for the division and identification of industrial stages. Industrial technology identifies the stage of industrial development from the changes in industrial technology level, technical efficiency, product added value, etc. Xu Shuang et al., from the perspective of industrial technology, established an indicator system including technology progress index and degree of technological progress to identify and analyze the development stage of the tin industry. However, since industrial technology indicators are difficult to quantify, there are few related studies on establishing indicators from the perspective of industrial technology.

[0096] In related research, most scholars have combined the three perspectives of industrial scale, industrial organization and industrial technology, comprehensively considered the characteristics of the industry under study, and established a targeted indicator system for industrial development characteristics. Zhang Li and Wei Wei focused on industrial scale and industrial organization, selected seven indicators such as passenger turnover, passenger transport volume, and number of routes, and used principal component analysis to construct an industrial comprehensive utility level function to identify the stages of China's civil aviation industry, dividing it into three stages. Chen Qiangqiang and other scholars combined these three perspectives to construct an indicator system including market structure, industrial scale, government role, and technological progress and efficiency when studying the life cycle of China's wine industry. Based on this, they divided and identified the stages of the development of the wine industry; Wang Jiaxing constructed the growth index of strategic emerging industries through market scale, scientific and technological innovation capability, profitability and other indicator factors, and further identified and judged the development stage of my country's strategic emerging industries. Shi Jinna and Liu Hui comprehensively considered industrial scale and industrial organization, constructed an indicator system including the number of corporate competitors, industrial scale and market structure, analyzed Chengdu's exhibition industry, and concluded that the industry has entered a period of rapid growth. Lv Qiuju sorted out and analyzed multiple indicators such as the industrial organization, market structure, scientific and technological innovation and government role of the walnut industry in Lin'an, Hangzhou, Zhejiang, and divided the development stages of the industry.

[0097] 1.1.2 Text Mining

[0098] As a key national strategic emerging industry, the development status and trend of the rail transit industry has always been the focus of national planning and media attention. Therefore, this embodiment further identifies the development characteristics of the rail transit industry by mining and analyzing texts such as relevant news reports.

[0099] (1) Text collection:

[0100] Using the Chrome browser plug-in Web-Scraper, according to the text collection principles and the corresponding crawling rules, the relevant news reports on the development of the rail transit industry on the China Rail Transit Network were crawled, and a total of 61,541 words of text were crawled.

[0101] (2) Word segmentation:

[0102] Words are composed of characters, which can be used independently and convey the basic language unit of content. However, it is difficult to divide and identify words by delimiters like English expressions, so it is necessary to perform word segmentation on the text. This embodiment uses Python's Chinese word segmentation package Jieba to perform word segmentation on the text to be processed. At the same time, in order to obtain meaningful text information, the above word segmentation also needs to be processed with stop words and filter words.

[0103] (3) Word frequency statistics and word frequency cloud drawing:

[0104] Word frequency analysis is a method to reflect the main content of a text based on the frequency of words and phrases in the text. After the above word segmentation and stop word filtering, the word frequency statistics of the text are as follows: Figure 1 The word frequency cloud of the text is an intuitive reflection of the frequency of words. The font size of the words can be used to preliminarily determine the frequency of words and phrases in the text. The wordcloud package in Python is used to draw the word frequency cloud of the text with the characteristics of the development of the rail transit industry based on the processed text, as shown in the figure below. Figure 2 shown.

[0105] According to the word frequency statistics of the text and the word frequency cloud presentation effect, it can be preliminarily concluded that technological innovation, design and construction, service quality, equipment manufacturing, etc. are the "keywords" for the development of the rail transit industry, which provides an analytical basis for the construction of the subsequent staged development characteristic indicator system of the rail transit industry.

[0106] (4) Keyword extraction based on TF-IDF:

[0107] The more prominent the characteristics of certain industrial development are, the more times the characteristic terms appear in related articles. However, the size of the word frequency cannot fully reflect the importance of words or phrases, because some words appear frequently in various articles, that is, they appear frequently in the entire corpus, so using such words as text keywords is obviously not representative. Therefore, it is also necessary to cooperate with the TF-IDF (term frequency-inverse document frequency) algorithm to identify and extract keywords for the development of rail transit industry. The calculation formulas for term frequency (TF) and inverse document frequency (IDF) are as follows:

[0108]

[0109]

[0110] TF-IDF=TF×IDF

[0111] Among them, N{t i ∈T} is the feature word t in the text T i The number of occurrences, T n is a summary of feature words, |D| is the total number of documents in the training corpus, m{t i ∈D m} contains t i The number of documents.

[0112] Using the sklearn package in python, the preprocessed word segmentation is converted into a word frequency matrix, the frequency of each word is calculated and all text keywords in the word bag are extracted. Finally, the TF-IDF value is calculated according to the word frequency matrix, and the ranking is from large to small. The ranking results are shown in Table 1. It can be seen from the observation results that in the process of the development stage evolution of the rail transit industry, "wisdom", "innovation" and "technology" have always been the "keywords" of its development characteristics and the focus of the development of the industry. At the same time, the presentation of key features such as "cooperation upgrade", "deepening transformation" and "ecological sustainability" also further reflects that in the process of the development of the rail transit industry, its core and the external interactive relationship are also constantly improving and upgrading, thereby promoting the good and sustainable development of the industry.

[0113] Table 1 TF-IDF value ranking results

[0114]

[0115] At the same time, in the process of constructing the phased development characteristic indicator system, this embodiment also follows the following principles:

[0116] (1) Systematic principle. The indicators in the indicator system should have a certain logical relationship, and there should be a certain subordinate relationship between the upper and lower indicators. The indicators at the same level should be independent and complementary to each other, so that the indicator system can reflect the overall situation of industrial development from different aspects and levels.

[0117] (2) Availability and operability principle. The availability and statistical feasibility of indicator data are important factors to consider in indicator selection, so as to provide accurate and comprehensive data support for subsequent statistical measurement and analysis. At the same time, the data is required to have a certain degree of operability, so as to facilitate subsequent quantitative research.

