Carbon footprint recognition method based on improved fuzzy C-means clustering algorithm

Abstract

The application provides a carbon code recognition method based on an improved fuzzy C-means clustering algorithm, and comprises the following steps: obtaining parameter data of a green development state, wherein the parameters include sample data sets of enterprise carbon emission intensity, enterprise vehicle scale, enterprise owned vehicle emission standard, and enterprise owned vehicle average age; using weighted Euclidean distance to measure the similarity of each parameter to obtain a similarity measurement function; constructing a gradient descent method minimization evaluation function based on the similarity measurement function, and using the function to solve the weight of each parameter; introducing the weight of the parameter into the fuzzy C-means clustering algorithm, and taking the final fuzzy C-means clustering algorithm based on the weight of the parameter as a model for recognizing the green development state of a freight transport enterprise; and using the fuzzy C-means clustering algorithm model based on the parameter weight to recognize the green development state of the freight transport enterprise. The application is helpful for promoting the green transformation and high-quality development of freight transport enterprises.

Description

Technical Field

[0001] This invention belongs to the field of enterprise green development management technology. This invention relates to a carbon efficiency code recognition method based on an improved fuzzy C-means clustering algorithm. Background Technology

[0002] As one of the three major energy consumers and carbon emission sources, the transportation industry has always been a key area for addressing climate change. Meanwhile, as a leading and fundamental sector of national economic development, the transportation industry will maintain a certain growth rate and scale for some time to come, especially with freight demand continuing to grow steadily. While maintaining economic growth, the green development of freight companies has become a new research direction. Currently, most research on the green development of freight companies is qualitative, only outlining the impact of influencing factors on their green development, lacking quantitative research. That is, there are currently no quantitative methods to identify the key influencing factors for the green development of freight companies, thus failing to effectively guide their green development. Summary of the Invention

[0003] This invention provides a carbon efficiency code recognition method based on an improved fuzzy C-means clustering algorithm.

[0004] This invention provides a carbon efficiency code recognition method based on an improved fuzzy C-means clustering algorithm, comprising:

[0005] S1. Obtain parameter data for freight companies, including: sample datasets of company carbon emission intensity, company vehicle scale, company vehicle emission standards, and average vehicle age; where company carbon emission intensity is the carbon emission per unit freight turnover of the freight company, in kilograms * 100 ton-kilometers; company vehicle scale is the number of trucks owned by the freight company, in vehicles; company vehicle emission standards are the number of trucks owned by the freight company that meet the National V emission standard and zero emission standard, in vehicles; and average vehicle age is the age of the trucks owned by the freight company, in years.

[0006] S2. Considering that the parameters of corporate carbon emission intensity, corporate vehicle scale, corporate vehicle emission standards, and corporate vehicle average age have different effects on the green development status of freight companies, the similarity of each parameter is measured using weighted Euclidean distance to obtain a similarity measurement function. Then, a gradient descent method is used to minimize the evaluation function based on the similarity measurement function, and the weights of each parameter are solved using the gradient descent method to minimize the evaluation function.

[0007] S3. The weights of each parameter are introduced into the fuzzy C-means clustering algorithm to obtain the final fuzzy C-means clustering algorithm model based on the parameter weights. The final fuzzy C-means clustering algorithm model based on the parameter weights is used as a model to identify the green development status of freight companies, which is called the "carbon efficiency code" identification method.

[0008] S4. Input data including the carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age of real freight companies into the fuzzy C-means clustering algorithm model based on parameter weights to output cluster center calculation results; based on the cluster center calculation results, identify the green development status of freight companies.

[0009] Furthermore, in step S1 of the carbon efficiency code recognition method based on the improved fuzzy C-means clustering algorithm, the enterprise's carbon emission intensity is calculated, including the following steps:

[0010] S1.1 Obtain the total mileage of freight company vehicles, energy consumption per 100 kilometers, energy conversion coefficient to standard coal equivalent, carbon emission coefficient of energy, and freight turnover.

