Bus carbon efficiency key causal chain identification method, electronic equipment and medium

Abstract

The invention discloses a bus carbon efficiency key causal chain identification method, electronic equipment and a medium, and the method comprises the steps: constructing a line-level bus carbon efficiency variable set which comprises a control variable, a covariant and unit passenger mileage carbon emission; the control variable is a bus operation strategy; the covariable is a bus carbon efficiency influence factor; assuming that the line-level bus carbon efficiency variable set obeys a directed acyclic graph structure formed by linear equations, and solving a causal effect matrix; establishing a bus carbon efficiency causal directed graph by taking a line-level bus carbon efficiency variable as a node and a causal effect as an edge weight; and performing post-nonlinear causal relationship test and conditional independence causal test on the bus carbon efficiency causal digraph to obtain an optimized bus carbon efficiency causal digraph, thereby obtaining a bus carbon efficiency key causal link.

Description

Technical Field

[0001] The present invention belongs to the technical field of urban transportation carbon efficiency, and in particular relates to a method, electronic equipment, and medium for identifying a key causal chain of public transportation carbon efficiency. Background Art

[0002] With the accelerating pace of urbanization and the deepening promotion of green, low-carbon development concepts, urban public transportation systems are playing an increasingly important role in reducing carbon emissions and optimizing energy utilization in the transportation sector. Carbon efficiency has become a core indicator for measuring the green operation and sustainable development of public transportation systems. Carbon efficiency is influenced by multiple factors, including road infrastructure configuration, multimodal transportation transfers, public transportation operation strategies, population density, and land use structure. Complex multi-level linkages and causal couplings exist between these variables.

[0003] Existing causal analysis methods for carbon efficiency primarily rely on traditional statistical inference tools such as structural equation models and regression analysis. These methods often require subjective assumptions about the causal order or structure between variables, limiting their analysis to local causal relationships between pairs of variables. This makes it difficult to fully reveal the multivariate, multidimensional, global causal structure of urban public transportation systems. Furthermore, faced with complex scenarios such as the fusion of multi-source heterogeneous data, and the nonlinear and dynamic nature of variables in public transportation systems, traditional methods are significantly insufficient in terms of applicability and modeling accuracy, making it difficult to accurately identify the key causal links that influence carbon efficiency improvements.

[0004] Therefore, there is an urgent need to develop technical means that can systematically characterize the multi-level complex causal relationships between variables, dynamically track the evolution of causal chains, and adapt to the characteristics of multi-source heterogeneous data. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the present invention provides a method, electronic equipment and medium for identifying the key causal chain of public transportation carbon efficiency.

[0006] In a first aspect, an embodiment of the present invention provides a method for identifying a key causal chain of public transportation carbon efficiency, the method comprising the following steps:

[0007] Constructing a route-level bus carbon efficiency variable set, the route-level bus carbon efficiency variable set includes control variables, covariates, and carbon emissions per passenger mile; the control variables are bus operation strategies; and the covariates are factors affecting bus carbon efficiency;

[0008] Assuming that the set of bus line-level carbon efficiency variables obeys a directed acyclic graph structure consisting of a set of linear equations, the causal effect matrix is ​​solved. A directed causal graph of bus carbon efficiency is established with variables as nodes and causal effects as edge weights.

[0009] The post-nonlinear causal relationship test and / or conditional independence causal test are performed on the public transportation carbon efficiency causal directed graph to obtain the optimized public transportation carbon efficiency causal directed graph, thereby obtaining the key causal links of public transportation carbon efficiency;

[0010] Calculate the global causal chain evolution index, strong causal chain stability coefficient and local causal effect variation coefficient to reveal the spatiotemporal evolution law of key causal chains;

[0011] Time series variables were selected from the line-level bus carbon efficiency variable set, and a full-input model and a variable removal model were established using a time series neural network. The difference in prediction errors between the two models was used to quantify the time series causal effect.

[0012] In a second aspect, an embodiment of the present invention provides an electronic device, including:

[0013] at least one processor; and

[0014] a memory communicatively connected to the at least one processor; wherein,

[0015] The memory stores one or more computer programs executable by the at least one processor. The one or more computer programs are executed by the at least one processor to enable the at least one processor to perform the above-mentioned method for identifying key causal chains of public transportation carbon efficiency.

