Influence factor evaluation method for special line logistics transportation demand
Through the PGTWR model combined with macro and micro analysis, the shortcomings of dedicated line logistics companies in spatiotemporal distribution characteristics analysis are solved, the accuracy of impact factor identification and evaluation is improved, theoretical guidance is provided, and enterprises are helped to optimize logistics planning and dedicated line layout, and solved the problem of Problems of supply and demand mismatch and duplicate configuration.
Patent Information
- Application Number
- CN202410488584.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-23
- Publication Date
- 2024-11-15
AI Technical Summary
Traditional dedicated line logistics companies lack professional theoretical knowledge and ignore the spatial and temporal distribution rules of dedicated line logistics transportation demand, resulting in mismatch between supply and demand and duplicate configuration of dedicated line. Existing research has failed to effectively consider the heterogeneity characteristics of spatiotemporal distribution.
Using the PGTWR model combined with macro and micro analysis, based on panel data and data visualization, the space-time visualization and autocorrelation test are performed by acquiring and preprocessing the transportation demand data, selecting bandwidth, building a spatiotemporal weight matrix, using the least squares method for parameter estimation, and finally Perform impact factor analysis and evaluation.
Improve the validity and accuracy of the identification and evaluation of the impact factors of dedicated logistics transportation demand, provide theoretical basis and technical solutions, help enterprises to carry out logistics planning and dedicated line integration, avoid the blindness of initial choice, and consider the impact factors among the Heterogeneity and dynamicity.
Smart Images

Figure CN118967205A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of evaluation of influencing factors of dedicated line logistics transportation demand, and in particular to an evaluation method of influencing factors of dedicated line logistics transportation demand. Background Art
[0002] Traditional dedicated logistics companies usually lack professional theoretical knowledge and rely on their own business experience to complete the layout of drop-off points and dedicated lines, ignoring the temporal and spatial distribution of dedicated logistics transportation demand, which leads to problems such as mismatch between supply and demand and duplicate configuration of dedicated lines in the dedicated business areas of enterprises. Existing studies have shown that logistics transportation demand and its influencing factors are heterogeneous and correlated in time and space, that is, the degree of influence of influencing factors on different regions is different, and the dedicated logistics transportation demand in adjacent regions will also affect the dedicated logistics transportation demand of the target. Although relevant studies have analyzed and evaluated the influencing factors of logistics demand, these methods all assume that the degree of influence of dedicated logistics transportation demand and its influencing factors in temporal and spatial distribution is consistent, and do not consider the heterogeneous characteristics of dedicated logistics transportation demand in temporal and spatial distribution. To solve this problem, the present invention proposes a method for evaluating the influencing factors of dedicated-line logistics transportation demand, and combines the PGTWR model to provide a set of theoretical basis and technical solutions from data processing of dedicated-line logistics transportation demand and its related influencing factors, selection of influencing factors to the final macro and micro analysis and evaluation of the influencing factors of dedicated-line logistics transportation demand, which can effectively improve the effectiveness and accuracy of identifying and evaluating the influencing factors of dedicated-line logistics transportation demand. Summary of the invention
[0003] The first purpose of the present invention is to solve the problems existing in the above-mentioned dedicated line logistics industry and to provide a method for evaluating the influencing factors of dedicated line logistics transportation demand. This method or scheme combines macro and micro analysis theories and data visualization analysis methods, and can provide a basic scheme and method guidance for dedicated line logistics companies and relevant management departments in the fields of dedicated line integration, expansion and dedicated line logistics planning.
[0004] The second purpose of the present invention is to solve the problems of the above-mentioned prior art and provide a method for evaluating the influencing factors of dedicated logistics transportation demand. The method is based on panel data and combined with the PGTWR model. Under the premise of taking into account the dedicated logistics transportation background, the actual transportation distance is used for the distance in the Gaussian function. At the same time, the direct and indirect mapping effects of the information of the influencing factors between the sample area and the target area are also taken into account.
