Multi-dimensional comprehensive transportation index synthesis and stage division method

Abstract

The invention relates to a multi-dimensional comprehensive transportation index synthesis and stage division method, which belongs to the field of traffic transportation statistics and comprises the following steps: (1) season adjustment: performing season adjustment on railway, highway, waterway and civil aviation passenger and freight transportation indexes; (2) carrying out nondimensionalization; (3) weight measurement and calculation; (4) index dynamic synthesis: performing index cross-period dynamic synthesis through a Fisher ideal index method; and (5) stage division: introducing a Markov mechanism conversion method, and identifying the probability distribution of the stage state of each stage of the index based on the historical change rule of the index. The method has the advantages that an index synthesis method flow that firstly, a seasonal adjustment method and a dimensionless method are used for converting a traffic transportation time sequence into a relative quantity index, multi-dimensional weight synthesis is carried out, and finally, dynamic weight synthesis is realized by relying on a Fisher ideal index is provided; and the actual business requirements of multiple heterogeneity indexes, multiple evaluation dimensions and long-term dynamic weight synthesis are effectively met.

Description

Technical Field

[0001] The present invention relates to a method for synthesizing and stage-dividing a multi-dimensional comprehensive transportation index, and belongs to the field of transportation statistics. Background Art

[0002] In terms of index synthesis, generally quantitative or qualitative methods are used to determine the synthesis weights of different transportation modes under specific evaluation criteria. Among them, quantitative methods mainly evaluate from the perspectives of value quantity and physical quantity. In terms of value quantity, the measurement of the proportion of added value of different transportation modes is mainly realized through input-output tables. In terms of physical quantity, the measurement of the proportion of explainable variance is mainly realized through the proportion of passenger and freight volumes, turnover volume scale or through principal component analysis, factor analysis, etc.; qualitative methods mainly obtain the proportion of weights of various transportation modes through questionnaire surveys.

[0003] In terms of index stage division, it is mainly divided into two methods: division based on time period and division based on index change. Division based on time period is mainly applicable to specific policy implementation periods or planning periods and has strong purposefulness; division based on index change mainly subjectively sets the division criteria for each stage through the high and low levels of index values, maximum and minimum points, or the high and low levels of index growth rates.

[0004] First, most of the existing means adopt single-dimensional synthesis, which is difficult to objectively reflect the characteristics of the transportation industry. The existing technologies and common means mainly conduct the synthesis of the transportation volumes of multiple transportation modes from a single perspective such as economic value or physical quantity scale. Transportation is a basic, leading, strategic industry and an important service industry in the national economy. Transportation services have obvious dual attributes of economy and public welfare. Single-dimensional synthesis is difficult to fully reflect the dual attribute characteristics of transportation services. Second, most of the common methods adopt static weight weighted synthesis. When applied to the transportation production sequence over a long period, it is difficult to reflect the dynamic changes caused by the internal structural adjustment of indicators, resulting in a systematic deviation between the synthesized weight of indicators and the actual situation. Third, the existing methods are difficult to effectively synthesize heterogeneous output indicators such as passenger and freight services. In the quantitative method, the weights of different modes such as railway, highway, waterway, and civil aviation can only be obtained by measuring weights based on the input-output method, but it is difficult to obtain the weights of passenger and freight indicators of each mode in earlier years. The measurement method based on physical quantity is difficult to scientifically synthesize different dimensions of passenger and freight indicators. Although the qualitative method can obtain the weights of passenger and freight by mode, its accuracy is relatively weak. Fourth, in terms of index stage division, the division boundaries are mainly set in advance according to standards such as policy implementation periods and planning periods or artificially based on the characteristics of index changes. There is a lack of corresponding unified standards and is greatly affected by subjectivity, which in turn affects the development of subsequent applications such as economic cycle changes. Summary of the Invention