[0016] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-mentioned method for identifying the key causal chain of public transportation carbon efficiency.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] The present invention provides a method for identifying key causal chains of public transportation carbon efficiency, constructing a line-level public transportation carbon efficiency variable set, wherein the line-level public transportation carbon efficiency variable set includes control variables, covariates, and carbon emissions per unit passenger mile; the control variables are public transportation operation strategies; the covariates are factors affecting public transportation carbon efficiency; assuming that the line-level public transportation carbon efficiency variable set obeys a directed acyclic graph structure consisting of a system of linear equations, a causal effect matrix is ​​solved; a public transportation carbon efficiency causal directed graph is established with variables as nodes and causal effects as edge weights; a post-nonlinear causal relationship test and / or conditional independence causal test are performed on the public transportation carbon efficiency causal directed graph to obtain an optimized public transportation carbon efficiency causal directed graph, thereby accurately identifying key causal links of public transportation carbon efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 This is a flow chart of a method for identifying a key causal chain of public transportation carbon efficiency provided by an embodiment of the present invention;

[0021] Figure 2 This is a technical roadmap for the method for identifying key causal chains of public transportation carbon efficiency provided by an embodiment of the present invention;

[0022] Figure 3 This is the identification result of the key causal chain of public transportation carbon efficiency provided by the embodiment of the present invention;

[0023] Figure 4 is a schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0025] It should be noted that, unless there is any conflict, the features in the following embodiments and implementations may be combined with each other.

[0026] like Figure 1 and Figure 2 As shown, an embodiment of the present invention provides a method for identifying a key causal chain of public transportation carbon efficiency, the method comprising the following steps:

[0027] Step S1: construct a route-level bus carbon efficiency variable set, which includes control variables, covariates, and carbon emissions per passenger mile; the control variables are bus operation strategies; and the covariates are factors affecting bus carbon efficiency.

[0028] In this example, by integrating multi-source heterogeneous data (including but not limited to bus smart cards, mobile phone signaling, road network and POI data), and taking bus routes as the analysis unit, a line-level bus carbon efficiency variable set S = {F, T, BEP} is established.

[0029] Among them, bus CO emissions per passenger-kilometre (BEP) is the carbon efficiency indicator, and the control variable T includes bus operation strategies, but is not limited to the setting of bus lanes, bus signal priority strategy, bus route electrification priority strategy, transfer hub and seamless connection facility construction, and bus fare discounts. The covariate set F={f1,f2,…,f NF} are factors affecting public transportation carbon efficiency, including but not limited to station density, population density, POI density (work, residence), operating speed, passenger load factor, road type ratio (expressway, trunk road, primary road, tertiary road), fleet size, vehicle type, departure interval, bus-subway connection index, turnover volume, etc.

[0030] Step S2: Assuming that the line-level bus carbon efficiency variable set obeys a directed acyclic graph structure composed of a set of linear equations, solve the causal effect matrix; establish a bus carbon efficiency causal directed graph with variables as nodes and causal effects as edge weights.

[0031] Specifically, step S2 includes the following sub-steps:

[0032] In step S201 , a variable prediction model is constructed, assuming that the route-level bus carbon efficiency variable set S obeys a directed acyclic graph (DAG) structure consisting of a set of linear equations, and that each noise term is non-Gaussian and independent. The variable prediction model is used to represent each variable as the weighted sum of its parent node variables plus non-Gaussian independent noise.

[0033] The variable prediction model is expressed as follows:

[0034]

[0035] Where Pa(i) is the variable s i The parent node set of s (i.e., direct dependent variable), j represents the jth variable, b i,j is the variable s i With variable s j The causal effect coefficient between i is the variable s i Non-Gaussian independent noise.

[0036] Step S202 : converting the prediction model into a matrix equation form, wherein the variable matrix is ​​equal to the product of the inverse matrix of the identity matrix minus the causal effect matrix and the non-Gaussian independent noise matrix.

[0037] Specifically, the prediction model is converted into a matrix form and transposed to obtain:

[0038] S=(IB) -1 e

[0039] Where S is the variable matrix, B is the causal effect matrix, I is the identity matrix, and e is the non-Gaussian independent noise matrix.

[0040] Step S203 : Decompose the variable matrix so that the non-Gaussian independent noise matrix is ​​the product of the unmixing matrix and the variable matrix.

[0041] To solve the causal effect matrix B, this example uses the improved Fast Independent Component Analysis (FastICA) to decompose the variable matrix S into:

[0042] e=WS

[0043] Among them, W is the unmixing matrix, that is, IB.

[0044] Step S204 : Calculate the unmixing matrix, and obtain the causal effect matrix B by subtracting the unmixing matrix from the identity matrix, that is, B=IW; and establish a public transportation carbon efficiency causal directed graph based on the causal effect matrix, with variables as nodes and causal effects as edge weights.

[0045] Specifically, the causal effect matrix B is used as the adjacency matrix of the bus carbon efficiency causal directed graph, and the variables are used as nodes and the causal effects are used as edge weights to establish the bus carbon efficiency causal directed graph.

[0046] Furthermore, in this example, in order to improve the reliability of causal relationships and eliminate noise interference, an edge weight threshold parameter β is introduced. Specifically, only when |b i,j When |≥β, variable i is considered to have a direct causal effect on variable j; otherwise, it is considered noise or a weak influence and is eliminated. In this example, the edge weight threshold parameter β is based on the distribution characteristics of all causal effect coefficients and takes the 5% quantile of their absolute values ​​to ensure that only representative causal relationship edges are retained in the public transportation carbon efficiency causal directed graph.