[0005] In order to achieve the above object, the present invention is implemented through the following technical solutions:
[0006] 1. A method for evaluating the influencing factors of dedicated logistics transportation demand includes the following steps:
[0007] S1. Obtain transportation demand data of dedicated logistics, national economic data (GDP per capita, added value of secondary industry, per capita disposable income, wholesale and retail sales, local fiscal expenditure, etc.), logistics infrastructure data (highway mileage, etc.), logistics supply and industry scale data (highway freight volume, number of new logistics companies), and pre-process the above data, including modification of relevant data layout format, logarithmic processing, and multicollinearity test;
[0008] S2. Spatiotemporal visualization of dedicated logistics transportation demand: The transportation demand data of dedicated logistics is visualized in terms of space and time through ArcGIS and Origin tools, and the spatiotemporal distribution characteristics and operation logic of dedicated logistics transportation demand are analyzed from a macro perspective.
[0009] S3. Conduct spatial autocorrelation tests on dedicated logistics transport demand and one or more influencing factors such as per capita GDP, added value of the secondary industry, per capita disposable income, wholesale and retail sales, highway mileage, highway freight volume, number of new logistics companies, and local fiscal expenditure to verify the feasibility of using the PGTWR model.
[0010] S4. Selection of optimal bandwidth; considering the importance of bandwidth for weight calculation, an adaptive bandwidth calculation method is used instead of a fixed bandwidth calculation method. At the same time, the optimal bandwidth value is selected based on the AICc criterion, CV criterion, and GCV criterion respectively. Based on the optimal bandwidth value, the spatiotemporal weight matrix is calculated.
[0011] S5. Construct a spatiotemporal weight matrix. Considering the time and space spillover effects between regions and the transportation background of dedicated logistics, the actual transportation distance is used instead of the Euclidean distance calculation method. Then, the direct weight matrix and the indirect weight matrix are calculated separately. Finally, the two are combined through the Kronecker product method to construct the spatiotemporal weight matrix.
[0012] S6, construction of PGTWR model. Based on S3, S4, and S5, the least square method based on spatiotemporal weights is used for parameter estimation. The model is evaluated by indicators such as the significant proportion of local parameter estimation, adjusted goodness of fit, and F statistic probability value, and finally the model regression results are analyzed.
[0013] S7. Visualization of solution results. Based on the results of the model parameter estimation, the influence of different influencing factors in different regions on the demand for dedicated logistics transportation can be obtained, and a visualization analysis can be performed based on the parameter changes of the obtained relevant parameters.
[0014] The invention discloses an evaluation method for influencing factors of dedicated line logistics transportation demand, characterized in that: the method described is mainly used to analyze and identify the influence degree of influencing factors on dedicated line logistics transportation demand in different regions, so as to provide a theoretical guidance and analysis method for dedicated line logistics enterprises or relevant departments to determine and integrate logistics planning and dedicated lines.
[0015] Compared with the prior art, the beneficial effects of the present invention are:
[0016] 1. In terms of data processing, the present invention selects influencing factors with stronger correlation and lower collinearity as input explanatory variables and explained variables of the PGTWR model by regressing and testing the original data and logarithmic data one by one, thereby avoiding the blindness in the initial selection of influencing factors and data processing, and improving the fitting effect of the PGTWR model.
[0017] 2. The data used in this invention is based on panel data. Combined with the PGTWR model, it can consider the heterogeneity, temporal dynamics and conductivity of influencing factors in different regions, identify and analyze the influence of influencing factors on the demand for dedicated logistics transportation, and provide a solution for dedicated logistics companies, related industries and departments to analyze the influencing factors of dedicated logistics transportation demand. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] Figure 1 It is a flow chart of the present invention.