[0011] S1.2 Construct the formula for calculating corporate carbon emission intensity:

[0012]

[0013] Among them, E i Indicates the carbon emission intensity of freight companies; S i Indicates the total mileage traveled by the freight company's vehicles; α j Indicates the vehicle's energy consumption per 100 kilometers; β j Indicates the energy conversion factor to standard coal equivalent; γ j Indicates the energy carbon emission factor; Q i This indicates the freight turnover of freight companies.

[0014] Furthermore, in step S2 of the carbon efficiency code recognition method based on the improved fuzzy C-means clustering algorithm, considering the different effects of parameters such as enterprise carbon emission intensity, enterprise vehicle scale, enterprise vehicle emission standards, and average vehicle age on the green development status of freight enterprises, the similarity of each parameter is measured using weighted Euclidean distance to obtain a similarity measurement function, including the following steps:

[0015] S2.1 Construct a similarity measurement function that considers parameter weights:

[0016]

[0017] Where ω=(ω1,ω2,…,ω m This indicates parameters including a company's carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age. This represents a similarity metric function that takes into account parameter weights, and its value is uniformly distributed in the interval [0,1] by adjusting β; β is a constant in the interval [0,1]. This represents the Euclidean distance based on the attribute weights.

[0018] Furthermore, in step S2.1 of the carbon code recognition method based on the improved fuzzy C-means clustering algorithm, β and The solution formula is:

[0019]

[0020] Where n represents the number of parameter data sets, d pq This represents the ordinary Euclidean distance between the p-th cluster center and the q-th data point.

[0021]

[0022] Where, ω k It describes the importance of the k parameter in clustering, x pk x qk This represents the coordinates of the cluster center and the coordinates of the data points for the k-parameters.

[0023] Furthermore, in the carbon code recognition method based on the improved fuzzy C-means clustering algorithm described above, the ordinary Euclidean distance d pq The formula is:

[0024]

[0025] Where, x pk x qk This represents the coordinates of the cluster center and the coordinates of the data points for the k-parameters.

[0026] Furthermore, in step S2 of the carbon code recognition method based on the improved fuzzy C-means clustering algorithm, a gradient descent minimization evaluation function is constructed based on the similarity metric function, as shown in the following formula:

[0027]

[0028] Where E(ω) represents the minimization evaluation function; This indicates a weighted similarity metric function; ρ pq This indicates that the weighted similarity measurement function is not considered.

[0029] Furthermore, in the carbon code recognition method based on the improved fuzzy C-means clustering algorithm described above, the weighted similarity metric function ρ is not considered. pq The formula is as follows:

[0030]

[0031] Furthermore, in the carbon code recognition method based on the improved fuzzy C-means clustering algorithm, the E(ω) minimization evaluation function is solved using a gradient descent algorithm, with the specific steps as follows:

[0032] (1) Based on the aforementioned β and Solve the formula to find the value of β;

[0033] (2) Initialize the weights ω = (ω1, ω2, ..., ω) of each parameter within [0, 1]. m );

[0034] (3) Calculate the weight correction Δω j :

[0035]

[0036]

[0037]

[0038]

[0039] (4) Where η is the learning rate;

[0040]

[0041] (5) When ω j +Δω j When the value is greater than 0, update the weight ω. j =ω j +Δω j ;

[0042] (6) Repeat steps (1) to (5) until convergence.

[0043] Furthermore, in step S3 of the carbon code recognition method based on the improved fuzzy C-means clustering algorithm, the objective function, membership matrix, and cluster center calculation formula of the final fuzzy C-means clustering algorithm model based on parameter weights are as follows:

[0044] Objective function:

[0045] Where U is the membership degree of each data point to its corresponding cluster center; v c μ is the c-th fuzzy cluster center; ij ∈[0,1] represents the membership degree of the j-th data point to the i-th cluster center; Let be the Euclidean distance between the i-th cluster center and the j-th data point based on parameter weights; m∈[1,∞] is a weighted exponent;

[0046] Membership matrix:

[0047] Cluster center:

[0048] Where b is the number of algorithm iterations.