[0005] To overcome the defects of the prior art, the present invention provides a method for synthesizing and stage-dividing a multi-dimensional comprehensive transportation index. The technical solution of the present invention is as follows: A method for synthesizing and stage-dividing a multi-dimensional comprehensive transportation index, comprising the following steps: (1) Seasonal adjustment: Perform seasonal adjustment on the passenger and freight transportation indicators of railways, highways, waterways, and civil aviation; (2) Dimensionless processing: Standardize each indicator to eliminate the dimension difference; (3) Weight measurement: Measure and synthesize the comprehensive weights in different periods from three dimensions of economic growth contribution, market service output, and social value recognition; (4) Index dynamic synthesis: Perform cross-period dynamic synthesis of indicators through the Fisher ideal index method; (5) Stage division: Introduce the Markov regime switching method, identify the probability distribution of the stage states of each period of the index based on the historical change law of the index itself, and then realize stage division.

[0006] The specific content of step (2) is as follows: Select the indexation method, that is, divide the current value by the base period value, multiply the obtained ratio by 100, and the value range is [0, +∞]. The specific formula is: ; where yi is the relative score value of the current transportation volume completion scale of the i-th sub-sector.

[0007] The specific content of step (3) is as follows: 3.1 Measurement of the economic growth contribution weight: Determine the weights of passenger and freight transportation services of each transportation mode through the input-output table and enterprise financial data; 3.2 Measurement of the service output scale weight: Determine the market recognized value of passenger and freight transportation services of each transportation mode through market freight rates and transportation volumes; 3.3 Measurement of the social value recognition weight: Evaluate the non-economic value of each transportation mode through questionnaire surveys; 3.4 Synthesize the comprehensive weight.

[0008] The specific content of step 3.1 is as follows: Introduce the main business net profit of enterprise financial indicators related to the calculation of industry added value as a reference index for splitting. Using relevant statistical survey data of the business operation data of representative enterprises in the industry and combining with the enterprise business volume data, the net profit generated per unit of transportation volume can be realized. Based on this, the net profit of passenger and freight production in each industry can be estimated, and the obtained result can be used as an approximate ratio for splitting the added value of passenger and freight. The specific measurement formula for a single transportation mode is as follows: ; where, is the estimated net profit of the i-th sub-sector, is the net profit of the main business of the j-th surveyed enterprise in the i-th sub-industry, is the current transportation volume of the i-th sub-industry, is the transportation business volume completed by the j-th surveyed enterprise in the i-th sub-industry; The splitting method of the added value of passenger and freight transportation by each transportation mode is as follows: ; among them, is the split value of the added value of the i-th sub-industry, is another sub-industry in the transportation mode where the i-th sub-industry is located. For example, if i is road passenger transportation, then j represents road freight transportation, represents the actual added value of the transportation modes corresponding to sub-industries i and j. Obviously, is equal to and The weight distribution of the added value of passenger and freight transportation by each transportation mode is as follows: , where i = 1 to 8 respectively represent railway passenger transportation, road passenger transportation, waterway passenger transportation, civil aviation passenger transportation, railway freight transportation, road freight transportation, waterway freight transportation, and civil aviation freight transportation.

[0009] Specifically, step 3.2 is as follows: The method for calculating the weight of the output scale of passenger and freight transportation services by each transportation mode is as follows: ; among them, is the transportation revenue of the i-th sub-industry.

[0010] Specifically, step 3.3 is as follows: The method for calculating the weight of social value recognition by each transportation mode is as follows: ; among them, is the weight of social value recognition of each sub-industry, is the first-level survey weight of the corresponding passenger and freight attributes of the sub-industry, is the second-level survey weight of the sub-industry in the corresponding passenger and freight fields and the corresponding transportation mode.

[0011] Specifically, step 3.4 is as follows: The weights of each sub-industry are weighted among three dimensions by the geometric mean method. The specific calculation method is as follows: ; where i takes values from 1 to 8, representing the passenger and freight transportation modes of railways, roads, waterways, and civil aviation respectively.