[0047] Furthermore, in this example, the process of calculating the unmixing matrix includes:

[0048] In step S20401, the route-level bus carbon efficiency variable set S is subjected to multi-scale centering (removing the mean), standardization (unit variance), and high-order centering (skewness and kurtosis are zeroed) to proactively eliminate possible variable offsets, scale imbalances, and abnormal high-order statistical features, thereby enhancing the robustness and generalization capabilities of subsequent matrix decomposition.

[0049] Step S20402, randomly initialize an NS ×N S The orthogonal matrix is ​​used as the unmixing matrix W. For each component i=1,2,…,N S , N S is the number of samples of the route-level bus carbon efficiency variable set S, and the parallel iterations are carried out until convergence.

[0050] Introducing an adaptive weighted combination of multiple nonlinear functions (such as tanh, cubic, exponential, ReLU, etc.), defined as:

[0051]

[0052] Among them, each weight It can be dynamically adjusted according to the current variable characteristics to achieve adaptive characterization of the distribution of different variables. is a nonlinear function index, Indicates the A nonlinear function, Represents the adaptive weighted combination nonlinear function constructed for the i-th variable.

[0053] Update the i-th component w of the unmixing matrix W according to the following formula i The formula is:

[0054]

[0055] in, is the updated unmixing vector, is the kth sample of the route-level bus carbon efficiency variable set S, and E represents the mean of all samples. is the total number of samples, is the derivative of the nonlinear function of variable i.

[0056] In step S20403, after each update, the unmixing matrix W is orthogonalized using Gram-Schmidt orthogonalization with an adaptive momentum factor to prevent the components from collapsing into the same direction. Adaptive momentum is introduced to improve convergence speed and global optimal robustness. The momentum update formula is as follows:

[0057]

[0058] Among them, β is the momentum coefficient, which is dynamically adjusted using the Adam optimizer, and τ is the number of iteration steps.

[0059] Step S20404, repeat the above operation until all components w i If the changes in are very small (such as the inner product of two adjacent iterations is close to 1), stop the iteration, otherwise continue.

[0060] Step S3: Perform a post-nonlinear causal relationship test and / or a conditional independence causal test on the public transportation carbon efficiency causal directed graph to obtain an optimized public transportation carbon efficiency causal directed graph, thereby obtaining a key causal link of public transportation carbon efficiency.

[0061] Among them, if there is a directed edge e between node i and node j in the bus carbon efficiency causal directed graph i,j , then it is necessary to perform a post-nonlinear causality test on this edge. The specific steps are as follows:

[0062] Step S301A: For any directed edge to be tested in the public transportation carbon efficiency causal directed graph, a forward nonlinear causal model and a reverse nonlinear causal model are constructed respectively; in the forward nonlinear causal model, a reversible nonlinear mapping is used to map the causal variable and the noise term to the result variable; in the reverse nonlinear causal model, a reversible nonlinear mapping is used to map the result variable and the noise term to the causal variable.

[0063] In the forward nonlinear causal model, if it is assumed that there is a forward nonlinear causal relationship i→j, then there are reversible nonlinear functions h1, h2 and noise term E, such that:

[0064]

[0065] Among them, the noise term E is related to the input variable s i Independence, i represents the cause variable, s j Represents the outcome variable.

[0066] In the reverse nonlinear causal model, if it is assumed that there is a reverse nonlinear causal relationship j→i, then there exists a reversible nonlinear function 、 and the noise term E', such that:

[0067]

[0068] Among them, the noise term E' is related to the input variable s j Independence, i represents the cause variable, s j Represents the outcome variable.

[0069] Step S302A: adopting a neural network to model the reversible nonlinear function, and training the neural network so that the noise term is approximately Gaussian distributed and independent of the input variable.

[0070] Furthermore, in this example, the neural network may adopt a multi-layer perceptron.

[0071] In the forward nonlinear causal model, the estimated noise term E is:

[0072]

[0073] Among them, H1 and H2 are approximately represented by neural networks. The training goal is to make the noise term E approximate Gaussian distribution and i independent.

[0074] The loss function is designed as:

[0075]

[0076] Where N sample is the total number of samples, k represents the kth sample, represents the noise term of the kth sample, s j,k represents the kth sample s j The observed values ​​of the variable.

[0077] In the reverse nonlinear causal model, the noise term E' is estimated as:

[0078]

[0079] Among them, G1 and G2 are approximately represented by neural networks. The training goal is to make the noise term E' approximate Gaussian distribution and j independent.

[0080] The loss function is designed as:

[0081]

[0082] Where, represents the noise estimate of the kth sample in the reverse nonlinear causal model, s i,k represents the kth sample s i Variable value.

[0083] Step S303A: Perform statistical tests on the independence between the input variables and the estimated noise terms of the forward nonlinear causal model and the reverse nonlinear causal model, respectively. If the forward independence test passes but the reverse independence test fails, the directed edge is retained; otherwise, the directed edge is deleted from the public transportation carbon efficiency causal directed graph.