[0019] Figure 2 The data processing flow chart of the present invention is
[0020] Figure 3 The time characteristic distribution diagram of the special line logistics transportation demand of the present invention is
[0021] Figure 4 The spatial distribution diagram of the starting and ending points of the dedicated logistics transportation of the present invention is
[0022] Figure 5 The spatial distribution diagram of the dedicated logistics transportation path of the present invention is
[0023] Figure 6 The spatial distribution diagram of the dedicated logistics transportation demand of the present invention is
[0024] Figure 7 This is the LISA clustering diagram of the dedicated logistics transportation demand from 2018 to 2022 of the present invention
[0025] Figure 8 This is the spatial distribution map of the per capita GDP regression coefficient from 2018 to 2022 of the present invention
[0026] Fig. 9This is the spatial distribution map of the regression coefficient of the secondary industry added value from 2018 to 2022 of the present invention
[0027] Fig.10 This is the spatial distribution map of the regression coefficient of local fiscal expenditure from 2018 to 2022 of the present invention
[0028] Fig.11 This is the spatial distribution map of the regression coefficients of wholesale and retail sales from 2018 to 2022 of the present invention
[0029] Fig.12 This is the spatial distribution map of highway mileage regression coefficients from 2018 to 2022 of the present invention
[0030] Fig.13 This is the spatial distribution diagram of the regression coefficient of highway freight volume from 2018 to 2022 of the present invention
[0031] Fig.14 This is the spatial distribution diagram of the regression coefficient of the number of new logistics enterprises from 2018 to 2022 of the present invention DETAILED DESCRIPTION
[0032] The following is a detailed explanation in combination with the technical solution and the flow chart of the present invention.
[0033] S1. The data in this example mainly include the demand for dedicated logistics transportation, the starting and ending points of transportation, and the data on per capita GDP, added value of the secondary industry, per capita disposable income, wholesale and retail sales, highway mileage, highway freight volume, the number of new logistics companies, government fiscal expenditure, etc. are obtained through the national economic and social development bulletins of various counties and cities. The above data are laid out in the format of panel data, the initial data is logarithmized, the original data and the logarithmic data are uniformly regressed one by one, the regression coefficients are compared, and the collinearity test is performed. Based on the fitting effect and VIF value, the final indicators are determined. The data processing process is as follows: Figure 2 Variance inflation factor (VIF) calculation method: The variable data was tested for multicollinearity, and the influencing factors with VIF values greater than 10 were eliminated. Considering that the regression coefficient of urban per capita disposable income was small, the secondary indicators were also eliminated, and the following data processing results were obtained. Since the logarithmic correlation between the dedicated logistics transportation demand and the logarithmic per capita GDP, the added value of the secondary industry, wholesale and retail sales, highway mileage, highway freight volume data and the local fiscal expenditure of the original data, and the number of new logistics companies is better, the above data are used to build the PGTWR model.
[0034] Table 1 Data processing results
[0035]
[0036] Note: ***, **, * respectively indicate passing the 1%, 5%, and 10% significance level tests
[0037] Table 2 Multicollinearity test results
[0038]
[0039] Note: ***, **, * respectively indicate passing the 1%, 5%, and 10% significance level tests
[0040] S2. Use ArcGIS's XY dedicated line and graded color functions to visualize the spatial characteristics of dedicated line logistics transportation demand, including the distribution of transportation origin and destination points, transportation route distribution, transportation demand and frequency distribution; use origin to visualize the time characteristics, including annual distribution characteristics analysis and monthly characteristics analysis, and finally get the following Figure 3-Figure 6 As shown in the figure, we can see that there is a relatively obvious imbalance in transportation resources in southern Xinjiang, which is manifested in the scattered demand for dedicated logistics transportation, small demand and relatively high transportation frequency. Therefore, the following micro-analysis will be carried out on the Aksu and Kashgar regions in this area.
[0041] S3. Perform global and local spatial autocorrelation tests on the demand for dedicated logistics and one or more influencing factors such as per capita GDP, added value of the secondary industry, wholesale and retail sales, highway mileage, highway freight volume, number of new logistics companies, and local fiscal expenditure. and The global and local spatial autocorrelation tests were conducted to verify the feasibility of using the spatiotemporal geographic weighted regression model. The global and local spatial autocorrelations are shown in Table 3 and Figure 7 shown.