[0118] (3) Scientific principle. The selected indicators must have a certain scientific theoretical basis, clear purpose, accurate definition, and be able to accurately measure the industrial development process. At the same time, they must also meet the data processing software's requirements for data.

[0119] (4) Dominant principle. Due to the complexity of industrial development, the presentation of its development characteristics must be multi-dimensional. The comprehensiveness of the indicator system is an inherent principle for indicator selection. However, including development characteristics will complicate the analysis process and make the analysis results unrepresentative. Therefore, it is necessary to select representative and dominant development characteristic indicators.

[0120] (5) Dynamic principle. The dynamic sustainability of industrial development requires that the indicators of its development characteristics also be dynamic. When selecting indicators, it is necessary to evaluate the applicability of the indicators from a dynamic perspective and update outdated indicators in a timely manner.

[0121] 1.2. Build results:

[0122] According to the analysis of relevant literature, relevant studies often select specific indicators of industrial development characteristics from three perspectives: industrial scale, industrial organization and industrial technology. At the same time, text mining results show that "technological innovation", "design and construction", "service quality" and "equipment manufacturing" are the "keywords" for the development of the rail transit industry. Therefore, based on the previous analysis and the principles of indicator selection, the construction results of the rail transit industry phased development characteristic indicator system are as follows: Figure 3 shown.

[0123] 1.2.1. Investment and construction:

[0124] The rail transit industry presents a chain-like structure. As an important upstream link of the industrial chain and the foundation of industrial development, the investment and construction of the industry preliminarily reflects the scale of the industry and its future development space. Taking into account the indicator selection principles and industrial characteristics, the number of exploration and design units, the mileage of new railway lines put into operation, the total output value of civil engineering and construction industry, the national railway fixed asset investment amount, and the operating mileage of traditional railways and urban rail transit are selected as specific indicators of industrial development investment and construction.

[0125] 1.2.2, Scientific research and innovation:

[0126] As an important driving force for promoting the innovation-driven development strategy, the scientific research and innovation of the rail transit industry is an important reflection of the progress of industrial technology, and it is also an important indicator of the evolution of the industrial development stage from the perspective of industrial technology. As the key content of "2025", the rail transit equipment manufacturing industry is not only an important force in promoting the advanced equipment manufacturing industry, but also a backbone force in promoting the overall innovation and development of the rail transit industry chain. As the technical core of the industrial chain, the scientific research and innovation of the rail transit equipment manufacturing industry is quite representative of the overall scientific research and innovation of the rail transit industry. The rail transit equipment manufacturing industry mainly includes special equipment, general equipment, transportation equipment, etc. Therefore, combined with the above analysis and considering the availability of data, this embodiment selects the R&D expenditure of industrial enterprises above a certain scale, the full-time equivalent of R&D personnel of industrial enterprises above a certain scale, and the number of patent applications of industrial enterprises above a certain scale as specific indicators of the industrial scientific research and innovation dimension.

[0127] 1.2.3. Service supply:

[0128] The industrial output effect is one of the important indicators for measuring the evolution of the industrial development stage, and the service supply quality is the key manifestation of the industrial output effect. As a national basic industry, the rail transit industry's main function is to meet the national production and people's travel needs. Therefore, this embodiment mainly evaluates the service supply quality of the rail transit industry from the downstream transportation link of the rail transit industry chain, and selects railway passenger and freight volume, passenger and freight turnover, average passenger and freight transportation distance, urban rail transit passenger volume, the number of rail transit vehicles and the number of railway transportation personnel as specific indicators of the service supply quality dimension.

[0129] 1.2.4. Industrial linkage:

[0130] As a basic industry, the rail transit industry has the characteristics of strong radiation effect and great industrial influence. The rail transit industry is in different stages of development, and the effect of its interaction with the outside world will also be different, and the difference in the interaction effect also reflects the difference in the scale of the industry to a certain extent. In terms of data acquisition, traditional industrial linkage indicators are generally calculated through the input-output table released by the National Bureau of Statistics. However, since the input-output table is not announced every year, and the industrial linkage indicators have strong volatility, it is difficult to complete the data through mathematical methods such as linear regression. Since this embodiment is a statistical measurement of stage data, the data integrity requirements are relatively high. At the same time, the linkage effect of the rail transit industry is mainly reflected in the driving effect of the supply of transportation services on the economy along the line. Therefore, considering the above factors, this embodiment selects the industrial contribution rate of the transportation industry (incremental added value of the transportation industry / incremental GDP) as the specific indicator of the industrial linkage dimension.

[0131] 2. Statistical measurement of characteristic indicators of rail transit industry development at different stages:

[0132] The industrial development process is not only reflected in the evolution of the industry through the entire development stage, but also in the phased changes in the characteristics of industrial development. There are obvious differences in the characteristics of industrial development at different stages. The phase change of development characteristics is an important external manifestation of industrial development and an important entry point for studying industrial development. Therefore, this embodiment divides and identifies the development stages of the rail transit industry based on the statistical measurement results of industrial development characteristics. At the same time, because there are large gaps in the data on industrial development characteristics in earlier years, considering the integrity of the indicator data and the accessibility of the data, this embodiment takes the development of my country's rail transit industry from 2005 to 2020 as the research object and divides and identifies its development stages.

[0133] On the basis of the above-mentioned characteristic indicator system for the development of the rail transit industry, the corresponding indicator data are collected, processed and statistically measured, and the development stages of the rail transit industry are identified based on this division. In terms of characteristic statistical measurement, taking into account the temporal orderliness of industrial development and the complex multidimensionality of development characteristics, the measurement method adopts the Fisher optimal partitioning method. This method is an important method of ordered cluster analysis, which aims to cluster ordered samples so that the sample differences within each category are minimized and the sample differences between categories are maximized. At the same time, the entropy weight method is used to structurally analyze the stage-by-stage development characteristics of the industry, so as to locate the characteristics of each stage on the basis of the preliminary division of the development stage and identify the characteristics of the characteristic structure of each stage. Based on the above statistical measurement, the final division and identification results of the development stages of the rail transit industry are obtained.