[0012] Specifically, step (4) is as follows: According to the basic expression form of the Fisher ideal index, the expression form of the index's intertemporal composite index is finally determined as follows: ; among them, Denote the comprehensive weight of the $i$-th transportation mode in the $j$-th time stage. The weights in different time periods are mainly obtained by setting specific years or cycles to carry out regular weight measurement. Denote in the time stage the relative score value of the transportation volume completed by the $i$-th transportation mode in the $t$-th period.

[0013] The specific content of the step (5) is as follows: Step 5-1, establish an identification model: Taking the two-state stage division of high growth rate and low growth rate as an example, in different stages, the year-on-year growth rate sequence of the synthesized comprehensive index its own volatility and its correlation with the growth rate of related indicators will show different characteristics. Establish a Markov regime-switching model: ; ; ; wherein, is the year-on-year growth rate of the comprehensive index in the $t$-th period, is the correlation coefficient between the growth rate of the comprehensive index and the growth rate of related indicators in different state stages, is the growth rate sequence of related indicators, the historical growth rate sequence of the comprehensive index or the growth rate sequence of other external indicators, is the residual, reflecting the degree of random fluctuation of the comprehensive index, and allowing differences in amplitude in different state stages; Step 4-2, calculate the probabilities of different state stages in each period. Iteratively estimate the probabilities in each period through Hamilton filtering, and continuously approximate and calculate the filtering probabilities corresponding to various states with the actually observed information. The calculation process is as follows: First, calculate the initial probability expectations of the index growth rate in each state stage at the initial stage ($t = 1$): ; ; wherein, represents the state transition probabilities corresponding to the Markov chain process, and the matrix is all the information that can be obtained in the period $t - 1$ and is helpful for stage judgment; Secondly, through iterative calculation period by period, for the $t$-th period, based on the known information in the $(t - 1)$-th period, estimate the preliminary probability expectations of each cycle state: ; Furthermore, update the state probability expectations of the current period with the newly added information in the $t$-th period, that is, calculate the proportion of the likelihood values in each state stage weighted by the initial probability expectations as the true probabilities corresponding to the assumptions in each stage of the current period: ; Among them, the function f is the probability distribution function corresponding to the growth rate value of the comprehensive index in the t-th period; Finally, iterate t according to the above steps until the cycle state of each period is obtained in the current period, the T-th period The corresponding filtering probability expression; Step 4-3, calculate the parameters, and through maximum likelihood estimation of the weighted models of each period, synchronously obtain the parameter estimation results of the entire parameter set, and at the same time the state stage of each period The corresponding filtering probability is obtained synchronously: ; Among them, is the likelihood function of the comprehensive index growth rate in the state stage j for the parameter set ; in addition, the expected duration of each state period continuation is calculated through the obtained state transition probability

[0014] The advantages of the present invention are: 1. It proposes a method flow for index synthesis that first uses the seasonal adjustment method and the dimensionless method to convert the transportation time series into relative quantity indicators, and through multi-dimensional weight synthesis, finally relies on the Fisher ideal index to achieve dynamic weight synthesis, effectively meeting the actual business needs of multi-heterogeneous indicators, multi-evaluation dimensions, and long-term dynamic weight synthesis.

[0015] 2. It proposes a method for dividing the relative weights of different transportation modes from three dimensions: economic growth contribution, transportation service market value, and public recognition, and comprehensively uses multiple data such as relevant statistics, monitoring, and questionnaires to measure the weights of economic growth contribution, transportation service market value, and public recognition of different transportation modes.

[0016] 3. It proposes a method for combining multi-dimensional weights such as economic growth contribution, transportation service market value, and public recognition into a comprehensive weight.