[0084] The neural network parameters are iteratively updated using the Adam optimization algorithm, and the grid search method is used to determine the optimal hyperparameters. i →s j With reverse s j →s i The independence between the input variables and the estimated noise term is statistically tested. The test method can be a mutual information test or a nonparametric test. In this example, the mutual information test is used as an example. i →s jp-value , the calculation expression is as follows:

[0085]

[0086]

[0087] Where, represents the mutual information independence test function used to test whether two variables are independent.

[0088] Reverse s j →s i p-value , the calculation expression is as follows:

[0089]

[0090]

[0091] Compare the p-values ​​of the forward and reverse tests to determine the causal direction: If (e.g. 0.05), and , it is believed that there exists s i →s j The significant nonlinear causal relationship of public transportation carbon efficiency causal directed graph retains the edge e i,j Otherwise, the causal relationship is considered invalid or unclear, and the edge e needs to be deleted from the public transportation carbon efficiency causal directed graph. i,j , that is, b i,j =0.

[0092] Step S304A: traverse each directed edge to be tested in the bus carbon efficiency causal directed graph, and output the line-level bus carbon efficiency causal graph after the post-nonlinear causal relationship test.

[0093] Among them, if there is a directed edge e between node i and node j in the bus carbon efficiency causal directed graph i,j , a conditional independence causal test needs to be performed on this edge, which includes the following steps:

[0094] Step S301B: for any directed edge to be tested in the bus carbon efficiency causal directed graph, extract all parent nodes of the two nodes in the bus carbon efficiency causal directed graph, merge the two groups of parent nodes and remove the two nodes themselves to form a condition set C. i,j .

[0095] Specifically, for any directed edge e to be tested in the causal directed graph of public transportation carbon efficiency, i,jFor the two connected nodes, node i and node j, extract all parent nodes (i.e. nodes that directly affect them) of node i and node j in the causal directed graph of public transportation carbon efficiency. Remove node i and node j from the union of their parent node sets to form the condition set C. i,j , the expression is as follows:

[0096]

[0097] Where Pa(i) is the variable s i The parent node set of , Pa(j) is the variable s j The parent node collection.

[0098] Step S302B: for each pair of samples, calculate the variable s i , variable s j , condition set C i,j The corresponding Gaussian kernel matrix.

[0099] Specifically, for each pair of samples k and k', the variable s is calculated separately. i , variable s j and condition set C i,j The corresponding elements of the Gaussian kernel matrix , and , the expression is as follows:

[0100]

[0101]

[0102]

[0103] in, 、 and The variables s i , variable s j and condition set C i,j Gaussian kernel width, s i,k Variable s representing the kth sample i The observed value, s i,k' Variable s representing the k'th sample j The observed value, C i,j,k Indicates that the kth sample is in the condition set C i,j The observed value vector of all variables on i,j,k' Indicates that the k'th sample is in the condition set C i,j The observed value vector of all variables in . To improve computational efficiency, the Gaussian kernel width is the median of the Euclidean distance of each sample pair.

[0104] Step S303B, based on condition set Ci,j The corresponding Gaussian kernel matrix , calculate the condition set C i,j The corresponding kernel projection matrix , used to eliminate the conditional set variables s in the kernel space i , variable s j The influence of the kernel projection matrix The expression is as follows:

[0105]

[0106] in, is the regularization parameter (to prevent the matrix from being irreversible, usually a small positive number such as 10 -3 ), where I represents the identity matrix.

[0107] Then, the kernel ridge regression method is used to transform the variable s i The corresponding residual kernel matrix M i and variable s j The corresponding residual kernel matrix M j Projection to condition set C i,j The space where the variable s is located i , variable s j The residual kernel matrix R i and R j , the expression is as follows:

[0108]

[0109]

[0110] Step S304B, based on variable s i Corresponding residual kernel matrix, variable s j The corresponding residual kernel matrix, calculate the variable s i and variable s j The conditional independence test statistic of the conditional independence test statistic is obtained by permutation test. If the p value is less than the threshold, it is considered that the conditional independence test statistic is i,j Next, the variable s i and variable s j There is a significant conditional dependency relationship between them, and the directed edge connecting node i and node j in the causal directed graph of public transportation carbon efficiency is retained.

[0111] Based on the variable s i The corresponding residual kernel matrix R i , variable s i The corresponding residual kernel matrix R j , calculate the variable and variable s j The conditional independence test statistic Ti,j , the expression is as follows:

[0112]

[0113] Where N sample represents the sample size, if , then T approaches 0.

[0114] The conditional independence test statistic T is obtained by permutation test i,j The corresponding p-value , the expression is as follows:

[0115]

[0116] Where, is the null hypothesis (under the condition set C i,j Next, the variable s i and variable s j The distribution of statistics under conditional independence.