[0134] 2.1 Clustering measurement - division of industrial development stages:

[0135] 2.1.1 Data source and compilation:

[0136] The data of the staged development characteristics of the rail transit industry are mainly sourced from the "China Statistical Yearbook" published by the National Bureau of Statistics, the "Railway Statistical Bulletin" published by the National Railway Administration, and the "Annual Statistics and Analysis Report on Urban Rail Transit" published by the China Urban Rail Transit Association. The results of the index data compilation are shown in Tables 2 to 5:

[0137] Table 2 Characteristic indicator data of investment and construction dimension

[0138]

[0139] Table 3 Characteristic index data of scientific research and innovation dimensions

[0140]

[0141] Table 4. Service supply dimension characteristic indicator data

[0142]

[0143] Table 5. Data on characteristic indicators of industrial linkage dimensions

[0144]

[0145] 2.1.2 Basic principles of Fisher optimal segmentation method:

[0146] Fisher's optimal segmentation method can cluster samples while ensuring the temporal order of sample sequences. This embodiment uses Fisher's optimal segmentation method to cluster the characteristics of industrial development stages and preliminarily divide industrial development stages. Fisher's optimal segmentation method is one of the ordered clustering methods. For n ordered samples, it is possible to find 2 n -1 division method. Among all the division methods, find one that minimizes the difference between samples in each segment and maximizes the difference between each segment. The sum of squared deviations represents the degree of difference between similar samples. The idea of ​​finding the optimal division is to determine the optimal number of classifications and the optimal division point through computational analysis, so that the sum of squared deviations between segments is minimized after segmentation.

[0147] The steps of Fisher's optimal segmentation method are as follows:

[0148] (1) Define the diameter of the class:

[0149] Suppose a class G contains samples {X (i) ,X (i+1) ,...X (j)}(j>i), denoted as G={i,i+1,...j}, the mean vector expression of this class is:

[0150]

[0151] Use D(i,j) to represent the diameter of this type. The commonly used diameter calculation formulas are:

[0152]

[0153] (2) Define the classification loss function:

[0154] b(n,k) is used to represent a method of dividing n ordered samples into k categories:

[0155] G1={i1,i1+1,...,i2-1},

[0156] G2={i2,i2+1,...,i3-1}, ...

[0157] G k ={i k ,ik +1,...,n}

[0158] The dividing point is 1=i1<i2<...<i k <n=i k +1-1 (i.e. i k =n+1);

[0159] The loss function of the above classification method is defined as:

[0160]

[0161] When n and k are fixed, the smaller L[b(n,k)] is, the smaller the sum of squares of the deviations of each category is, and the closer it is to the optimal segmentation. Therefore, we seek a segmentation method b(n,k) that minimizes the loss function L[b(n,k)], and denote this segmentation method as p[n,k];

[0162] (3) Recursive formula:

[0163]

[0164]

[0165] The above two recursive formulas are the core, and their connotation is that if we want to find the optimal segmentation result of n samples into k classes, we should build on the optimal segmentation of j-1 (j=2,3,...,n) samples into k-1 classes;

[0166] (4) Method for finding the optimal solution:

[0167] If the number of classifications k (1 < k < n) is known, find the classification method p (n, k) so that it is minimized in terms of the loss function. The method is as follows:

[0168] First find the point j k , so that the recursive formula reaches a minimum, that is:

[0169] L[p(n,k)]=L[P(j k -1,k-1)]+D(j k ,n)

[0170] So we get the kth class G k ={i k ,i k-1 +1,...,j k-1}, then find j k-1 , so that it satisfies:

[0171] L[P(j k-1 ,k-1)]=L[P(j k -1,k-2]+D( jk-1 ,jk-1 )

[0172] Get the k-1th class G k-1 ={i k-1 ,i k-1 +1,...,j k -1}. A similar method is used to obtain all classes G1, G2, ..., G K , which is the optimal solution, namely P(n,k)={G1,G2,...,G K}.

[0173] 2.1.3. Empirical analysis of Fisher optimal segmentation method:

[0174] (1) Standardization of indicator data:

[0175] Different types of indicator data have differences in units and dimensions, so the indicators need to be standardized to eliminate the impact of units and dimensions on the results. This embodiment uses the z-scores normalization method to standardize the data. The formula is as follows:

[0176]

[0177] where x ij is the original data, x′ ij is the standardized data, is the average of the jth index, s ij is the standard deviation.

[0178] (2) Determine the optimal number of categories:

[0179] According to the above-mentioned Fisher optimal segmentation principle, the minimum loss function corresponding to each classification number is solved by Matlab software, and the following is made: Figure 4 The relationship between it and the number of classifications is shown in the figure.

[0180] from Figure 4 It can be seen that the larger the number of classifications, the smaller the minimum loss function, that is, the smaller the difference between samples within a segment, the larger the difference between segments. However, the increase in the number of classifications will increase the number of sample analysis objects, and the significance of classification will be reduced. It can be seen from the figure that when the number of classifications is 3, the minimum loss function has an obvious inflection point and the minimum loss function changes significantly. Therefore, 3 is selected as the optimal number of classifications, that is, the industrial development process is divided into three stages.

[0181] (3) Determine the division results:

[0182] After determining the number of categories, the optimal split point is obtained through MATLAB software programming according to the optimal solution method, and all the categories of the optimal solution are obtained. The preliminary division results of the stage are shown in Table 6:

[0183] Table 6 Results of industrial development stage division

[0184]

[0185] 2.2. Structural measurement - identification of industrial development stages:

[0186] The characteristic structure refers to the combination of the development characteristic indicators in each stage. The difference of the characteristic structure is to study the stage changes of industrial development characteristics from an overall perspective. On the basis of the development stages that have been divided above, the characteristics of the development characteristic structure of each stage are identified, and the characteristics of each stage are identified and positioned. First, the entropy weight method is used to weight the industrial stage development characteristic indicator system established in the previous article. The characteristic manifestation values ​​of the four indicators of investment and construction, scientific research and innovation, service supply and industrial linkage are calculated according to the comprehensive weight value. According to the stages divided above, the characteristic manifestation values ​​of the years included in each stage are averaged to calculate the development characteristic manifestation values ​​of each stage. Finally, according to the measurement results, the stage changes of the development characteristic structure are analyzed.