[0017] 4. It proposes to apply the Markov regime switching method to the measurement of the state probability of each period of the transportation comprehensive index, and then realize the quantitative identification of the index stage division. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is the flow schematic diagram of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0019] ​The present invention will be further described below in conjunction with specific embodiments, and the advantages and features of the present invention will become clearer with the description. However, these embodiments are merely exemplary and do not constitute any limitation to the scope of the present invention. Those skilled in the art should understand that without departing from the spirit and scope of the present invention, modifications or substitutions can be made to the details and forms of the technical solutions of the present invention, but such modifications and substitutions all fall within the protection scope of the present invention.

[0020] See Figure 1 , the present invention relates to a method for synthesizing and stage dividing a multi-dimensional comprehensive transportation index, including the following steps: (1) Seasonal adjustment: Perform seasonal adjustment on the passenger and freight transportation indicators of railways, highways, waterways, and civil aviation; the scale of passenger and freight transportation volume in transportation is greatly affected by seasons. Before carrying out index synthesis, it is necessary to perform seasonal adjustment on the passenger and freight transportation indicators of each railway, highway, waterway, and civil aviation to effectively reduce the influence of seasonal factors on the index trend and periodic fluctuations and improve data comparability.

[0021] (2) Dimensionless processing: Standardize each indicator to eliminate the dimension difference; (3) Weight measurement: Measure and synthesize the comprehensive weights in different periods from three dimensions of economic growth contribution, market service output, and social value recognition; to realize the evaluation of the comprehensive weights in three dimensions of the economic growth contribution, service output scale, and social value recognition of passenger and freight transportation services, it is first necessary to measure the weights of each dimension one by one. On this basis, the geometric mean method is used to perform secondary weighting and standardization on the three groups of weights to obtain the comprehensive weights.

[0022] (4) Index dynamic synthesis: Perform cross-period dynamic synthesis of indicators through the Fisher ideal index method; this method can take into account the changes in weights, quantities, and their structures in the base period and reporting period of the index at the same time, and can pass the time reversal test, factor reversal test, and circular test, and can effectively meet the continuity requirements after index synthesis.

[0023] (5) Stage division: Introduce the Markov regime switching method, identify the probability distribution of the stage states of each period of the index based on the historical change law of the index itself, and then realize stage division. Whether the index is in a high-speed or low-speed stage is the primary question that stage division needs to answer. Secondly, it is necessary to answer how likely it is for the current stage to continue or change. At the same time, considering that the transportation industry is closely related to the macro economy, it is also necessary to meet the actual need of introducing reference indicators of related industries. For this reason, this solution realizes the probability measurement of the index state based on the historical change law of the numerical values of each period of the index and the historical change law of related indicators (if necessary), and evaluates the possibility of index stage change through the state transition probability.

[0024] The specific content of step (2) is as follows: Before conducting index synthesis, it is necessary to standardize the different dimensions of each index. Considering that the synthesis of long-term cycle indicators requires ensuring the historical continuity of the dimensionless indicators, the exponential method is selected, that is, dividing the current value by the base period value, multiplying the obtained ratio by 100, and the value range is [0, +∞). The specific formula is: ; where yi is the relative score value of the current transportation volume completion scale of the i-th sub-industry.

[0025] The specific content of step (3) is as follows: 3.1 Measurement of the contribution weight of economic growth: Determine the weights of passenger and freight transportation services of each transportation mode through the input-output table and enterprise financial data; that is, the relative importance of the contribution of passenger and freight transportation services of each transportation mode to the industry added value. This requires obtaining the composition of the added value of passenger and freight transportation of each transportation mode from the input-output table. Since 2017, it can be extracted from sub-industries such as railway passenger transportation, railway freight transportation and transportation auxiliary activities, urban public transportation and highway passenger transportation, road freight transportation and transportation auxiliary activities, water passenger transportation, water freight transportation and transportation auxiliary activities, air passenger transportation, air freight transportation and transportation auxiliary activities in the input-output table. However, in the input-output tables before 2017, the industry classification only included railway transportation, road transportation, water transportation and air transportation, and it was necessary to further split them into the proportion of passenger transportation added value and the proportion of freight transportation added value. Therefore, it is necessary to introduce the main business net profit of enterprise financial indicators, which is highly relevant to the calculation of industry added value, as a reference index for splitting. Using the relevant statistical survey data of the operating data of representative enterprises in the industry and combining with the enterprise business volume data, the net profit generated per unit of transportation volume can be realized. Based on this, the net profit of passenger and freight transportation production by sub-industry can be calculated, and the obtained result can be used as an approximate ratio for splitting the added value of passenger and freight transportation.

