Grain trade data mining method and system based on cascading failure model

Abstract

The embodiment of the application discloses a kind of grain trade data mining method and system based on cascading failure model, wherein, method includes: obtaining the grain trade data of multiple regions;Grain trade data is uniformly converted into kilocalorie equivalent and is carried out multiple linear regression analysis, and linear regression model is obtained;Remove the regression coefficient and driving factor of low significance in model and take antilog, and obtain the first data model for reflecting the degree of trade preference between regions;The second data model of original export distribution proportion is corrected, and the third data model for reflecting the distribution proportion is obtained;Cascade failure model is run based on third data model and grain trade directed graph, simulate the change of grain supply in each region under different parameters, and real-time display the diffusion process and important data in grain trade directed graph.This embodiment can fully mine effective information in trade data, and be applied in the simulation analysis of cascading failure model, provide better data support.

Description

Technical Field

[0001] This invention relates to the field of data mining technology, and in particular to a method and system for mining grain trade data based on a cascading failure model. Background Technology

[0002] Grain trade networks are a crucial foundation for cross-regional food security. By analyzing the structure of grain trade networks (e.g., measuring network size, density, in-degree, and proximity centrality), we can fully leverage the characteristics of trade data to provide decision support for the construction and adjustment of trade infrastructure, such as building transport roads, optimizing transport routes, and improving transport distances.

[0003] Currently, the most commonly used method for trade network analysis is simulation analysis based on cascading failure models. However, most cascading failure models, when simulating real-world trade scenarios, typically focus on failures caused by node overload or underload, and analyze the simple linear propagation of shocks within the network. This fails to provide detailed and comprehensive data support for the construction and adjustment of trade infrastructure. One reason for this is the lack of data mining methods capable of extracting more effective data information from existing trade data and applying it to cascading failure model-based simulation analysis, thereby providing more effective data support for the construction and adjustment of trade infrastructure. Summary of the Invention

[0004] This invention provides a method and system for mining grain trade data based on a cascading failure model, which fully extracts effective information from trade data and applies it to simulation analysis based on the cascading failure model, providing more effective data support for the construction and adjustment of trade infrastructure.

[0005] In a first aspect, embodiments of the present invention provide a method for mining grain trade data based on a cascading failure model, comprising:

[0006] Obtain grain trade data from multiple regions;

[0007] The grain trade data was uniformly converted into kilocalories, and multiple linear regression analysis was performed on the converted data to obtain a linear regression model with the logarithm of inter-regional grain trade volume as the dependent variable and non-inertial driving factors of trade volume as independent variables. The non-inertial driving factors include whether the regions are adjacent, geographical distance, grain production, and grain reserves.

[0008] By removing the regression coefficients and driving factors with low significance from the linear regression model and taking the antinomial of the linear regression model, a first data model that reflects the degree of trade preference between regions is obtained, excluding inertial and random factors.

[0009] Using the first data model, the original second data model, which reflects the distribution ratio of export volume among various related import regions, is modified, and the degree of modification is adjusted by an exponential weight parameter to obtain a third data model that reflects the distribution ratio.

[0010] Based on the third data model and the directed graph of grain trade among the multiple regions, a cascade failure model is run to simulate the changes in grain supply in each region under different parameters. During the simulation, the diffusion process of export shocks in the directed graph of grain trade and important data are displayed in real time.

[0011] Secondly, embodiments of the present invention provide a grain trade data mining system based on a cascading failure model, comprising:

[0012] The data acquisition module is used to acquire grain trade data from multiple regions;

[0013] The data modeling module is used to convert the grain trade data into kilocalories and perform multiple linear regression analysis based on the converted data to obtain a linear regression model with the logarithm of inter-regional grain trade volume as the dependent variable and non-inertial driving factors of trade volume as independent variables. These non-inertial driving factors include whether regions are adjacent, geographical distance, grain yield, and grain reserves. The module removes regression coefficients and driving factors with low significance from the linear regression model and takes the antilogarithm of the model to obtain a first data model that reflects the degree of inter-regional trade preference, excluding inertial and random factors. Using this first data model, the original second data model, which reflects the distribution ratio of export volume among related importing regions, is modified, and the degree of modification is adjusted using exponential weight parameters to obtain a third data model that reflects the distribution ratio.

[0014] The data simulation module is used to run a cascade failure model based on the third data model and the directed graph of grain trade between the multiple regions to simulate changes in grain supply in each region under different parameters.