[0049] Furthermore, in step S4 of the carbon efficiency code recognition method based on the improved fuzzy C-means clustering algorithm, data including the carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age of real freight companies are input into the parameter-weighted fuzzy C-means clustering algorithm model to output the cluster center calculation results, including the following step:

[0050] (1) Set the initial cluster category c;

[0051] (2) Set the fuzzy weighted index m;

[0052] (3) Initialize the membership matrix μ ij ;

[0053] (4) Update cluster center v i ;

[0054] (5) Calculate and update the membership matrix μ ij ;

[0055] (6) Calculate the objective function of the final fuzzy C-means clustering algorithm model based on the weight of the parameters. If the change is less than the set threshold, the calculation stops. If the convergence condition is not met, return to step (4) for iterative calculation.

[0056] Furthermore, in step S4 of the carbon efficiency code recognition method based on the improved fuzzy C-means clustering algorithm, the green development status of freight enterprises is identified based on the cluster center calculation results, including:

[0057] (1) Based on experience, the optimal value of c is determined to be 3, which represents the green state, blue state and red state, respectively, representing the good, average and poor green development status of freight companies;

[0058] (2) The four parameters of each enterprise—carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age—are combined into a four-dimensional vector. This vector is used as the input sample data, and the cluster centers are calculated using the method described in step one, as follows:

[0059]

[0060] In this matrix, rows 1 to 3 represent the cluster centers corresponding to the green, blue, and red states, respectively; the four values ​​in each row represent the enterprise's carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age.

[0061] With the establishment of more and more freight data platforms, the integration of relevant data of freight enterprises has become possible. This invention aims to extract key parameters based on a large amount of relevant data of freight enterprises, and through the improvement of the fuzzy C-means clustering algorithm, realize a quantitative algorithm for identifying the green development status of freight enterprises. It can be used to find out the factors affecting the green development of freight enterprises and provide support for the green and high-quality development of freight enterprises. Attached Figure Description

[0062] Figure 1 This is a flowchart of a carbon code recognition method based on an improved fuzzy C-means clustering algorithm, as described in this invention.

[0063] Figure 2 This is a diagram of the parameter data integration module for freight companies.

[0064] Figure 3 This is a flowchart of the calculation based on the improved fuzzy C-means clustering algorithm.

[0065] Figure 4 This is a flowchart illustrating the calculation and application process for freight companies. Detailed Implementation

[0066] The technical solutions of the embodiments of the present invention will be explained and described below with reference to the accompanying drawings. However, the following embodiments are only preferred embodiments of the present invention and not all of them. Other embodiments obtained by those skilled in the art based on the embodiments in the implementation methods without creative effort are all within the protection scope of the present invention.

[0067] Before introducing the solution of this embodiment, the application background of this embodiment will be introduced.

[0068] As the background technology points out, the transportation industry is one of the three major energy consumers and carbon emission sources, and freight is a key carbon emission source in the transportation industry. With economic development, the demand for freight transportation will continue to increase, which will inevitably lead to a continued increase in pollution emissions such as carbon emissions. It is urgent to use information and quantitative methods to identify the green development status of freight companies, so as to provide a basis for guiding the green development of freight companies.

[0069] The essence of green development for freight companies is to achieve both low energy consumption and low emissions while promoting sustainable economic development. Therefore, factors related to green development in freight companies include not only carbon emissions but also the size of the company's fleet, vehicle emission standards, and average vehicle age. Thus, achieving green development for freight companies requires establishing a carbon efficiency code identification method. Due to the diversity and difficulty in collecting the required data, there is currently no quantitative method for identifying the green development status of freight companies. Therefore, it is necessary to integrate data from multiple platforms and construct a model using fuzzy algorithms to identify the green development status of freight companies.