[0042] S3, analyze the spatiotemporal distribution characteristics of the dedicated logistics transportation demand from a macro perspective through ArcGIS and Origin tools. The next steps S3-S6 will analyze from a micro perspective, using the AICc criterion, GCV criterion and CV criterion to select the optimal bandwidth value, through the formula, the standard is the smaller AICc, GCV, CV values.
[0043] S4. Then, in order to make up for the spatial spillover effect between regional influencing factors ignored by the traditional GTWR model, direct weight matrix and indirect weight matrix are introduced, and the actual transportation distance of dedicated logistics is used instead of the calculation methods such as Euclidean distance and projection distance. The actual transportation distance of dedicated logistics can be obtained by using Amap API, calling the requests model through Python programming to request website data. Afterwards, they were respectively
[0044] The indirect spatial weight matrix and the indirect temporal weight matrix are calculated, and then the indirect spatiotemporal weight matrix is calculated by the Kronecker product, such as Similarly, through Get the direct spatiotemporal weight matrix.
[0045] Finally, according to the formula The final spatiotemporal weight matrix is obtained.
[0046] S5, construction of PGTWR model. Based on S2, S3, and S4, a weight matrix is introduced on the basis of the original least squares method, so that it can perform local analysis.
[0047] β l =(X {∈l} T STW T {∈l) STW {∈l) X {∈l} ) -1 X {∈l} T STW T {∈l) STW {∈l) y {∈l} , and finally constructed the PGTWR model lny {i,t} =β 0 (i,t)+β RJGDP (i,t)lnX RJGDP,{i,t} +β DECY (i,t)lnX DECY,{i,t} +...+β CZZC (i,t)lnX CZZC,{i,t} +μ {i,t} .
[0048] S6. Finally, the local regression coefficient results and model evaluation results of the PGTWR model of this example are solved using the matlab program package, and the model is evaluated using indicators such as local coefficient estimates, adjusted goodness of fit, and F statistics. The model evaluation results are shown in Table 4. In order to prove the superiority of the present invention, the fitting effects are compared with those of OLS, GWR, and GTWR, as shown in Table 5.
[0049] As shown in the model evaluation results, although the modified fitting effects of mixed effects and individual-fixed effects are better, the modified goodness of fit of individual effects also reaches 0.98, and its local parameter estimation significant proportion and log-likelihood value are also larger than those of mixed effects and individual-fixed effects, while the random perturbation variance, AICc, CV, and GCV estimates are also smaller than the above two effects. When the modified goodness of fit is not much different, it is necessary to compare evaluation indicators such as the significant proportion of local parameter estimates and random perturbation variance, so it can be proved that the PGTWR model under individual fixed effects has a better fitting effect for the dedicated logistics transportation demand and influencing factors. In order to prove the superiority of the PGTWR model adopted by the present invention in the evaluation of the influencing factors of the dedicated logistics transportation demand, it is compared with OLS, GWR, and traditional GTWR, and its modified goodness of fit is also greater than the above three traditional models.
[0050] S7. Finally, the PGTWR model based on individual fixed effects is used to evaluate the influencing factors of the transportation demand of dedicated logistics, and the results are visualized, as shown in Figure 8-Figure 14 As shown in Table 6, the maximum, minimum, median, mean and variance of each influencing factor were calculated by SPSS. Since the average value can usually represent the overall level of a set of data, the seven influencing factors are ranked according to the size of the average value, namely, GDP per capita>local fiscal expenditure>value added of the secondary industry>number of new logistics enterprises>highway freight volume>wholesale and retail sales>highway mileage, and the results are shown in Table 6.