[0187] 2.2.1. Determination of weights of development characteristic indicators:

[0188] In order to measure and analyze the characteristic structure of the staged development of the industry more objectively and scientifically, the entropy weight method is used to assign weights to the development characteristic indicator system. The entropy weight method is an objective value assignment method. According to the degree of variation of each indicator, the entropy weight of each indicator is calculated using information entropy, and then the weight of each indicator is corrected according to the obtained entropy weight, and a more objective indicator weight is further obtained. Generally speaking, if the information entropy of an indicator is smaller, the greater the degree of variation of the indicator value, the more information is provided, the greater the role it plays in the comprehensive evaluation, and the greater the weight, and vice versa. The steps of entropy weighting method are as follows:

[0189] (1) Positive indicators:

[0190] Different indicators have different meanings. The indicator types generally include: the larger the better, the smaller the better, and the optimal point. The indicator type is determined according to the effect of the indicator value on the target layer. In order to facilitate subsequent evaluation, all indicators are generally converted into the larger the better. However, the indicators selected in this embodiment are all the larger the better, so this step can be omitted.

[0191] (2) Data standardization:

[0192] Because the units and magnitudes of the indicator values ​​are different, they need to be standardized before comparison. This step uses the maximum and minimum value standardization method. The formula is as follows:

[0193] Positive indicators:

[0194] Negative indicators:

[0195] Among them, xij' is the standardized value of the indicator data, xij is the original value of the indicator data, min,j is the minimum value of the jth indicator in the original data of the indicator, x max,j The maximum value of the jth indicator of the original data of the indicator;

[0196] (3) Calculate the information entropy E of the jth indicator j :

[0197]

[0198] where p ij is the proportion of the i-th and j-th evaluation units in the total amount of the indicator, p ij , the calculation formula is as follows:

[0199]

[0200] (4) Calculate the index weight w ij :

[0201]

[0202] According to the above calculation steps, the weight results of the staged development characteristic indicator system are shown in Table 7:

[0203] Table 7 Weights of the index system for characteristic staged development of industries

[0204]

[0205] 2.2.2 Feature structure measurement analysis

[0206] By multiplying the weight of each indicator in the rail transit industry stage development characteristic indicator system with the standardized data of each year, the characteristic manifestation values ​​of the four indicators of investment and construction, scientific research and innovation, service supply and industrial linkage are obtained, as shown in Table 8:

[0207] Table 8 Characteristic appearance values

[0208]

[0209]

[0210] Based on the previous stage division, the stage feature manifestation value is calculated, and the results are shown in Table 9 below:

[0211] Table 9 Stage characteristic manifestation values

[0212]

[0213] The percentage of the development characteristics of the rail transit industry is shown in the figure below: Figure 5-Figure 8 As shown, it reflects the development characteristics and structure of each stage. Figure 5 It can be seen that in the industrial development stage I, the display values ​​of investment and construction, scientific research and innovation, service supply and industrial linkage accounted for 28%, 13%, 54% and 8% respectively; Figure 6 It can be seen that in the industrial development stage II, the proportions of investment and construction, scientific research and innovation, service supply and industrial linkage are 40%, 17%, 41% and 2% respectively; Figure 7 It can be seen that in the industrial development stage III, the proportions of investment and construction, scientific research and innovation, service supply and industrial linkage are 42%, 20%, 37% and 1% respectively. It can be seen that the proportion of the characteristic indicators of each dimension will change in stages, that is, the development characteristic structure of the rail transit industry shows a trend of stage changes. The changes are as follows: Figure 8 shown.

[0214] Based on the above results, the phased changes in the proportion of the characteristic manifestation value of the rail transit industry development are analyzed as follows:

[0215] (1) The proportion of investment and construction characteristics first increased significantly and then stabilized:

[0216] It can be seen from the change diagram of the characteristic structure of industrial development that the proportion of investment and construction characteristic manifestation value has increased significantly during the evolution of the industry from stage I to stage II. This is because the evolution process from stage I to stage II is in a critical promotion period. During this development period, the promotion of urbanization and the adjustment and upgrading of industrial structure are the key points of economic and social development in this stage. At the same time, the further implementation of the strategy of expanding domestic demand is also an important work content at that time, and this specific era development background will bring about the continued stable growth of passenger and freight transportation demand and the optimization and upgrading of transportation structure. In terms of freight transportation, coal transportation continues to maintain the pattern of transporting coal from the west to the east and from the north to the south. The demand for small-volume, high-value, and high-frequency freight transportation shows a rapid growth trend. At the same time, the demand for bulk freight transportation such as oil and ore remains high. In terms of passenger transportation, demand is becoming more diversified and the demand level is becoming richer. The demand for passenger transportation in all aspects, including urban, intercity and international passenger transportation, is growing rapidly. Therefore, providing safer, more convenient, comfortable and efficient transportation services has become the key to industrial development. It can be seen that the growth and improvement of transportation demand in quantity and quality require the follow-up of rail transit network construction to a certain extent, so as to further strengthen investment and construction activities and further manifest the characteristics of investment and construction. At the same time, the urban rail transit industry has become a sunrise industry, and its investment and construction efforts have been further strengthened. The "Review of Urban Rail Transit Development during the Period" released by the National Development and Reform Commission pointed out that since the construction of the first subway line in 1965, my country's urban rail transit industry has entered a rapid development stage after decades of development. "The mileage of lines opened in the past 5 years has doubled compared with the past 45 years, and the passenger transportation volume has shown a continuous upward trend; the construction scale has continued to grow, and the cumulative investment has exceeded one trillion yuan. At the same time, many cities have formed a multi-line network operation pattern. During the period, the number of lines put into operation reached 2019 kilometers, with an average annual increase of 404 kilometers. As of the end of 2015, 26 cities in the region had opened urban rail transit operations, with a total of 116 lines and a total length of 3618 kilometers. Among them, there are 11 cities with lines of more than 100 kilometers, 20 cities with more than two urban rail transit lines, and a certain scale of operating line networks have been built in many cities across the country. At the same time, the diversification of investment models has also helped to increase the scale of investment and construction to a certain extent, thereby driving the emergence of investment and construction characteristics.