[0070] In one embodiment, this application provides a carbon code recognition method based on an improved fuzzy C-means clustering algorithm. Please refer to the appendix. Figure 1 This includes the following steps:

[0071] Step 1: Obtain parameter data for freight companies. These parameters include: sample datasets of company carbon emission intensity, company vehicle size, company vehicle emission standards, and average vehicle age. Specifically, company carbon emission intensity refers to the carbon emissions per unit of freight turnover, expressed in kilograms per 100 ton-kilometers; company vehicle size refers to the number of trucks owned by the freight company, expressed in vehicles; company vehicle emission standards refer to the number of trucks meeting the National V emission standard and zero-emission standards, expressed in vehicles; and average vehicle age refers to the age of the trucks owned by the freight company, expressed in years.

[0072] Step 2: Considering that parameters such as corporate carbon emission intensity, corporate vehicle scale, corporate vehicle emission standards, and average vehicle age have different effects on the green development status of freight companies, the similarity of each parameter is measured using weighted Euclidean distance to obtain a similarity measurement function. Then, a gradient descent method is used to minimize the evaluation function based on the similarity measurement function, and the weights of each parameter are solved using the gradient descent method to minimize the evaluation function.

[0073] Step 3: Introduce the weights of each parameter into the fuzzy C-means clustering algorithm to obtain the final fuzzy C-means clustering algorithm model based on the parameter weights; use the final fuzzy C-means clustering algorithm model based on the parameter weights as a model to identify the green development status of freight companies, which is called the "carbon efficiency code" identification method.

[0074] Step 4: Input data including the actual carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age of freight companies into the parameter-based weighted fuzzy C-means clustering algorithm model to output cluster center calculation results; based on the cluster center calculation results, identify the green development status of freight companies.

[0075] For the method of obtaining freight company parameter data in step one, please refer to the appendix. Figure 2 Parameters such as vehicle size, vehicle emission standards, and average vehicle age are all sourced from the transportation administration database system. Carbon emission intensity is calculated using vehicle energy consumption data, vehicle GPS data, and freight turnover data. These three data sources are respectively derived from vehicle tracking surveys, vehicle network control systems, and enterprise data tables. All parameter data are ultimately integrated into the Zhejiang Provincial Transportation Carbon Peak Platform.

[0076] In step one, the formula for calculating the enterprise's carbon emission intensity is as follows:

[0077]

[0078] Among them, E i Indicates the carbon emission intensity of freight companies; S i Indicates the total mileage traveled by the freight company's vehicles; α j Indicates the vehicle's energy consumption per 100 kilometers; β j Indicates the energy conversion factor to standard coal equivalent; γ j Indicates the energy carbon emission factor; Q i This indicates the freight turnover of freight companies.

[0079] S i Total mileage of freight company vehicles, Q i Freight companies obtain real data on freight turnover through the vehicle network control system and the company's unified reporting system. Tables 1, 2, and 3 are parameter tables for vehicle energy consumption per 100 kilometers, energy conversion to standard coal equivalent, and energy carbon emission coefficient, respectively. Table 1, vehicle energy consumption per 100 kilometers, is based on theoretical values ​​according to national standards and was obtained through a special survey conducted by the Zhejiang Provincial Department of Transportation. Tables 2 and 3, energy conversion to standard coal equivalent, and energy carbon emission coefficient parameter tables, are obtained with reference to the "Zhejiang Province Greenhouse Gas Inventory Compilation Guidelines (2020 Revised Edition)".

[0080] Table 1 Vehicle Energy Consumption Parameters per 100 km

[0081]

[0082]

[0083] Table 2 Energy Conversion Coal Standard Coupon Parameter Table

[0084]

[0085] Table 3 Energy Carbon Emission Coefficient Parameters

[0086]

[0087] In step two, considering the varying impacts of different parameters on the green development status of freight companies, it is necessary to calculate the weights of each parameter. The specific steps are as follows:

[0088] S2.1 The four parameters of a freight company—carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age—are combined into a four-dimensional vector. The ordinary Euclidean distance between these vectors is calculated using the following formula:

[0089]

[0090] Where, x pk x qk This represents the coordinates of the cluster center and the coordinates of the data points for the k-parameters.

[0091] S2.2 The value of β is obtained according to the following formula, and its value is calculated to be 0.41.