[0026] 3.2 Measurement of the weight of service output scale: Determine the market recognized value of passenger and freight transportation services of each transportation mode through market freight rates and transportation volumes; measure the weight of the service output scale of passenger and freight transportation services of each transportation mode, that is, the market recognized value of passenger and freight transportation services of each transportation mode, which is also the proportion of total transportation revenue under market conditions. Calculate the total revenue of each transportation mode mainly through the method of multiplying the freight rate by the transportation volume. The transportation volume of each sub-industry comes from the statistical indicators of the industry statistical report system, and the freight rate of each sub-industry It is necessary to measure through multiple methods from multiple channels. Among them, the unit prices of railway passenger and freight transportation are mainly calculated based on the passenger and freight operating revenues and traffic volumes in the financial statements of the National Railway Group; the unit price of highway passenger transportation is mainly measured through the mileage of typical passenger transport lines and the record-filing of ticket prices; the unit price of highway freight transportation is mainly measured through freight platforms and questionnaires of truck drivers; the unit price of waterway passenger transportation is mainly measured through the mileage of typical waterway passenger transport lines and the record-filing of ticket prices; the unit price of waterway freight transportation is mainly measured through the main business revenues and traffic volumes in the financial statements of typical waterway freight enterprises, and is verified through questionnaire surveys; the unit prices of civil aviation passenger and freight transportation are mainly measured through the passenger and freight revenues and the scales of passenger and freight transportation in the financial reports of listed airlines.

[0027] 3.3 Measuring the weights of social value recognition. Evaluate the non-economic values of each transportation mode through questionnaire surveys; measure the weights of social value recognition for each transportation mode, that is, the non-economic values of each transportation mode, including social values such as public welfare services, green energy conservation, safety, and service guarantee for basic people's livelihoods, which are mainly evaluated through questionnaire surveys of the general public. The questionnaire survey realizes the measurement of relative importance through a two-stage method. In the first stage, the public is surveyed on the relative importance and proportion of the social public welfare values of passenger transportation services and freight transportation services to obtain the first-layer weights, where m is 1 or 2, representing passenger transportation services and freight transportation services respectively, and the sum of the two is 1; on this basis, the relative importance and proportion of the social public welfare values of railways, highways, waterways, and civil aviation in passenger transportation services and freight transportation services are respectively surveyed to obtain the second-layer weights, where n is from 1 to 4, representing railways, highways, waterways, and civil aviation respectively, and the sum of the weights of the 4 modes in both passenger and freight transportation is 1.

[0028] 3.4 Synthesize the comprehensive weights. The geometric mean method is mainly selected because it is relatively less affected by extreme values of single-dimensional weights and has strong stability. After geometric averaging, the weights of each sub-industry are difficult to meet the condition of summing up to 100%. For the convenience of subsequent indicator synthesis, it is necessary to further correct the weights of each sub-industry through the equal-proportion expansion coefficient method to achieve the goal of summing up to 100%.