[0217] As the industry transitions from stage II to stage III, the proportion of investment and construction value tends to stabilize, but is still on a growth trend. There are two main reasons behind this: First, the railway network construction has shifted from high-speed development to high-quality development track. During the "period, the full completion of the main skeleton of the "four vertical and four horizontal" high-speed railway and the continuous improvement of the main channels of the "eight vertical and eight horizontal" high-speed railway have continuously guided the construction of the railway network to a high-quality direction. The opening and operation of a large number of new lines with important production strategic significance, such as the intelligent Beijing-Zhangjiakou High-speed Railway and the Haoji Railway, a heavy-duty channel for transporting coal from north to south, have continuously optimized the national road network layout and significantly improved the quality of the road network. Intelligent railways have gradually become a construction focus, and the railway network in the central and western regions has been continuously improved, and hubs and supporting facilities have been continuously strengthened. In the process of ensuring the stable construction of the railway network, the development focus has gradually shifted to the construction of an intelligent and efficient railway network, rather than simply pursuing the number of road networks. It can be seen that the construction of the railway network has gradually transformed from pursuing the construction of new lines to focusing on high-quality development such as road network upgrading and structural optimization. Therefore, the proportion of the apparent value of industrial investment and construction has become increasingly stable. Second, with the continuous advancement of urbanization and the continuous improvement and improvement of urban construction, urban rail transit has entered a stage of rapid development, which has caused the proportion of the apparent value of investment and construction characteristics to continue to show an upward trend.

[0218] (2) The proportion of scientific research and innovation characteristics has steadily increased:

[0219] From the diagram of the changes in the characteristic structure of industrial development, it can be seen that with the evolution of the industrial development stage, the proportion of the value of scientific research and innovation characteristics has shown a steady upward trend. Because the rail transit equipment manufacturing industry in the midstream is the technical core of the entire rail transit industry chain, the reason why the proportion of the value of the scientific research and development characteristics of the rail transit industry has shown a steady upward trend is mainly analyzed from the perspective of the development of the rail transit equipment manufacturing industry.

[0220] First of all, the global industrial development trend. The world is experiencing a new round of technological innovation, the most representative of which is the industrial innovation in intelligent manufacturing, energy materials and information networks. Under the influence of this wave, the global rail transit equipment manufacturing industry is bound to usher in a new round of all-round transformation opportunities, which will also guide the future new rail transit to a safer, more efficient, more environmentally friendly and intelligent direction. Under such an industrial background, my country's rail transit equipment manufacturing industry has been included in the top ten development areas in 2025. It is believed that the technological progress and innovation of this industry are of great significance to my country's regional economic development and the transformation and upgrading of the manufacturing industry. At the same time, the "Guidelines for the Implementation of High-end Equipment Innovation Project" issued by the Ministry of Industry and Information Technology further pointed out that in order to achieve the steady and comprehensive development of my country's rail transit equipment manufacturing industry, efforts should be made in the following four aspects: intelligent equipment and application, key technology research and development, rail transit train control system and rail transit network vehicles. Therefore, the promotion of the industrial environment and the national development strategy has enabled the rail equipment manufacturing industry to continuously increase its scientific and technological research and development efforts and consciously enhance its technological innovation capabilities.

[0221] Secondly, the current development status of my country's rail transit equipment manufacturing industry. At present, the scientific research and innovation achievements of my country's rail transit industry are remarkable. At present, an equipment manufacturing system that runs through multiple production links has been formed, which can comprehensively and efficiently complete tasks such as design and development, manufacturing and production, and testing services. It has independent research and development and innovation capabilities, advanced and complete supporting equipment, and production and operation have shown a certain scale. But at the same time, the industry's technical shortcomings in scientific research and innovation are still very obvious. For example, in the research and development and manufacturing of components and key parts, there is still a certain gap with the world's first-class level. The research and development technology of some key system components is still limited and dependent on other countries. For example, the manufacturing and development of bearing systems has been monopolized by German companies, and the research and development and manufacturing processes of some signal systems still rely on the introduction and reference of foreign advanced technologies. It can be seen that the road to innovation and development of my country's rail transit equipment manufacturing industry is still arduous and long, and the investment in its scientific research and innovation will inevitably be further increased.

[0222] Based on the above analysis, it can be predicted that in the future development, the scientific research and innovation value of the development characteristics of the rail transit industry will still account for a large proportion and show a continuous upward trend.

[0223] (3) The proportion of service supply characteristics is on a downward trend:

[0224] Service supply characteristics often correspond to the mid- and downstream transportation service links. From the stage change diagram of industrial development characteristics structure, it can be seen that the proportion of service supply characteristics presents a downward trend. Combined with the analysis of relevant literature, the reasons for this trend are as follows:

[0225] First, we should have an in-depth understanding of the characteristics of the rail transit industry chain structure. By combing through the relevant literature on the development of the rail transit industry, we can see that the early research on the development stage of the rail transit industry mainly focused on the transportation service link. For example, when Yang Jie et al. studied the development trend of the railway industry of Shenhua Group, they only used freight turnover as a parameter indicator to construct an industrial growth model and conduct a stage analysis of the group's industrial development. When Li Xuemei et al. discussed the development trend of my country's railway industry based on the theory of industrial life cycle, they mainly selected transportation indicators as the research indicators. After statistical analysis of the indicators, they proposed the ideal development trend of the industry and divided the development stages of my country's railway industry. It can be seen that in the early research on the development stage of the railway industry or rail transit industry, the indicators selected were generally transportation indicators. This is not only because the statistics of transportation indicators are relatively comprehensive, the data missing degree is small, and the coverage period is large, but also because the early understanding of the chain structure of the rail transit industry is still not deep enough. The research on the development of the rail transit industry mainly focuses on the downstream transportation service link. However, in recent years, the emphasis on industrial chain construction in the national development plan has further promoted the improvement and development of the rail transit industry chain, and the chain form and value of the rail transit industry have been gradually recognized and valued. The Third Plenary Session of the 18th CPC Central Committee proposed to unleash the vitality of all labor, knowledge, technology, management, and capital, and to fully tap all sources of social wealth, accelerate the transformation of economic development methods, accelerate the construction of an innovative country, and promote more efficient, fairer, and more sustainable economic development. In terms of specific implementation, taking Changzhou as an example, the local municipal party committee and government proposed the decision to "accelerate the pace of industrial chain construction". Focusing on the city's industrial development priorities, further implement the chain-filling, chain-building, and chain-strengthening to form Changzhou's characteristic industrial chain and comprehensively enhance the overall competitiveness of the industry. The local rail transit industrial chain is the key content of the construction of the ten major industrial chains.