[0092]

[0093] Where n represents the number of parameter data sets, d pq This represents the ordinary Euclidean distance between the p-th cluster center and the q-th data point.

[0094] S2.3 The four parameters of a freight company—carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age—are combined into a four-dimensional vector. The weighted Euclidean distance between these vectors is calculated using the following formula:

[0095]

[0096] Where, ω k It describes the importance of the k parameter in clustering, x pk x qk This represents the coordinates of the cluster center and the coordinates of the data points for the k-parameters.

[0097] Based on the above results, construct a similarity measurement function that considers parameter weights, and calculate the similarity of four parameters: carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age of enterprises.

[0098]

[0099] Where ω=(ω1,ω2,…,ω m This indicates parameters such as a company's carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age. This represents a similarity metric function that takes into account parameter weights, and its value is uniformly distributed in the interval [0,1] by adjusting β; β is a constant in the interval [0,1]. This represents the Euclidean distance based on the attribute weights.

[0100] Based on this, S2.5 uses the similarity measurement formula to construct the evaluation function formula as follows:

[0101]

[0102] Where E(ω) represents the minimization evaluation function; This indicates a weighted similarity metric function; ρ pq This indicates that the weighted similarity measurement function is not considered.

[0103] The following steps are used to calculate the weight values ​​of four parameters for a company: carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age.

[0104] (1) Solve for the value of β;

[0105] (2) Initialize the weights of the four parameters—carbon emission intensity, vehicle size, vehicle emission standard, and average vehicle age—within [0, 1], assigning them the same weight ω = (0.25, 0.25, 0.25, 0.25);

[0106] (3) Calculate the weight correction Δω j :

[0107]

[0108]

[0109]

[0110]

[0111] (4) Where η is the learning rate;

[0112]

[0113] (5) When ω j +Δω j When the value is greater than 0, update the weight ω. j =ω j +Δω j ;

[0114] (6) Repeat steps (1) to (5) until convergence.

[0115] The iteration terminates when E(ω) converges, at which point the value of E(ω) is 0.526, and the weights of the four parameters, carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age, are obtained as ω = (0.48, 0.15, 0.22, 0.13).

[0116] Furthermore, the fuzzy C-means algorithm is a partition-based clustering algorithm. Its core idea is to maximize the similarity between objects grouped into the same cluster and minimize the similarity between different clusters. It is often used for traffic operation status classification. Utilizing this algorithm, and considering the differences in the influence of various parameters, the weights of these parameters are introduced into the fuzzy C-means clustering algorithm to identify the green development status of freight companies. Its principle and calculation process can be found in [reference needed]. Figure 3 As shown.

[0117] In step three, the weights of the parameters are introduced into the fuzzy C-means clustering algorithm. The specific steps are as follows:

[0118] S3.1 Construct the objective function:

[0119] Where U is the membership degree of each data point to its corresponding cluster center; v c μ is the c-th fuzzy cluster center; ij ∈[0,1] represents the membership degree of the j-th data point to the i-th cluster center; is the Euclidean distance between the i-th cluster center and the j-th data point based on parameter weights; m∈[1,∞] is a weighted exponent.

[0120] S3.2 Constructing the membership matrix:

[0121] S3.3 Constructing Cluster Centers:

[0122] Where b is the number of algorithm iterations.

[0123] In step four, data on the actual carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age of freight companies are input into a fuzzy C-means clustering algorithm model based on parameter weights. The output results identify the green development status of freight companies. The specific steps are as follows:

[0124] S4.1 Set the initial cluster category c. Based on relevant research experience, the optimal value range of c is [3,6]. In this paper, we set c=3, which represents the green state, blue state, and red state, respectively, representing the good, average, and poor green development status of freight companies.

[0125] S4.2 Set the fuzzy weighting index m. Based on relevant research experience, the optimal range of m is [1.5, 2.5]. Let m = 2.