[0029] The specific content of step 3.1 is as follows: Introduce the main business net profit of enterprise financial indicators related to the calculation of industry added value as the reference index for splitting. Using the relevant statistical survey data of the operating data of representative enterprises in the industry and combining the business volume data of enterprises, the net profit generated per unit of transportation volume can be realized. Based on this, the net profit of passenger and freight production by sub-industry can be calculated, and the result can be used as an approximate ratio for splitting the added value of passenger and freight transportation; the specific measurement formula for a single transportation mode is as follows: ; where is the estimated net profit of the i-th sub-industry, is the main business net profit of the j-th surveyed enterprise in the i-th sub-industry, is the current transportation volume of the i-th sub-sector, is the transportation business volume completed by the j-th surveyed enterprise in the i-th sub-sector; The methods for splitting the added value of passenger and freight transportation by various transportation modes are as follows: ; among them, is the split value of the added value of the i-th sub-sector, is another sub-sector in the transportation mode where the i-th sub-sector is located. For example, if i is road passenger transportation, then j represents road freight transportation, represents the actual added value of the transportation modes corresponding to sub-sectors i and j. Obviously, is equal to and The weight distribution of the added value of passenger and freight transportation by various transportation modes is as follows: , where i = 1 to 8 respectively represent railway passenger transportation, road passenger transportation, waterway passenger transportation, civil aviation passenger transportation, railway freight transportation, road freight transportation, waterway freight transportation, and civil aviation freight transportation.

[0030] The specific content of step 3.2 is as follows: The method for measuring the weight of the output scale of passenger and freight transportation services by various transportation modes is as follows: ; among them, is the transportation income of the i-th sub-sector.

[0031] The specific content of step 3.3 is as follows: The method for measuring the weight of social value recognition by various transportation modes is as follows: ; among them, is the weight of social value recognition of each sub-sector, is the first-layer survey weight corresponding to the passenger and freight attributes of this sub-sector, is the second-layer survey weight of this sub-sector in the corresponding passenger and freight fields and corresponding transportation modes.

[0032] The specific content of step 3.4 is as follows: The weights of each sub-sector are weighted among three dimensions through the geometric mean method. The specific calculation method is as follows: ; where i takes values from 1 to 8, representing the passenger and freight transportation modes of railways, roads, waterways, and civil aviation respectively.

[0033] The specific content of step (4) is as follows: According to the basic expression form of the Fisher ideal index, the expression form of the index's intertemporal composite index is finally determined as follows: ; among them, represents the comprehensive weight of the i-th transportation mode in the j-th time stage. The weights in different time periods are mainly obtained through regular weight measurement by setting specific years or cycles; represents in the time stage The relative score value of the transportation volume completed by the \(i\)-th transportation mode in the \(t\)-th period.

[0034] The specific content of the said step (5) is as follows: Step 5-1, establish an identification model: Taking the two-state stage division of high growth rate and low growth rate as an example, in different stages, the year-on-year growth rate sequence of the synthesized comprehensive index its own volatility and its trend correlation with the growth rate of related indicators will show different characteristics. Establish a Markov regime-switching model: ; ; ; Among them, is the year-on-year growth rate of the comprehensive index in the \(t\)-th period, is the correlation coefficient between the growth rate of the comprehensive index and the growth rate of related indicators in different state stages, is the growth rate sequence of related indicators, the historical growth rate sequence of the comprehensive index or the growth rate sequence of other external indicators, is the residual, reflecting the degree of random volatility of the comprehensive index, and allowing differences in amplitude in different state stages; Step 4-2, calculate the probabilities of different state stages in each period. Iteratively estimate the probabilities in each period through Hamilton filtering. The core idea of Hamilton filtering is to continuously approximate and calculate the filtering probabilities corresponding to various states based on the newly added information observed in each period. The calculation process is as follows: First, calculate the initial probability expectations of the index growth rate in the initial period (\(t = 1\)) in each state stage: ; ; Among them, represents the state transition probabilities corresponding to the Markov chain process, and the matrix is all the information that can be obtained in the period \(t - 1\) and is helpful for stage judgment; Secondly, through iterative calculation period by period, for the \(t\)-th period, based on the known information in the \(t - 1\)-th period, estimate the preliminary probability expectations of the states in each cycle: ; Furthermore, update the state probability expectations of the current period with the newly added information in the \(t\)-th period, that is, calculate the proportion of the likelihood values in each state stage weighted by the initial probability expectations as the true probabilities corresponding to the assumptions in each stage of the current period: ; Among them, the function \(f\) is the probability distribution function corresponding to the growth rate value of the comprehensive index in the \(t\)-th period; Finally, iterate t according to the above steps until the cycle state of each period is obtained in the current period T. The corresponding filtering probability expression; Step 4-3: Calculate the parameters. By performing maximum likelihood estimation on the weighted models of each period, the parameter estimation results of the entire parameter set are obtained synchronously, and the state stages of each period The corresponding filtering probabilities are obtained synchronously: ; where is the likelihood function for the parameter set when the growth rate of the composite index is in state stage j; in addition, the expected duration of each state period is calculated through the obtained state transition probability .