[0015] The data visualization module is used to display the diffusion process and key data of export shocks in the directional graph of grain trade in real time during the simulation.

[0016] In summary, the international grain trade data processing method based on an improved cascading failure model provided by this invention offers more effective information and richer auxiliary decision-making data for the simulation and analysis of the cascading failure model through effective data mining. Specifically, this is reflected in the following aspects:

[0017] (1) A multi-dimensional trade flow redistribution mechanism was constructed through data mining to enhance the model's realism: Existing technologies have relatively simple mechanisms for redistributing surplus trade volume, failing to systematically integrate the influence of multiple complex factors, and neglecting the unstable effects brought about by inertial and random factors. This embodiment quantifies the impact of multi-dimensional factors on trade flow through a multiple linear regression model, extracting important data to reflect the degree of trade preference. And this data is compared with the impact assignment weight parameters. This mechanism, applied in conjunction with cascading failure models, enables trade flow redistribution based on multi-dimensional stable driving factors. It transforms post-shock trade adjustments from simple proportional allocations to optimized allocations that better align with objective data patterns. This enhances the realism of model simulations and the accuracy of predictions, providing more effective guidance for the construction and adjustment of trade infrastructure.

[0018] (2) Improving prediction accuracy by comprehensively simulating dynamic shock propagation based on mined data: Existing technologies for simulating trade shocks generally simplify the shock propagation mechanism, focusing primarily on the static structure analysis of the network, making it difficult to capture the dynamic cascading effects of trade disruptions. This embodiment introduces an initial export reduction rate based on mined data. Impact transmission rate Reserve absorption rate and impact allocation weight parameters These four key parameters construct a more dynamic and realistic cascading failure model. This parameterized dynamic modeling approach allows this embodiment to meticulously depict the complex process of initial supply reduction in exporting regions propagating through trade networks, and effectively simulate the flow, attenuation, and secondary propagation of shocks between nodes. For example, through parameters... This allows for the quantification of the retransmission of shocks from importing regions to their trading partners, rather than simply terminating at the current point. This comprehensive modeling of dynamic propagation mechanisms significantly improves the predictive accuracy and explanatory power of dynamic responses in trade networks, providing more effective guidance for the construction and adjustment of trade infrastructure, thereby enabling more effective responses to real-world trade disruption scenarios.

[0019] (3) Unifying data measurement standards to improve data comparability and analytical accuracy: Existing technologies for processing grain trade data typically use indicators such as weight or value directly, failing to uniformly convert them to kilocalorie equivalents. This results in insufficient cross-regional data comparability and makes it difficult to accurately assess the nutritional value and actual supply capacity of different grain commodities. This embodiment innovatively converts all grain production, trade, and reserve data into kilocalorie equivalents. This standardized measurement method effectively solves the problem of inconsistent units of measurement caused by differences in grain types, making the overall analysis of grain supply and demand balance, food security risk assessment, and comparison of nutritional contributions among different grains more scientific and accurate. It eliminates the analytical complexity caused by differences in crop types and substitution effects, ensuring global data comparability, thereby improving the accuracy and applicability of data analysis. Attached Figure Description

[0020] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a flowchart of a grain trade data mining method based on a cascading failure model provided in an embodiment of the present invention;

[0022] Figure 2 This is a schematic diagram of the structure of a grain trade data mining system based on a cascading failure model provided in an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0024] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0025] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0026] Figure 1 This is a flowchart illustrating a grain trade data mining method based on a cascading failure model, provided by an embodiment of the present invention. This method is suitable for data analysis and simulation of cross-regional trade networks and is executed by electronic equipment. Figure 1 As shown, the method specifically includes:

[0027] S110: Obtain grain trade data from multiple regions.

[0028] This embodiment first identifies multiple regions as the analysis objects and obtains the trade volume of various grains in each region during the same time period, which serves as the data source for the entire method. The regions can be at any level, such as city, province, country, or other regional classification scales. This embodiment does not impose specific limitations, as long as all regions participating in the method are at the same level or the same classification scale. Trade data includes import and export data, where imports and exports are not limited to inter-country flows. Imports refer to flows from other regions to this region, and exports refer to flows from this region to other regions.

[0029] In one specific implementation, taking a national-level region as an example, with 2020-2023 as the target period, data sources are obtained as follows: production, trade, and reserve data for nine cereal commodities (barley, maize, millet, miscellaneous grains, oats, rice, rye, sorghum, and wheat) are obtained from the Food and Agriculture Organization of the United Nations (FAOSTAT) database and the USDA Production, Supply and Distribution (USDA-PSD) database. Reserve data for EU countries are allocated according to the proportion of cereal production.