[0126] S4.3 Initialize the membership matrix μ ij ;

[0127] S4.4 Update Cluster Center v i ;

[0128] S4.5 Calculate and update the membership matrix μ ij ;

[0129] S4.6 Calculate the objective function of the final fuzzy C-means clustering algorithm model based on parameter weights. If the change is less than the set threshold (usually the iteration stop threshold ε = 0.001), the calculation stops. If the convergence condition is not met, return to step (4) for iterative calculation.

[0130] A clustering analysis of over 200,000 historical freight company data points can be performed using an improved fuzzy C-means clustering algorithm, outputting the following cluster center results:

[0131]

[0132] In this matrix, rows 1 to 3 represent the cluster centers corresponding to the green, blue, and red states, respectively; the four values ​​in each row represent the enterprise's carbon emission intensity, vehicle scale, vehicle emission standards, and average vehicle age.

[0133] Currently, based on an improved fuzzy C-means clustering algorithm, the green development status of over 600 freight companies has been identified through the Zhejiang Provincial Transportation Carbon Peaking Platform. Those identified as having a green development status are receiving policy support and low-interest green loans from relevant industry authorities and financial institutions. The specific process is as follows... Figure 4 As shown.

Claims

1. A carbon code recognition method based on an improved fuzzy C-means clustering algorithm, characterized in that, include: S1. Obtain parameter data for freight companies, including: company carbon emission intensity, company vehicle size, company vehicle emission standards, and a sample dataset of the average age of the company's vehicles; wherein, company carbon emission intensity is the carbon emission per unit of freight turnover, expressed in kilograms. 100 ton-kilometers; Enterprise vehicle scale, i.e., the number of trucks owned by the freight company, in units of vehicles; Enterprise vehicle emission standards, i.e., the number of trucks owned by the freight company that meet the National V emission standard and the zero emission standard, in units of vehicles; Average vehicle age of the enterprise vehicles, i.e., the age of the trucks owned by the freight company, in units of years. S2. Considering that the parameters of corporate carbon emission intensity, corporate vehicle scale, corporate vehicle emission standards, and corporate vehicle average age have different effects on the green development status of freight companies, the similarity of each parameter is measured using weighted Euclidean distance to obtain a similarity measurement function. Then, a gradient descent method is used to minimize the evaluation function based on the similarity measurement function, and the weights of each parameter are solved using the gradient descent method to minimize the evaluation function. S3. The weights of each parameter are introduced into the fuzzy C-means clustering algorithm to obtain the final fuzzy C-means clustering algorithm model based on the parameter weights. The final fuzzy C-means clustering algorithm model based on the parameter weights is used as a model to identify the green development status of freight companies, which is called the "carbon efficiency code" identification method. S4. Input data including the carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age of real freight companies into the fuzzy C-means clustering algorithm model based on parameter weights to output cluster center calculation results; based on the cluster center calculation results, identify the green development status of freight companies.

2. The carbon code recognition method based on the improved fuzzy C-means clustering algorithm according to claim 1, characterized in that, In step S1, the enterprise's carbon emission intensity is calculated, including the following steps: S1.1 Obtain the total mileage of freight company vehicles, energy consumption per 100 kilometers, energy conversion coefficient to standard coal equivalent, carbon emission coefficient of energy, and freight turnover. S1.2 Construct the formula for calculating corporate carbon emission intensity: ； in, Indicates the carbon emission intensity of freight companies; This indicates the total mileage traveled by the freight company's vehicles; Indicates the vehicle's energy consumption per 100 kilometers; Indicates the energy conversion factor to standard coal equivalent; Indicates the carbon emission coefficient of energy; This indicates the freight turnover of freight companies.

3. The carbon code recognition method based on the improved fuzzy C-means clustering algorithm according to claim 1, characterized in that, In step S2, considering the different effects of parameters such as corporate carbon emission intensity, corporate vehicle scale, corporate vehicle emission standards, and average vehicle age on the green development status of freight companies, a weighted Euclidean distance is used to measure the similarity of each parameter, resulting in a similarity measurement function, including the following steps: S2.1 Construct a similarity measurement function that considers weights: ， in, This includes parameters such as a company's carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age. This represents a similarity metric function that takes into account parameter weights, and its value is adjusted... Evenly distributed within [0,1]; It is a constant in the interval [0,1]. This represents the Euclidean distance based on the attribute weights.