[0226] Second, the improvement of the rail transit industry chain has become increasingly effective. The mid- and upper-stream links, including design consulting, scientific and technological research and development, equipment manufacturing, etc., have always been a weak but important link in the industry chain. In particular, the development of the rail transit equipment manufacturing industry is an important aspect of enhancing the competitive advantage of the manufacturing industry and promoting the high-quality development of the manufacturing industry. The government, enterprises and society have jointly made efforts to "strengthen the chain", "fill the chain" and "extend the chain". Various cities have gradually improved the rail transit industry chain, gradually changed the situation of focusing on a single link in industrial construction, especially the operation and maintenance service link, gradually improved the upstream, mid- and downstream industrial links, continued to consolidate the strong links, supplemented and improved the weak links, and all links throughout the industrial chain, from R&D planning, design consulting, construction and construction to equipment manufacturing, operation and maintenance services and value-added services, have been improved and strengthened to varying degrees.

[0227] However, it is worth noting that although the proportion of the apparent value of service supply characteristic indicators shows a downward trend, the supply of operation and maintenance services in the industrial chain, especially transportation services, is still a key link in the operation of the industry. The decline in the proportion of its apparent value is a side reflection of the further completion of the industrial chain and the increasing balance of various links, and is not equivalent to a decline in the importance of the operation and maintenance service link to industrial development.

[0228] (4) The proportion of the value of the industrial linkage characteristics fluctuates steadily:

[0229] The industrial linkage index mainly reflects the value-added service link of the industrial chain. As the terminal link of the rail transit industry, the development of this link is still in the exploratory development stage, so it shows a stable fluctuation trend. For example, as a key construction project in the value-added service link, the Chengdu Municipal Government has issued relevant policies and measures in recent years, such as the "Implementation Opinions of the Chengdu Municipal People's Government on the Comprehensive Development of Rail Transit Stations", "Special Plan for the Comprehensive Development of Rail Transit Stations", and "Guidelines for the Integrated Urban Design of Chengdu Rail Transit Stations", which provide guidelines for the top-level design of rail transit construction funds and TOD comprehensive development, and provide specific operational implementation methods for TOD comprehensive development, further assisting the construction of the value-added service link of the rail transit industry. With the continuous and rapid development of rail transit, the rapid advancement of new urbanization, the pursuit of high-quality development in cities, and the compound promotion of infrastructure investment and financing reforms, the value-added service link derived from the rail transit industry, which is in the development and exploration stage, is about to usher in its development climax.

[0230] 3. Analysis of the identification results of the development stages of rail transit industry:

[0231] The cluster measurement work mainly divides the rail transit industry into stages, while the structural measurement work mainly identifies and locates the characteristics of each stage, that is, identifies the presentation characteristics of its development characteristics at each stage. Based on the above statistical measurement results of the staged development characteristics of the industry, the development history of my country's rail transit industry from 2005 to 2020 is divided into three stages. The results of the identification of the industrial development stage division are as follows:

[0232] 3.1. Phase I (2005-2012): “Chain Formation” Phase:

[0233] The time span of the development phase I is relatively large, from 2005 to 2012. In order to build a modern railway network with a reasonable layout, wide coverage, high efficiency, convenience, safety and economy, the State Council reviewed and approved the National Medium- and Long-Term Railway Network Plan in 2004. The starting point of phase I is the first year of the implementation of the Medium- and Long-Term Railway Network Plan, which is also a critical period for the implementation of the "National Railway Network Plan". The construction of the railway network focuses on expanding the scale, improving the structure, improving the quality, rapidly expanding the transportation capacity, and rapidly improving the equipment level, so that the railway transportation capacity can meet the needs of national economic and social development. In the review of the development of the railway "National Railway Plan" released by the National Railway Administration, it is mentioned that during the "National Railway Plan" period, a number of high-speed railways were put into operation and built, which greatly shortened the distance between regions, saved transportation time and cost, and further promoted economic and trade cooperation between regions, promoted balanced development between regions, and provided strong guarantees for the coordinated development of regions and the advancement of industrialization. The construction and improvement of western railways have improved the infrastructure conditions in the western region, enhanced the region's self-development capabilities, and accelerated the pace of economic and social development in old, remote, poor and remote areas. Railways have always put national and social interests first, and always ensured their backbone role in the transportation of key materials, public welfare transportation, and emergency transportation, which effectively guaranteed the stable operation of the national economy and the production and living needs of the people. At the same time, the construction and operation of urban rail transit has also entered the initial development stage. Many cities have built and opened new lines, and the construction of urban rail transit lines has driven the rapid and sustainable development of cities. Therefore, this development stage is mainly characterized by the comprehensive improvement of the transportation capacity of the railway network, the optimization of the transportation network structure, and the rapid development of the investment, construction and operation of urban rail transit. At this time, the characteristics of the rail transit industry chain are that the development focus is mainly concentrated on the operation and maintenance links of the middle and lower reaches, and the service supply characteristics are obvious, but the chain form of the overall industry is still not obvious, and some industrial links need to be improved.