[0051] Table 3 Global spatial autocorrelation results
[0052]
[0053]
[0054] Table 4 Model evaluation results
[0055]
[0056] Table 5 Model comparison results
[0057]
[0058] Table 6 Detailed results of impact factor evaluation
[0059]
[0060]
Claims
1. A method for evaluating the influencing factors of dedicated logistics transportation demand includes the following steps: S1. Obtain the transportation demand data of dedicated logistics, national economic data (GDP per capita, added value of the secondary industry, per capita disposable income, wholesale and retail sales, local fiscal expenditure, etc.), logistics infrastructure data (highway mileage, etc.), logistics supply and industry scale data (highway freight volume, number of new logistics companies), and pre-process the above data, including modification of the relevant data layout format, logarithmic processing, and multicollinearity test. S2. Spatiotemporal visualization of dedicated logistics transportation demand: The spatial and temporal characteristics of dedicated logistics transportation demand data are visualized through ArcGIS and Origin tools respectively, and the spatiotemporal distribution characteristics and operation logic of dedicated logistics transportation demand are analyzed from a macro perspective. S3. Conduct spatial autocorrelation tests on dedicated logistics transport demand and one or more influencing factors such as per capita GDP, added value of the secondary industry, per capita disposable income, wholesale and retail sales, highway mileage, highway freight volume, number of new logistics companies, and local fiscal expenditure to verify the feasibility of using the PGTWR model. S4. Selection of optimal bandwidth; considering the importance of bandwidth for weight calculation, an adaptive bandwidth calculation method is used instead of a fixed bandwidth calculation method. At the same time, the optimal bandwidth value is selected based on the AICc criterion, CV criterion, and GCV criterion respectively. Based on the optimal bandwidth value, the spatiotemporal weight matrix is calculated. S5. Construct a spatiotemporal weight matrix. Considering the time and space spillover effects between regions and the transportation background of dedicated logistics, the actual transportation distance is used instead of the Euclidean distance calculation method. Then, the direct weight matrix and the indirect weight matrix are calculated separately. Finally, the two are combined through the Kronecker product method to construct the spatiotemporal weight matrix. S6, PGTWR model construction; based on S3, S4, S5, the least square method based on spatiotemporal weights is used for parameter estimation. The model is evaluated by indicators such as the significant proportion of local parameter estimation, adjusted goodness of fit, and F statistic probability value, and finally the model regression results are analyzed. S7. Visualization of solution results. Based on the results of the model parameter estimation, the influence of different influencing factors in different regions on the demand for dedicated logistics transportation can be obtained, and a visualization analysis can be performed based on the parameter changes of the obtained relevant parameters.
2. A method for evaluating influencing factors of dedicated logistics transportation demand according to claim 1, characterized in that: The transportation demand of dedicated logistics is taken as the explained variable; It will be used to measure the regional economic development level, residents' consumption level, logistics infrastructure, logistics supply and industry scale, and policy environment indicators, such as per capita GDP, added value of the secondary industry, per capita disposable income, wholesale and retail sales, highway mileage, highway freight volume, number of new logistics companies, and one or more influencing factors in local fiscal expenditure will serve as explanatory variables.
3. According to the method for evaluating the influencing factors of dedicated logistics transportation demand described in claim 1, the S1 is specifically: Stack the explanatory variables and explained variables in the format of panel data; In order to reduce the impact of heteroscedasticity and volatility of the data, the data were logarithmized, and then the original data and the logarithmic data were regressed one by one, and the variables with stronger correlation were selected, and multicollinearity tests were performed, and the cycle was iterated until the influencing factors that met the experimental requirements were selected. The criteria are: Strong correlation and low collinearity; Variance inflation factor (VIF) was used to test for multicollinearity. In formula (1), VIF represents the variance inflation factor. The larger its value is, the more serious the multicollinearity between variables is. Traditionally, it is believed that VIF>7 indicates that there is a serious collinearity problem in the regression model.