[0030] S120. The grain trade data is uniformly converted into kilocalorie equivalents, and multiple linear regression analysis is performed based on the converted data to obtain a linear regression model with the logarithm of inter-regional grain trade volume as the dependent variable and non-inertial driving factors of trade volume as independent variables. The non-inertial driving factors include whether the regions are adjacent, geographical distance, grain production, and grain reserves.

[0031] The grain trade data obtained in the previous step is usually expressed in units of weight or value (such as US dollars), leading to insufficient comparability of cross-regional data and making it difficult to accurately assess the nutritional value and actual supply capacity of different grain commodities. Therefore, this embodiment converts all grain production, trade, and reserve data into kilocalorie equivalents. This standardized measurement method effectively solves the problem of inconsistent units of measurement caused by differences in grain types, making the overall analysis of grain supply and demand balance, food security risk assessment, and comparison of the nutritional contributions of different grains more scientific and accurate.

[0032] In the specific implementation method described above, which uses countries as regions, the trade volume of various grains can be converted to kilocalories according to the FAO standard (FAO 2001) and aggregated by region. That is, for every two regions, the converted trade volumes of all grains between the two regions are added together to obtain the final grain trade volume between the two regions. Furthermore, to smooth out annual fluctuations, the data can be averaged annually from 2020 to 2023.

[0033] In addition to grain data, this embodiment also extracts non-inertial drivers of trade volume for linear regression analysis. Non-inertial drivers refer to factors unrelated to trade inertia; for example, the previous year's trade volume is an inertial driver and should be excluded from this analysis.

[0034] In the above-mentioned specific implementation method based on the country as the region, 16 non-inertial driving factors were extracted based on five dimensions: geography, politics and culture, circulation policy, economy and resources, and country risk (adjacent, geographical distance, whether it is a landlocked country, whether it is an island country, common language, political differences, World Trade Organization membership, preferential trade agreement, free trade agreement, partial scope agreement, customs union, economic integration agreement, grain production, grain reserves, GDP, and INFORM risk index), as shown in Table 1; and specific data for each non-inertial driving factor were obtained from databases such as Dynamic Gravity Dataset (DGD), United Nations General Assembly Voting Data (UNGAVD), FAOSTAT, USDA-PSD, World Bank, and the Knowledge Centre for Disaster Risk Management (DRMKC).

[0035] Table 1

[0036]

[0037] It should be noted that, in order to ensure the accuracy and effectiveness of data mining, this embodiment considers political, cultural, and economic factors in addition to natural factors. However, this does not mean that this embodiment does not follow natural laws. On the contrary, it is precisely because this embodiment objectively and comprehensively considers non-inertial driving factors in various dimensions that it can fully explore relevant data patterns and more accurately and realistically reflect the trade status of the region. This embodiment follows data patterns, which is also a way of following natural laws.

[0038] Furthermore, based on the above data, this embodiment uses a multiple linear regression model to analyze the impact of driving factors on trade volume, and the following linear regression model can be constructed:

[0039] (1)

[0040] in, Indicates region To the region Grain trade volume, region To the region This indicates the direction of trade. For export regions, For importing regions; This indicates taking the natural logarithm. This indicates taking the logarithm of a constant; Indicates with region and / or region The relevant first A type of non-inertial driving factor that does not require taking a constant logarithm. express Quantity, express The corresponding regression coefficients; Indicates with region and / or region The relevant first Non-inertial driving factors that require taking the constant logarithm express Quantity, express The corresponding regression coefficients; This represents the intercept.

[0041] As can be seen, the linear regression model above divides the independent variables into two categories: one category consists of non-inertial driving factors that require logarithmic calculation (generally long-tailed data), and the other category consists of non-inertial driving factors that do not require logarithmic calculation. This division is based on the magnitude and variation patterns of the data itself, and is unrelated to which dimension the data belongs to: geography, politics and culture, circulation policy, economy and resources, or national risk. This embodiment employs this classification to adapt to the characteristics of the data itself in order to achieve linear regression analysis. After all, taking logarithms for all driving factors or not taking logarithms for any of them will not result in a well-fitted linear regression model. This also further demonstrates that the data patterns mined and followed in this embodiment are unrelated to whether the data belongs to non-natural dimensions such as politics, culture, policy, or economy.