4. The carbon code recognition method based on the improved fuzzy C-means clustering algorithm according to claim 3, characterized in that, In step S2.1, and The solution formula is: ， in, Indicates the number of parameter data sets. Indicates the first The cluster centers and the first Ordinary Euclidean distance of 1 data point; ， in, It describes the importance of the k parameter in clustering. , This represents the coordinates of the cluster center and the coordinates of the data points for the k-parameters.

5. The carbon code recognition method based on the improved fuzzy C-means clustering algorithm according to claim 4, characterized in that, Ordinary Euclidean distance The formula is: ； in, , This represents the coordinates of the cluster center and the coordinates of the data points for the k-parameters.

6. The carbon code recognition method based on the improved fuzzy C-means clustering algorithm according to claim 4 or 5, characterized in that, In step S2, a gradient descent minimization evaluation function is constructed based on the similarity metric function, as shown in the following formula: ， in, This represents the function that minimizes the evaluation criteria. This indicates that a weighted similarity measurement function is considered. This indicates that the weighted similarity measurement function is not considered.

7. The carbon code recognition method based on the improved fuzzy C-means clustering algorithm according to claim 6, characterized in that, Without considering weighted similarity measurement function The formula is as follows: 。 8. The carbon code recognition method based on the improved fuzzy C-means clustering algorithm according to claim 6, characterized in that, The evaluation function is minimized using the gradient descent method, and the specific steps are as follows: (1) According to the above and Solve the formula, solve The value; (2) Initialize the weights of each parameter in [0, 1]. ; (3) Calculate the weight adjustment amount : ； ； ； ； (4) Among them, It is the learning rate; ； (5) When Update weights at that time. ; (6) Repeat steps (1) to (5) until convergence.

9. The carbon code recognition method based on the improved fuzzy C-means clustering algorithm according to claim 1, characterized in that, In step S3, the objective function, membership matrix, and cluster center calculation formula of the final fuzzy C-means clustering algorithm model based on parameter weights are as follows: Objective function: , in, It is the membership degree of each data point to its corresponding cluster center; It is the c-th fuzzy cluster center; ∈[0,1] represents the membership degree of the j-th data point to the i-th cluster center; Let be the Euclidean distance between the i-th cluster center and the j-th data point based on parameter weights; m∈[1,∞] is a weighted exponent; Membership matrix: , Cluster center: , Where b is the number of algorithm iterations.

10. The carbon code recognition method based on the improved fuzzy C-means clustering algorithm according to claim 9, characterized in that, In step S4, data including the carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age of actual freight companies are input into the parameter-based weighted fuzzy C-means clustering algorithm model to output the cluster center calculation results, including step one: (1) Set the initial cluster category c; (2) Set the fuzzy weighting index m; (3) Initialize the membership matrix ; (4) Update cluster centers ; (5) Calculate and update the membership matrix ; (6) Calculate the objective function of the final fuzzy C-means clustering algorithm model based on the weight of the parameters. If the change is less than the set threshold, the calculation stops. If the convergence condition is not met, return to step (4) for iterative calculation.

11. The carbon code recognition method based on the improved fuzzy C-means clustering algorithm according to claim 10, characterized in that, In step S4, based on the cluster center calculation results, the green development status of freight enterprises is identified, including: (1) Based on experience, the optimal value of c is determined to be 3, which represents the green state, blue state and red state, respectively, representing the good, average and poor green development status of freight companies; (2) The four parameters of each enterprise—carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age—are combined into a four-dimensional vector. This vector is used as the input sample data. Using the method described in step one, the cluster centers are calculated as follows: ； In this matrix, rows 1 to 3 represent the cluster centers corresponding to the green, blue, and red states, respectively; the four values ​​in each row represent the enterprise's carbon emission intensity, vehicle size, vehicle emission standards, and average vehicle age.

Images