4. A method for evaluating the influencing factors of dedicated logistics transportation demand according to claim 1, characterized in that The spatial and temporal characteristics of the dedicated logistics transportation demand data are visualized through ArcGIS and Origin tools, respectively, and the spatiotemporal distribution characteristics and operation logic of the dedicated logistics transportation demand are analyzed from a macro perspective. The S2 mentioned above is specifically: The annual regional dedicated logistics transportation demand data is visualized using ArcGIS and Origin tools to analyze the spatiotemporal heterogeneity and transportation rules of transportation demand from a macro perspective; The spatial characteristics of dedicated logistics transportation demand are visualized using ArcGIS's XY dedicated line and graded color functions, including the distribution of transportation origin and destination points, transportation route distribution, transportation demand and frequency distribution. The purpose is to analyze the relationship between the dedicated logistics service scope, frequency and demand. Origin is used to visualize time characteristics, including annual distribution characteristics analysis and monthly distribution characteristics analysis. The purpose is to analyze the annual and monthly trend changes in dedicated logistics transportation demand.
5. A method for evaluating the influencing factors of dedicated logistics transportation demand according to claim 1, characterized in that The spatial autocorrelation test is conducted on one or more influencing factors such as the dedicated logistics transportation demand and per capita GDP, added value of the secondary industry, per capita disposable income, wholesale and retail sales, highway mileage, highway freight volume, number of new logistics enterprises, and local fiscal expenditure. The S3 is specifically: Spatial autocorrelation test is a necessary step in adopting the PGTWR model. In order to explore whether there is a potential dependence relationship between the dedicated logistics transportation demand and one or more influencing factors such as per capita GDP, added value of the secondary industry, per capita disposable income, wholesale and retail consumption, highway mileage, highway freight volume, number of new logistics companies, and local fiscal expenditure in the sample area and the surrounding areas, the global and local Moran's I indexes are used to express them respectively. Where X in formula (2) and formula (3) i , W ij ,I Global ,I Local They represent the variable value of the i-th region, the variable mean, the weight matrix between the i-th region and the j-th region, the global Moran index, and the local Moran index. Global ∈[-1,1],I Global When it is a positive value, it means that the spatial distribution of the variable in the region shows a certain agglomeration, and the closer it is to 1, the stronger it is; otherwise, it shows a discrete distribution and imbalance, and its value is 0, which means that the spatial distribution of the variable in the region is random. Local >0 means that the level of a certain indicator in the region is high and the indicators in the surrounding areas are also high, or the level of a certain indicator in the region is low and the indicators in the surrounding areas are also low; I Local >0 means that the level of a certain indicator in the region is high, while the indicator in the surrounding areas is low, or the level of a certain indicator in the region is low, while the indicator in the surrounding areas is high.
6. According to the method for evaluating the influencing factors of the dedicated logistics transportation demand described in claim 1, the S4 is specifically: Considering the importance of bandwidth for weight calculation, and the fact that fixed bandwidth is not suitable for dispersed areas, adaptive bandwidth calculation is used instead of fixed bandwidth calculation. At the same time, the optimal bandwidth value is selected based on the AICc criterion, CV and GCV criteria respectively. The selection criteria are the smaller AICc value, CV value and GCV value. d , calculate the spatiotemporal weight matrix STW.