[0042] In the above specific implementation method using countries as regions, the final linear regression model is as follows: (2)

[0043] in, It represents the country To the country The volume of grain trade (in kilocalories). For the intercept term, arrive The regression coefficients for each driving factor are given. Represents the country and country Adjacency Represents the country From capital to country The geographical distance of the capital Represents the country Is it a landlocked country? Represents the country Is it a landlocked country? Represents the country Is it an island nation? Represents the country Is it an island nation? Represents the country and country Is there a common official language? Indicates the countries whose voting records were extracted from the United Nations. and country political distance, Represents the country and country Are they all WTO members? Represents the country and country Are there any preferential trade agreements? Represents the country and country Is there a free trade agreement? Represents the country and country Are there any partial trade agreements? Represents the country and country Is it a customs union? Represents the country and country Is there an economic integration agreement? Represents the country Grain production, Represents the country Grain production, Represents the country Grain reserves, Represents the country Grain reserves, Represents the country GDP Represents the country GDP Represents the country INFORM Country Risk Index Represents the country The INFORM Country Risk Index.

[0044] S130. Remove the regression coefficients and driving factors with low significance from the linear regression model, and take the antinomial of the linear regression model to obtain a data model that reflects the degree of trade preference between regions, excluding inertial and random factors.

[0045] Based on the linear regression model in equation (1), this embodiment first removes regression coefficients and driving factors with significance below a set threshold, retaining only statistically significant regression coefficients and driving factors. This is to further eliminate random factors affecting trade volume and screen out more stable driving factors. Then, the inverse of the removed model is taken to obtain the following data model:

[0046] (3)

[0047] in, Indicates An index with a base; Indicates all The first one retained in the set formed Non-inertial driving factors, express Quantity, express The corresponding regression coefficients in the linear regression model; Indicates all The non-inertial driving factors retained in the formed set, express Quantity, express The corresponding regression coefficients in the linear regression model. Since the data model at this point has excluded trade inertia factors and low-significance random factors, what remains are the driving factors that have strong stability over trade volume. This comprehensive reflection of the combined effect of these driving factors on trade volume is representative of the region. To the region The degree of trade preference. This degree of preference is relatively objective and stable, and can better reflect the probability of trade between regions occurring without any trade inertia.

[0048] In the above specific implementation method based on countries as regions, the final data model is as follows:

[0049]

[0050] (4)

[0051] S140. Using the data model, the original data model reflecting the distribution ratio of export volume among various related import regions is modified, and the degree of modification is adjusted by exponential weight parameters to obtain the final data model of the distribution ratio.

[0052] This is a core data point mined in this embodiment. This embodiment uses this data to modify the data model in the traditional trade network cascading failure model that reflects the distribution ratio of export volume among various related import regions, in order to obtain a more objective and accurate distribution ratio. For ease of distinction and description, this embodiment refers to the data model corresponding to equations (3) and (4) as the first data model, the data model in the traditional trade network cascading failure model that reflects the distribution ratio of export volume among various related import regions as the second data model, and the final data model of the modified distribution ratio as the third data model.

[0053] Optionally, the second data model is:

[0054] (5)

[0055] in, Indicates the region before the correction. Export volume in related import regions The allocation ratio in the model. Therefore, this embodiment modifies the data model in the following way:

[0056] First, in the first data model With the second data model Weighted multiplication, and through exponential weight parameters Adjustment Weights are used to predict regional distributions. To various related import regions Non-inertial trade volume This trade volume includes the degree of trade preference. This consideration is more likely to reflect a relatively stable volume of non-inertial trade. This allows for adjustment of the weighting of considerations regarding trade preferences. The value can be determined based on the actual situation and needs.

[0057] Then, based on the relevant import regions This yields a third data model that reflects the allocation ratio:

[0058] (6)

[0059] in, Indicates the corrected region Export volume in related import regions The allocation ratio within the data. This trade flow redistribution mechanism based on data mining makes post-shock trade adjustments no longer a simple proportional allocation, but rather an optimized allocation that better aligns with data patterns.

[0060] S150. Based on the third data model and the directed graph of grain trade among the multiple regions, run the cascade failure model to simulate the changes in grain supply in each region under different parameters, and display the diffusion process and important data of export shocks in the directed graph of grain trade in real time during the simulation.

[0061] Based on the third data model, this embodiment also improves the cascading failure model by adding adjustable parameters such as the export reduction ratio in exporting regions, the impact propagation ratio and reserve absorption ratio in importing regions, and the exponential weighting parameter in the third data model. The introduction of these parameters constructs a more dynamic and realistic cascading failure model.