[0117] like (take 0.05), then reject the null hypothesis and think that under the given condition set C i,j Next, the variable s i With variable s j There is a significant conditional dependency relationship between them, retaining the edge e in the causal directed graph of bus carbon efficiency. i,j .like , then the null hypothesis cannot be rejected, and it is believed that under the given condition set C i,j Next, the variable s i With variable s j Conditionally independent, the edge e should be deleted from the bus carbon efficiency causal directed graph i,j (ie b i,j =0).

[0118] Step S4: extract key causal links of public transportation carbon efficiency based on the optimized public transportation carbon efficiency causal directed graph and visualize them.

[0119] In this example, the causal directed graph of public transportation carbon efficiency that has passed the post-nonlinear causal relationship test and / or the conditional independence causal test is visualized. To highlight the strength of different causal relationships, the K-means clustering algorithm is first used to cluster all edge weights b. i,j Perform clustering to determine high causal effect threshold .when When , a thick solid line is used to connect node i and node j to highlight the strong causal relationship; when When b i,j= 0, no connection line is drawn between node i and node j, and invalid causal paths are eliminated.

[0120] By analyzing the carbon efficiency causal graph, we identify the key causal links that can significantly improve the carbon efficiency of the public transportation system. Taking the morning peak (7:00–8:00) as an example, Figure 3 As shown in the causal relationship diagram, the present invention reveals the key causal chain for improving the carbon efficiency of public transportation: "bus lane setting → subway connection index → ​​bus connection index → ​​fleet size → turnover → carbon efficiency". Specifically, the scientific configuration of bus lanes can effectively improve the connection level between multi-modal transportation and significantly optimize the seamless transfer capacity between bus and rail transportation. This process further guides the reasonable adjustment of fleet size and turnover by improving the subway connection index and bus connection index, forming a multi-level intermediary mechanism, thereby having an indirect and significant positive impact on the core indicator of bus carbon efficiency BEP. This causal chain clearly reveals that urban transportation intervention measures achieve a comprehensive regulation path for carbon efficiency through multi-level and complex coupling relationships between variables. Based on the above analysis, the present invention recommends: giving priority to focusing on the core links represented by the intermediary variables and formulating a systematic and full-process comprehensive intervention policy. Specifically, it includes: (1) systematically improving the multi-modal transportation connection system, optimizing the transfer convenience and connection efficiency between bus and rail transportation; (2) scientifically planning the layout of bus lanes, improving bus operation speed and service level, and stimulating the positive linkage of subsequent causal chain links.

[0121] It is worth emphasizing that the quantitative visualization of the causal directed graph of public transportation carbon efficiency not only provides a theoretical basis for the targeted and precise implementation and priority sorting of urban transportation intervention measures, but also serves as an important decision-making tool for policy simulation, intervention effect evaluation and dynamic optimization, thereby significantly improving the scientificity, systematicness and sustainability of the carbon efficiency improvement strategy of the public transportation system.

[0122] Step S5: Calculate the Global Causal Chain Evolution Index (GCCEI), the High Causal Chain Stability Coefficient (HCCSC), and / or the Local Causal Effect Variation (LCEV) to analyze the key causal links of public transportation carbon efficiency.

[0123] The calculation process of the global causal chain evolution index includes: calculating the average value of the difference in edge weights between two nodes during peak hours and off-peak hours; the expression is as follows:

[0124]

[0125] Where N S is the total number of variables, and The edge weights of the peak and flat-peak adjacency matrices, respectively. This indicator reflects the average change in the overall causal structure. The larger the value, the stronger the causal network dynamics.

[0126] For causal chains identified as "strong causal relationships" in any time period, the overlap ratio between the two time periods is defined as the strong causal chain stability coefficient HCCSC. The calculation process of the strong causal chain stability coefficient includes: obtaining the intersection of the strong causal relationship links of public transportation carbon efficiency during peak hours and the strong causal relationship links of public transportation carbon efficiency during off-peak hours, obtaining the union of the strong causal relationship links of public transportation carbon efficiency during peak hours and the strong causal relationship links of public transportation carbon efficiency during off-peak hours, and taking the ratio of the intersection to the union as the strong causal chain stability coefficient. The expression is as follows:

[0127]

[0128] Where, 、 The HCCSC is the set of “strong causal relationships” causal chains during peak and off-peak periods, respectively. A higher HCCSC indicates a better cross-period stability of the strong causal path structure.