7. A method for evaluating the influencing factors of dedicated logistics transportation demand according to claim 1, characterized in that From a microscopic perspective, considering the time and space spillover effects between two regions and the transportation background of dedicated logistics, the actual transportation distance is used instead of the Euclidean distance calculation method, and then the direct matrix and the indirect matrix are calculated separately, and finally the two are combined into a spatiotemporal weight matrix through the Kronecker product method. The S5 described is specifically: The actual transportation distance calculation formula is: radlat 1,2 =(lat1*π) / 180-(lat2*π) / 180 (4) radlon 1,2 =(lone1*π) / 180−(lone2*π) / 180 (5) radlat1=lat1*π / 180 (6) Among them, formula (4) to formula (9) are the calculation methods of the actual distance of Amap. Dedicated logistics transportation from starting point n0 to end point n d spatial distance. The calculation formula of the time-space connection weight matrix is mainly shown in formulas (10)-(12): Where STW in formula (10) l,spillover 、TW l,spillover , SW l,spillover They represent the indirect standardized spatiotemporal weight matrix, time weight matrix, and space weight matrix of the target county / city area l of the dedicated logistics, respectively. is the Kronecker product symbol. In formula (11), f(.) represents the kernel function, which is used to characterize the first law of geography, that is, the farther the distance, the smaller the spatial influence effect or spatial spillover effect. The present invention adopts the Gaussian function. In formula (12), The global Moran'I index calculated based on all cross sections for the period corresponding to the transport origin and the period corresponding to the transport destination, t d , t o They respectively represent the period number corresponding to the starting point and the period number corresponding to the end point of the sample counties and cities under study. The calculation formula of the spatiotemporal direct weight matrix is mainly shown in equations (13)-(15): Where STW in formula (13) l,direct The spatiotemporal weight matrix represents the direct spatiotemporal impact of the sample county and city area {∈l} of the dedicated logistics on the target county and city area l, SW l,direct The sample county and city regions {∈l} representing the dedicated logistics are arranged according to the time period according to the regional cross-sections n {∈l} The spatial weight matrix that has a direct spatial impact on the target county or city area l. In addition, diag(.) is a new vector formed by extracting the main diagonal elements of the matrix in brackets. In formula (15), b is the bandwidth value, and sevc is the critical value of the spatial impact effect that can be tolerated by the spatial impact boundary. Finally, the calculation formula of the spatiotemporal weight matrix is shown in (16): Where STW in formula (10) {∈l) It represents the spatiotemporal weight matrix used when the sample county and city area {∈l} in dedicated logistics is holographically mapped to the target county and city area l. Num {∈l} The unit matrix of order, and Num {∈l} =N {∈l} *T {∈l} It represents the product of the number of cross sections of the county or city under study and the number of time periods to which it belongs.
8. A method for evaluating the influencing factors of dedicated logistics transportation demand according to claim 1, characterized in that: On the basis of S3, S4 and S5, the least square method based on time-space weight is used for parameter estimation. The model is evaluated by indicators such as the significant proportion of local parameter estimation, the goodness of fit of adjustment, and the probability value of the F statistic. Finally, the model regression results are analyzed. The specific S6 is as follows: The least squares method is used for parameter estimation. In addition, in order to verify the effectiveness of the model, the significant proportion of local parameter estimation, the adjusted goodness of fit, the probability value of the F statistic, etc. are used for verification. Finally, based on the results of parameter estimation, the relevance of the explanatory variables to the demand for dedicated logistics transportation is analyzed to determine the highly correlated influencing factors. b l =(X {∈l} T STW T {∈l) STW {∈l) X {∈l} ) -1 X {∈l} T STW T {∈l) STW {∈l) y {∈l} (17) flax {i,t} =β0(i,t)+β RJGDP (i,t)lnX RJGDP,{i,t} +β DECY (i,t)lnX DECY,{i,t} +...+β CZZC (i,t)lnX CZZC,{i,t} +μ {i,t} (18) Wherein, equations (17) and (18) represent the least squares estimation parameters and PGTWR model respectively.
9. A method for evaluating the influencing factors of dedicated logistics transportation demand according to claim 1, characterized in that According to the results of the model parameter estimation, the influence of different influencing factors in different regions on the dedicated logistics transportation demand can be obtained, and a visual analysis is performed based on the parameter changes of the obtained relevant parameters. The S7 is specifically: Visualization of relevant parameters, by sorting out the regression coefficients of each influencing factor, and using the graded color module in ArcGIS, the spatial distribution of the regression coefficients of each influencing factor from 2018 to 2022 is visualized. SPSS was used to statistically describe the regression coefficients of the influencing factors, calculate the maximum value, minimum value, median, mean value and variance, and rank the influence of the factors according to the size of the mean value and comprehensive consideration of other indexes.