[0234] 3.2. Phase II (2013-2016): “Supplementing the supply chain” phase:

[0235] The time span of development stage II is from 2013 to 2016. By observing its development characteristic structure, it can be seen that this stage is in the transitional "chain filling" stage. The proportion of investment and construction value representing the upstream of the industrial chain has increased significantly, while the proportion of scientific research and innovation value representing the midstream of the industrial chain has also shown a stable upward trend. The proportion of service supply value representing the midstream and downstream of the industrial chain has shown a further downward trend, while the industrial linkage index is in a fluctuating state. Analysis shows that the rail transit industry in stage II has further improved its chain form, and the cooperation between various links in the industrial chain has become closer. The progress in the scientific and technological research and development link has further promoted the evolution of the rail transit industry stage. Development stage II is in a critical period of planning. The "" period is an important strategic opportunity period for my country to build a strong equipment manufacturing country. The transformation and upgrading of the mid-to-high-end equipment manufacturing industry is the key content of promoting the development of strategic emerging industries and seizing strategic development opportunities. As an important part of the mid-to-high-end equipment manufacturing industry, the equipment manufacturing industry in the midstream of the rail transit industry chain has also further implemented the key tasks of industrial transformation and upgrading. At the same time, the in-depth implementation of the "Medium- and Long-Term Railway Network Planning", the development of the economy and society, and the acceleration of urbanization and industrialization have put forward higher requirements for the rail transit industry, especially its core technology rail transit equipment manufacturing industry. Therefore, in the development stage II, the scientific research and innovation indicators representing the midstream of the industrial chain are further revealed, the industrial chain is further improved and balanced, and the innovative development of the midstream links is a significant feature.

[0236] 3.3. Phase III (2017-2020): “Strong Chain” Phase:

[0237] The time span of development stage III is 2017-2020. By observing its development characteristic structure, it can be seen that this stage is in the later "strengthening chain" stage. The proportion of investment and construction characteristics representing the upstream of the industrial chain has increased slightly, while the proportion of scientific research and innovation indicators representing the midstream of the industrial chain has maintained a steady upward trend. The proportion of service supply in the midstream and downstream of the industrial chain has maintained a certain downward trend, while the industrial linkage index is in a fluctuating state. The industrial chain at this stage further promotes the innovative and efficient development of key industrial links on the basis of pursuing balanced development, thereby further promoting the improvement of the overall competitiveness of the industrial chain.

[0238] First, the proportion of investment and construction characteristics representing the upstream is further increasing. The revised version of the "National Railway Medium- and Long-Term Plan" in 2016 proposed that "by 2020, a number of major national railway landmark projects will be completed and put into operation, and the railway network will reach 150,000 kilometers, including 30,000 kilometers of high-speed railways; covering more than 80% of large cities geographically; by 2025, the railway network will reach about 175,000 kilometers, including 38,000 kilometers of high-speed railways. It can be foreseen that national policy support and market recovery will continue to promote the expansion of my country's railway investment and construction. At the same time, under the background of national capacity reduction, deleveraging, and structural adjustment, the role of urban rail transit in driving the national economy and urban construction has been further recognized. Urban rail transit construction is gradually extending to second- and third-tier cities, and urban rail transit construction has entered a period of rapid development. As for the scientific research and innovation indicators representing the midstream links, their apparent value The proportion remains stable and upward. It is clear that the rail transit equipment manufacturing industry is the key content of my country's high-end equipment manufacturing industry. Under the guidance of favorable national policies and the strong market demand, my country's rail transit equipment manufacturing industry is entering a period of rapid growth. By 2020, the market demand for the sales output value of the rail transit equipment industry will exceed 650 billion yuan. This provides broad prospects for the sustained and rapid development of the rail transit equipment industry. Technological innovation results continue to emerge, becoming an important representative of my country's high-end equipment entering the international market. The proportion of service supply and industrial linkage indicators that characterize the middle and lower reaches shows a stable and fluctuating trend. As a traditional key link in the rail transit industry, operation and maintenance services continue to provide high-quality and efficient transportation services for the operation of the national economy and production and life activities with the upgrading and improvement of the transportation network and the technological upgrading of transportation equipment.

[0239] This embodiment first preliminarily identifies the staged development characteristic indicators of the rail transit industry through relevant literature combing and text mining, and constructs the staged development characteristic indicator system of the rail transit industry according to the corresponding indicator selection principles. The Chinese rail transit industry from 2005 to 2020 is selected as the research object, and the data is collected and processed according to the constructed indicator system, and then the staged development characteristic indicator system is statistically measured: first, the Fisher optimal segmentation method is used to preliminarily divide the development stage of the rail transit industry, and then the entropy weight method is used to analyze the stage changes of the development characteristic structure, and the characteristics of each stage are located to further identify the development stage of the rail transit industry. Based on the above statistical measurement results, the division and identification results of the development stage of the rail transit industry are comprehensively obtained: it is divided into three stages: the "chain formation" stage (2005-2012), the "chain supplement" stage (2013-2016) and the "chain strengthening" stage (2017-2020).