[0129] For each directed edge The calculation process of the local causal effect variation coefficient includes: obtaining the difference between the edge weight between node i and node j during the peak period and the edge weight between node i and node j during the off-peak period, obtaining the maximum value of the edge weight between node i and node j during the peak period and the edge weight between node i and node j during the off-peak period, and taking the ratio of the difference to the maximum value as the local causal effect variation coefficient; the expression is as follows:

[0130]

[0131] Where, To prevent small positive numbers with zero denominators, represents the edge weight between node i and node j corresponding to the peak period, represents the edge weight between node i and node j during off-peak periods. LCEV reflects the sensitivity and variability of specific causal relationships across different time periods. Causal chain evolution and stability indicators, such as the GCCEI and HCCSC, are calculated for each city and for each key analysis period (e.g., peak and off-peak periods). Each city's indicator values ​​are spatially mapped to its geographic coordinate system to form a city-scale spatial distribution layer. Spatial autocorrelation methods, such as Moran's I and the Local Moran's Index (LISA), are used to assess the geographic clustering or dispersion of the GCCEI and HCCSC, identifying areas of significant positive or negative spatial correlation. Spatial hotspot detection methods, such as the Getis-Ord Gi*, are used to identify areas with significantly high (hotspots) and low (coldspots) GCCEI and HCCSC values, thereby locating cities or regions where the evolution of carbon efficiency causal chains is most active or stable.

[0132] Step S6: Analyze the Granger causality of bus carbon efficiency variables at the line level.

[0133] Specifically, to address the limitations of traditional Granger causality tests in multivariate, multi-line bus systems (e.g., they are only applicable to single threads or variable pairs, and have difficulty accurately revealing nonlinear and complex interactions between variables in the system), this paper proposes a Granger causality analysis method for analyzing bus carbon efficiency variables at the line level. This example can automatically model nonlinear, high-order causal, and temporal structures between variables, improving the ability to discriminate multivariate, multi-line causal relationships and their applicability. The specific implementation steps are as follows:

[0134] Step S601: Filter out the time series variable set of interest from the route-level bus carbon efficiency variable set S. ,in, Indicates the total number of time series variables. Each time series variable Both have The data structure represents the number of lines and number of time periods The time series variables include but are not limited to population density, operating speed, passenger load factor, departure interval, turnover volume and other bus operation and carbon efficiency related indicators.

[0135] For each time series variable , design and train a neural network prediction model. The input includes the values ​​of all relevant variables on all routes at different historical moments, and the output is the predicted value of the target variable for all routes at the current moment. The neural network structure can be selected based on the data characteristics, using architectures suitable for time series modeling, such as long short-term memory networks (LSTMs), temporal convolutional networks (TCNs), or multi-layer perceptrons (MLPs). The prediction relationship can be expressed as:

[0136]

[0137] in, For the The predicted value of the i-th variable at time t for each line, is the neural network mapping, and z is the maximum lag order.

[0138] To determine the jth time series variable For the i-th time series variable To determine whether there is a Granger temporal causal relationship, the following two types of neural network models are established for comparison:

[0139] The input is the historical lag information of all time series variables, and a full input model is established :

[0140]

[0141] Where, It represents the predicted value of the l-th line and the i-th variable at time t using a full-input neural network model that includes historical lag information of all time series variables. Represents the value set of all time series variables at each line and historical moment.

[0142] Input is to remove the tested variable All historical information of variables after the historical lag term, and the establishment of a variable removal model :

[0143]

[0144] Where, Indicates that the tested variable is not included The neural network predicts the value of the l-th line and the i-th variable at time t under the condition of historical lag information. Indicates the variable being tested The collection of all historical lag information.

[0145] Both models use Mean Square Error (MSE) as the loss function, which is as follows:

[0146]

[0147]

[0148] By comparing the difference in prediction errors between the two types of models, the quantitative index of Granger causal effect is defined:

[0149]

[0150] This indicator reflects the introduction of Historical information To test the improvement of prediction accuracy. The null hypothesis is set : Introduced The historical value will not increase significantly The prediction accuracy of Perform statistical tests and obtain the distribution and corresponding p-value under the null hypothesis by repeated sampling, which is recorded as . Set the statistical significance threshold (such as 0.05). , then the variable right There is a significant Granger temporal causality between all bus routes; otherwise, it is determined that there is no significant Granger causality between the two.

[0151] In summary, the present invention provides a method for identifying key causal chains of public transportation carbon efficiency, constructs a line-level public transportation carbon efficiency variable set, and the line-level public transportation carbon efficiency variable set includes control variables, covariates, and carbon emissions per unit passenger mile; the control variables are public transportation operation strategies; the covariates are factors affecting public transportation carbon efficiency; it is assumed that the line-level public transportation carbon efficiency variable set obeys a directed acyclic graph structure composed of a set of linear equations, and the causal effect matrix is ​​solved; a public transportation carbon efficiency causal directed graph is established with variables as nodes and causal effects as edge weights; the public transportation carbon efficiency causal directed graph is subjected to a post-nonlinear causal relationship test and / or a conditional independence causal test to obtain an optimized public transportation carbon efficiency causal directed graph, thereby accurately identifying key causal links of public transportation carbon efficiency.