[0035] As described above, only the preferred specific embodiments of the present invention are provided, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent substitutions or changes should be covered within the protection scope of the present invention.

Claims

1. A multi-dimensional comprehensive transportation index synthesis and stage division method, characterized in that The following steps are involved: (1) Seasonal adjustment: seasonal adjustment of passenger and freight transport indicators by railway, highway, waterway, and civil aviation; (2) Dimensionlessness: Standardize each indicator to eliminate dimensional differences; (3) Weight calculation: The comprehensive weights of different periods are calculated from three dimensions: contribution to economic growth, market service output, and social value recognition; (4) Dynamic synthesis of indicators: Dynamic synthesis of indicators across time periods is performed using the Fisher ideal index method; (5) Stage division: The Markov mechanism conversion method is introduced to identify the probability distribution of the stage state of the index in each period based on the historical change law of the index itself, thereby realizing stage division; The step (2) is specifically as follows: select the indexation method, that is, divide the current value by the base period value, multiply the resulting ratio by 100, and the value range is [0, +∞]. The specific formula is: ; among them, y i is the relative score value of the current transportation volume completion scale of the i-th sub-sector; The step (3) is specifically as follows: 3.1 Calculation of economic growth contribution weights: Determine the weights of passenger and freight transport services of each mode of transport through input-output tables and corporate financial data; 3.2 Calculation of service output scale weights, and determination of the market recognition value of passenger and freight transport services of each mode of transport through market freight rates and transport volumes; 3.3 Calculation of social value recognition weights, evaluating the non-economic value of each mode of transport through questionnaire surveys; 3.4 Synthetic comprehensive weight.

2. The multi-dimensional comprehensive transportation index synthesis and stage division method according to claim 1, wherein The step 3.1 is specifically as follows: The net profit of main business of enterprises related to the industry value-added accounting is introduced as a reference indicator for splitting. The statistical survey data related to the operating data of representative enterprises in the industry, combined with the business volume data of enterprises, can realize the net profit generated by unit transportation volume. Based on this, the net profit of passenger and freight production by industry is calculated. The result can be used as the approximate ratio of passenger and freight value-added splitting; the specific calculation formula for a single mode of transportation is as follows: ; among which, is the estimated net profit of the i-th sub-industry, is the net profit of the main business of the j-th surveyed enterprise in the i-th sub-industry, is the current transportation volume of the i-th sub-industry, is the transportation business volume completed by the j-th surveyed enterprise in the i-th sub-industry; The value-added breakdown of passenger and freight transport by mode of transport is as follows: ; among them, is the split value of the added value of the i-th sub-sector, is another sub-sector in the transportation mode where the i-th sub-sector is located. For example, if i is road passenger transport, then j represents road freight transport, represents the actual added value of the transportation modes corresponding to sub-sectors i and j. Obviously is equal to and the sum of; the weight distribution of the added value of passenger and freight transport for each transportation mode is as follows: , where i = 1 to 8 respectively represent railway passenger transport, road passenger transport, waterway passenger transport, civil aviation passenger transport, railway freight transport, road freight transport, waterway freight transport, and civil aviation freight transport.