[0240] The above-mentioned embodiments only express the specific implementation of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A method for dividing and identifying industrial development stages based on characteristic statistical measurement, characterized in that: The following steps are involved: Step 1: Construct the characteristic index system of the staged development of the rail transit industry: select the characteristic indexes of the staged development of the rail transit industry by combining literature review and text analysis, and construct the corresponding index system; The indicator system includes four primary indicators: investment and construction, scientific research and innovation, service supply and industrial linkage; Among them, the first-level indicator investment and construction consists of six second-level indicators: the number of exploration and design units, the total output value of civil engineering construction industry, the amount of railway fixed asset investment, the mileage of new railway lines put into operation, the national railway operating mileage, and the national urban rail operating mileage; The first-level indicator of scientific research and innovation is composed of three second-level indicators: R&D expenditure of industrial enterprises above designated size, full-time equivalent of R&D personnel of industrial enterprises above designated size, and number of patent applications of industrial enterprises above designated size; The first-level indicator service supply consists of nine second-level indicators, namely, railway passenger volume, urban rail passenger volume, railway freight volume, railway passenger turnover, railway freight turnover, average railway passenger distance, average railway freight distance, number of rail transit vehicles, and number of railway transport employees. The first-level indicator industrial linkage consists of one second-level indicator, industrial contribution rate; Step 2, statistically measure the characteristic indicators of industrial stage development: based on the characteristic indicator system of rail transit industry development, collect, process and statistically measure the corresponding indicator data, and identify the industrial development stage based on this division, wherein the statistical measurement work is to first divide the industrial development stage through cluster measurement, and then identify the industrial development stage through structural measurement, and use Fisher's optimal segmentation method to cluster the characteristics of the industrial development stage, preliminarily divide the industrial development stage, and use the entropy weight method to structurally analyze the characteristics of the industrial stage development, so as to locate the characteristics of each stage on the basis of the preliminary division of the development stage, and identify the characteristics of the characteristic structure of each stage; based on the statistical measurement, the final division and identification results of the industrial development stage are obtained; In step 2, the structural analysis of the staged development characteristics of the rail transit industry by using the entropy weight method includes: firstly, weighting the established staged development characteristic index system of the rail transit industry by using the entropy weight method, calculating the characteristic manifestation value of the index according to the comprehensive weight value, taking the average of the characteristic manifestation values ​​of the years included in each stage according to the divided stages, calculating the development characteristic manifestation value of each stage, and finally analyzing the staged changes of the development characteristic structure according to the measurement results; The method of using the entropy weight method to weight the established rail transit industry phased development characteristic indicator system specifically includes the following steps: (1) Positive indicators; (2) Data standardization: Select the maximum and minimum value standardization method, the formula is as follows: Positive indicators: Negative indicators: Among them, xij' is the standardized value of the indicator data, xij is the original value of the indicator data, min,j is the minimum value of the jth indicator in the original data of the indicator, x max,j The maximum value of the jth indicator of the original data of the indicator; (3) Calculate the information entropy E of the jth indicator j : where p ij is the proportion of the i-th and j-th evaluation units in the total amount of the indicator, p ij , the calculation formula is as follows: (4) Calculate the indicator weight w ij : According to the above calculation steps, the weight results of the rail transit industry stage development characteristic index system are obtained; In step 1, the characteristic indicators of the staged development of the rail transit industry are selected through literature review, including: By reviewing and sorting out relevant literature, combining the three perspectives of industry scale, industry organization and industry technology, and comprehensively considering the characteristics of the industry under study, a targeted industry development characteristic indicator system is established; The step 1 of selecting characteristic indicators of the staged development of the rail transit industry through text analysis includes: further identifying the development characteristics of the industry by mining and analyzing the text of relevant news reports, specifically including the following steps: (1) Text collection: Using the Chrome browser plug-in Web-Scraper, we crawled relevant news reports on the development of the rail transit industry according to the principles of text collection and collection and set corresponding crawling rules; (2) Word segmentation: Use the Python Chinese word segmentation package Jieba to segment the text to be processed; at the same time, in order to obtain meaningful text information, stop words and filter words are processed on the segmented words; (3) Word frequency statistics and word frequency cloud drawing: Use the wordcloud package in Python to draw the word frequency cloud of the rail transit industry development characteristics text based on the processed text; (4) Keyword extraction based on TF-IDF: The TF-IDF algorithm is used to identify and extract keywords for industrial development. The calculation formulas for word frequency TF and inverse document frequency IDF are as follows: TF-IDF=TF×IDF Among them, N{t i ∈T} is the feature word t in the text T i The number of occurrences, T n is a summary of feature words, |D| is the total number of documents in the training corpus, m{t i ∈D m } contains t i The number of documents; In step 2, the Fisher optimal segmentation method specifically includes the following steps: (1) Define the diameter of the class: Suppose a class G contains samples {X (i) ,X (i+1) ,...X (j) }, j>i, denoted as G={i,i+1,...j}, the mean vector expression of this class is: D(i,j) is used to represent the diameter of this type, and the diameter calculation formula is: (2) Define the classification loss function: b(n,k) is used to represent a method of dividing n ordered samples into k categories: G1={i1,i1+1,...,i2-1}, G2={i2,i2+1,...,i3-1}, ... G k ={i k ,i k +1,...,n} The dividing point is 1=i1<i2<...<i k <n=i k +1-1, i.e. i k =n+1; The loss function of the above classification method is defined as: When n and k are fixed, the smaller L[b(n,k)] is, the smaller the sum of squares of the deviations of each category is, and the closer it is to the optimal segmentation. We seek a segmentation method b(n,k) that minimizes the loss function L[b(n,k)], and denote this segmentation method as p[n,k]; (3) Recursive formula: The above two recursive formulas are the core, and their connotation is that if we want to find the optimal segmentation result of n samples into k classes, we should build on the optimal segmentation of j-1, j=2,3,...,n samples into k-1 classes; (4) Method for finding the optimal solution: If the number of classifications k, 1<k<n is known, find the classification method p(n,k) so that it is minimized in terms of the loss function. The method is as follows: First find the point j k , so that the recursive formula reaches a minimum, that is: L[p(n,k)]=L[P(j k -1,k-1)]+D(j k ,n) So we get the kth class G k ={i k ,i k-1 +1,...,j k-1 }, then find j k-1 , so that it satisfies: L[P(j k-1 ,k-1)]=L[P(j k -1,k-2]+D( jk-1 ,j k-1 ) Get the k-1th class G k-1 ={i k-1 ,i k-1 +1,...,j k -1}, we get all classes G1,G2,...,G K , which is the optimal solution, namely P(n,k)={G1,G2,...,G K }; It also includes an empirical analysis of Fisher's optimal partitioning method, which specifically includes the following steps: (1) Standardization of indicator data: The indicators are standardized to eliminate the influence of units and dimensions on the results. The z-scores normalization method is used to standardize the data. The formula is as follows: where x ij is the original data, x′ ij is the standardized data, is the average of the jth index, s ij is the standard deviation; (2) Determine the optimal number of categories: Use Matlab software to solve the minimum loss function corresponding to each classification number, and make a trend chart of its relationship with the classification number; (3) Determine the division results: After determining the number of categories, the optimal split point is obtained through MATLAB software programming according to the optimal solution method, and all categories of the optimal solution are obtained.

Images