[0152] Accordingly, the present application also provides an electronic device, comprising: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned method for identifying the key causal chain of public transportation carbon efficiency. Figure 4 As shown in the figure, a hardware structure diagram of any device with data processing capability is provided for the method for identifying the key causal chain of public transportation carbon efficiency provided by the embodiment of the present invention, except Figure 4 In addition to the processor, memory, and network interface shown, any device with data processing capabilities in which the apparatus in the embodiment is located may also include other hardware, generally based on the actual functions of the device with data processing capabilities, which will not be described in detail.

[0153] Accordingly, the present application also provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the above-mentioned method for identifying key causal chains of public transportation carbon efficiency. The computer-readable storage medium may be an internal storage unit of any device with data processing capabilities described in any of the aforementioned embodiments, such as a hard disk or memory. The computer-readable storage medium may also be an external storage device, such as a plug-in hard disk, a smart media card (SMC), an SD card, a flash card, etc. equipped on the device. Furthermore, the computer-readable storage medium may also include both an internal storage unit and an external storage device of any device with data processing capabilities. The computer-readable storage medium is used to store the computer program and other programs and data required by any device with data processing capabilities, and may also be used to temporarily store data that has been output or is to be output.

[0154] The above embodiments are intended only to illustrate the design concepts and features of the present invention. Their purpose is to enable those skilled in the art to understand the contents of the present invention and implement them accordingly. The scope of protection of the present invention is not limited to the above embodiments. Therefore, any equivalent changes or modifications made based on the principles and design concepts disclosed in the present invention are within the scope of protection of the present invention.

Claims

1. A method for identifying key causal chains of public transportation carbon efficiency, characterized by: The method comprises: Constructing a route-level bus carbon efficiency variable set, the route-level bus carbon efficiency variable set includes control variables, covariates, and carbon emissions per passenger mile; the control variables are bus operation strategies; and the covariates are factors affecting bus carbon efficiency; Assuming that the set of bus line-level carbon efficiency variables obeys a directed acyclic graph structure consisting of a set of linear equations, the causal effect matrix is ​​solved. A directed causal graph of bus carbon efficiency is established with variables as nodes and causal effects as edge weights. The post-nonlinear causal relationship test and / or conditional independence causal test are performed on the bus carbon efficiency causal directed graph to obtain the optimized bus carbon efficiency causal directed graph, thereby obtaining the key causal links of bus carbon efficiency.

2. The method for identifying key causal chains of public transportation carbon efficiency according to claim 1 is characterized in that: The process of establishing a causal directed graph of public transportation carbon efficiency includes: Assuming that the route-level bus carbon efficiency variable set obeys a directed acyclic graph structure consisting of a set of linear equations, a variable prediction model is constructed. The variable prediction model is used to represent each variable as the weighted sum of its parent node variables plus non-Gaussian independent noise. Converting the prediction model into a matrix equation form, where the variable matrix is ​​equal to the product of the inverse matrix of the identity matrix minus the causal effect matrix and the non-Gaussian independent noise matrix; Decompose the variable matrix so that the non-Gaussian independent noise matrix is ​​the product of the unmixing matrix and the variable matrix; The unmixing matrix is ​​calculated and the causal effect matrix is ​​obtained by subtracting the unmixing matrix from the identity matrix. Based on the causal effect matrix, a causal directed graph of public transportation carbon efficiency is established with variables as nodes and causal effects as edge weights.

3. The method for identifying key causal chains of public transportation carbon efficiency according to claim 2 is characterized in that: The process of solving the unmixing matrix includes: initializing an orthogonal matrix as the unmixing matrix; selecting a nonlinear function for weighted combination based on the line-level bus carbon efficiency variable; updating each component in the unmixing matrix based on the weighted combination of nonlinear functions, and forming new components through momentum update; judging whether the update amplitude of each component in the unmixing matrix is ​​lower than a preset threshold. If so, the iteration is stopped and the final unmixing matrix is ​​output.

4. The method for identifying key causal chains of public transportation carbon efficiency according to claim 1 is characterized in that: The process of conducting a post-nonlinear causal relationship test on the public transportation carbon efficiency causal directed graph includes: For any directed edge to be tested in the causal directed graph of public transportation carbon efficiency, a forward nonlinear causal model and a reverse nonlinear causal model are constructed respectively. In the forward nonlinear causal model, a reversible nonlinear mapping is used to map the causal variable and the noise term to the outcome variable. In the reverse nonlinear causal model, a reversible nonlinear mapping is used to map the outcome variable and the noise term to the causal variable. A neural network is used to model the reversible nonlinear mapping, and the noise term is made to be approximately Gaussian distributed and independent of the input variable by training the neural network; Conduct statistical tests on the independence between the input variables and the estimated noise terms of the forward nonlinear causal model and the reverse nonlinear causal model respectively. If the forward model passes the independence test but the reverse model fails, the current directed edge is retained; otherwise, the current directed edge is deleted from the public transportation carbon efficiency causal directed graph. Traverse each directed edge to be tested in the bus carbon efficiency causal directed graph, and output the line-level bus carbon efficiency causal graph after the post-nonlinear causal relationship test.