3. The multi-dimensional comprehensive transportation index synthesis and stage division method according to claim 1, characterized in that The specific step 3.2 is as follows: The method for calculating the weight of passenger and freight service output scale of each mode of transportation is as follows: ; wherein, is the transportation revenue of the i-th sub-sector.

4. The multi-dimensional comprehensive transportation index synthesis and stage division method according to claim 1, characterized in that The specific content of step 3.3 is as follows: The measurement method for the social value recognition weights of each transportation mode is as follows: ; among them, is the social value recognition weight of each sub-industry, is the first-layer survey weight of the corresponding passenger and freight attributes of this sub-industry, is the second-layer survey weight of this sub-industry in the corresponding passenger and freight fields and for the corresponding transportation mode.

5. The multi-dimensional comprehensive transportation index synthesis and stage division method according to claim 1, characterized in that The specific step 3.4 is: weighting the weights of each sub-industry among the three dimensions by the geometric mean method. The specific calculation method is as follows: ; where i ranges from 1 to 8, respectively representing the passenger and freight transportation modes of railways, highways, waterways, and civil aviation.

6. The multi-dimensional comprehensive transportation index synthesis and stage division method according to claim 1, characterized in that The step (4) is specifically as follows: According to the basic expression of Fisher's ideal index, the expression of the index's cross-period composite index is finally determined as follows: ; among them, represents the comprehensive weight of the i-th transportation mode in the j-th time stage. The weights in different time periods are mainly obtained by setting specific years or cycles to carry out regular weight measurement; represents in the time stage the relative score value of the transportation volume completed by the i-th transportation mode in the t-th period.

7. The multi-dimensional comprehensive transportation index synthesis and stage division method according to claim 1, characterized in that The step (5) is specifically as follows: Step 5-1, establish an identification model: Taking the two-state stage division of high growth rate and low growth rate as an example, in different stages, the year-on-year growth rate sequence of the synthesized comprehensive index its own volatility and its trend correlation with the growth rate of related indicators will show different characteristics, and establish a Markov regime-switching model: ； ； ； Among them, is the year-on-year growth rate of the composite index in the t-th period, is the correlation coefficient between the growth rate of the composite index and the growth rate of the associated indicator in different state stages, is the historical growth rate sequence of the associated indicator, the composite index, or the growth rate sequence of other external indicators, is the residual, reflecting the severity of the random fluctuation of the composite index, and allowing differences in amplitude in different state stages; Step 4-2, calculate the probabilities of different states in each period, iteratively estimate the probabilities of each period through Hamilton filtering, and continuously approximate the filtering probabilities corresponding to various states based on the actually observed information. The calculation process is as follows: First, calculate the initial probability expectation of the exponential growth rate in each state stage at the initial stage (t=1): ； ； Among them, represents the state transition probabilities corresponding to the Markov chain process, and the matrix is all the information that can be obtained at time t - 1 and is helpful for carrying out stage judgment; Secondly, through iterative calculations for each period, for the t-th period, based on the known information of the (t - 1)-th period, the preliminary probability expectations of the states of each cycle are estimated: ; Furthermore, the newly added information in the tth period is used to update the expected state probability of the current period, that is, by calculating the proportion of the likelihood values ​​of each state stage in the current period weighted by the initial probability expectation, as the corresponding true probability under the assumption of being in each stage in the current period: ； Among them, function f is the probability distribution function of the growth rate value of the comprehensive index corresponding to the tth period; Finally, iterate t according to the above steps until the cycle states of each period are obtained in the current period, i.e., the T-th period. The corresponding filtering probability expression; Step 4-3: Calculate the parameters. By performing maximum likelihood estimation on each period's weighted model, obtain the parameter estimation results of the entire parameter set synchronously, and simultaneously determine the state stage of each period. The corresponding filtering probabilities are obtained synchronously: ； Among them, is the likelihood function for the parameter set when the growth rate of the composite index is in state stage j; in addition, the expected duration of each state period continuation is calculated through the obtained state transition probability .

Images