5. The method for identifying key causal chains of public transportation carbon efficiency according to claim 1 is characterized in that: The process of conditional independence causal test on the public transportation carbon efficiency causal directed graph includes: For any directed edge to be tested in the bus carbon efficiency causal directed graph, we extract all the parent nodes of the two nodes in the bus carbon efficiency causal directed graph, merge the two groups of parent nodes and remove the two nodes themselves to form the condition set C. i,j ; For each pair of samples, calculate the variable s separately i , variable s j , condition set C i,j The corresponding Gaussian kernel matrix; Based on the condition set C i,j The corresponding Gaussian kernel matrix, calculate the condition set C i,j The corresponding kernel projection matrix; The variable s i The corresponding Gaussian kernel matrix, variable s j The corresponding Gaussian kernel matrix is ​​projected onto the condition set C i,j The space where the variable s is located i Corresponding residual kernel matrix, variable s j The corresponding residual kernel matrix; Based on the variable s i Corresponding residual kernel matrix, variable s j The corresponding residual kernel matrix, calculate the variable s i and variable s j The conditional independence test statistic of the conditional independence test statistic is obtained by permutation test. If the p value is less than the threshold, it is considered that the conditional independence test statistic is i,j Next, the variable s i and variable s j There is a significant conditional dependency relationship between them, and the directed edge connecting node i and node j in the causal directed graph of public transportation carbon efficiency is retained; Traverse each directed edge to be tested in the bus carbon efficiency causal directed graph, and output the line-level bus carbon efficiency causal graph after the conditional independence causal test.

6. The method for identifying key causal chains of public transportation carbon efficiency according to claim 1 is characterized in that: The method further includes: visualizing the public transportation carbon efficiency causal directed graph that passes the post-nonlinear causal relationship test and / or the conditional independence causal test; including: Cluster all directed edge weights to set a high causal effect threshold; When the weight of the directed edge between node i and node j is greater than or equal to the high causal effect threshold, it indicates that there is a strong causal relationship between node i and node j in terms of public transportation carbon efficiency. The first mark is used to highlight the directed edge between node i and node j. When the weight of the directed edge between node i and node j is less than or equal to the high causal effect threshold, it means that there is a medium-intensity causal relationship between node i and node j in terms of public transportation carbon efficiency, and the second marker is used to highlight the directed edge between node i and node j; When the weight of the directed edge between node i and node j is equal to 0, it means that there is no causal relationship between node i and node j in terms of public transportation carbon efficiency, and the directed edge between node i and node j is not drawn.

7. The method for identifying key causal chains of public transportation carbon efficiency according to claim 1 is characterized in that: The method further includes: calculating a global causal chain evolution index, a strong causal chain stability coefficient, and / or a local causal effect variation coefficient to analyze key causal links of public transportation carbon efficiency; The calculation process of the global causal chain evolution index includes: calculating the average value of the edge weight difference between each node during peak hours and off-peak hours; The calculation process of the strong causal chain stability coefficient includes: obtaining the intersection of the strong causal relationship link of public transportation carbon efficiency during peak hours and the strong causal relationship link of public transportation carbon efficiency during off-peak hours, obtaining the union of the strong causal relationship link of public transportation carbon efficiency during peak hours and the strong causal relationship link of public transportation carbon efficiency during off-peak hours, and using the ratio of the intersection to the union as the strong causal chain stability coefficient; The calculation process of the local causal effect coefficient of variation includes: obtaining the difference between the edge weight between node i and node j during peak hours and the edge weight between node i and node j during off-peak hours, obtaining the maximum value of the edge weight between node i and node j during peak hours and the edge weight between node i and node j during off-peak hours, and taking the ratio of the difference to the maximum value as the local causal effect coefficient of variation.

8. The method for identifying key causal chains of public transportation carbon efficiency according to claim 1 is characterized in that: The method further comprises: Filter out the time series variable set containing the time dimension from the route-level bus carbon efficiency variable set; Determining whether a j-th time series variable has a Granger time series causal relationship with an i-th time series variable includes: constructing a first neural network model, wherein the input of the first neural network model is the historical lag information of all time series variables; constructing a second neural network model, wherein the input of the second neural network model is the historical information of all time series variables after removing the historical lag term of the j-th time series variable; using the difference between the loss function corresponding to the first neural network model and the loss function corresponding to the second neural network model as a quantitative indicator of the Granger causal effect; performing a statistical test on the quantitative indicator of the Granger causal effect to obtain a p-value; if the p-value is less than a statistical significance threshold, determining that the j-th time series variable has a Granger time series causal relationship with the i-th time series variable.

9. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores one or more computer programs executable by the at least one processor, and the one or more computer programs are executed by the at least one processor to enable the at least one processor to perform the method for identifying key causal chains of public transportation carbon efficiency according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When executed by a processor, the computer program implements the method for identifying key causal chains of public transportation carbon efficiency according to any one of claims 1 to 8.

Images