[0062] The meanings and values ​​of the four optional key parameters are as follows:

[0063] Export reduction rate This indicates the percentage decrease in export volume from the exporting region relative to the initial export volume. The value range is 10%-100%, and different values ​​can be taken in each simulation, with an increment step of 10%.

[0064] Impact propagation ratio This indicates the proportion of the time that an importing region passes on the reduction in imports to its trading partners. The value range is 10%-90%, and different values ​​can be used in each simulation, with an increment step of 10%.

[0065] Reserve absorption ratio This indicates the proportion of the region's reserves that absorb the shock. The value range is 10%-90%, and different values ​​can be used in each simulation, with an increment step of 10%.

[0066] Impact allocation weight parameters (i.e., the exponential weighting parameter in the point model): used to control the influence of driving factors in the impact distribution. Different values ​​can be taken in each simulation, such as 0, 0.25, 0.5, 0.75, 1, 1.5, and 2.

[0067] This embodiment, based on the third data model and the directed graph of grain trade among multiple regions, runs a cascading failure model to simulate changes in grain supply in different regions caused by different combinations of the aforementioned parameters. The directed graph of grain trade is a data structure for inter-regional trade networks and can also be visualized. This graph uses each region as a node, with production, trade, and reserve data for each region as node attributes, trade relationships between regions as edges, trade direction as edge direction, and trade volume as edge weight. If there is trade volume between two regions, an edge is constructed between the nodes of the two regions, pointing from the exporting region to the importing region, with the trade volume value serving as the edge weight. This embodiment obtains this graph based on the data acquired and transformed in S110 and S120.

[0068] Based on this directed graph, this embodiment performed multiple cascading failure model simulations. Each simulation used a single parameter combination, and the model's operation steps included:

[0069] S1. Select an export region and adjust the export volume of that region according to the parameter combination. Reduce, to obtain the reduced export volume of the region. The initial trade volume in the node attributes remains unchanged in the remaining regions.

[0070] S2. In the directed graph of grain trade, traverse each import region pointed to by each edge starting from the export region, and consider them as the associated import regions of the export region. .

[0071] S3, Combine the parameters Substitute these values ​​into the third data model to calculate the allocation ratio; then compare these ratios with the reduced export volume. Multiplying them together yields the allocation for each related importing region. .

[0072] S4, Read the parameter combination For each related importing region The allocation amount in this region In the process, a portion of the grain trade spreads to other regions pointed to by the edges originating from the respective related importing regions in the directed graph, while the other portion... Priority will be given to absorption by the region's total reserves. (in, (for the region's total reserves), the remaining portion This will directly impact the supply within the region.

[0073] S5. If the remaining portion of an import-related region is greater than 0, then the related import region is considered to be affected by the export region selected in S1.

[0074] S6. The allocated amount of each disseminated item will be used as the basis for each related importing region. For each decrease in exports, repeat operations S2 to S6 as described above; continue this process until the decrease in imports in the affected region falls below a set percentage of the initial imports, or reaches a boundary node of the grain trade directed graph, at which point the propagation terminates.

[0075] Using the same method, each exporting region is taken as the starting point of the impact (i.e., the source region of the reduction in exports), and simulations are performed separately for each parameter combination, resulting in multiple diffusion processes and simulation results. During each simulation, a visualization module can be used to display regions that have participated in the calculation, are currently participating in the calculation, and are not participating in the calculation, using different colors. When the export impact spreads to a certain region, that region is displayed in the corresponding color, thus dynamically demonstrating the diffusion process of the export impact and providing decision-makers with a more intuitive simulation effect.

[0076] Furthermore, data mining can be conducted based on the simulation results to provide more specific data support for the construction of trade infrastructure.

[0077] Optionally, after each simulation, the region can be quantified according to the following formula. Influence :

[0078] (7)

[0079] in, Indicates the region The number of regions whose supply is affected when exports decrease. For the region The number of direct export regions (i.e., regions) (Number of related import regions). For the region The decrease in total trade volume in a directed trade graph caused by a decrease in exports. Let be the initial total trade volume in the directed trade graph. For each simulation... Taking the average, we can obtain the regional... Ultimately, the impact.

[0080] For regions whose ultimate influence exceeds a set threshold, it can be recommended as an external objective to strengthen trade infrastructure construction and adjustment, that is, to strengthen infrastructure construction and adjustment between this region and these regions, as an auxiliary decision-making suggestion provided to experts.

[0081] Optionally, the region can also be quantified using the following formula. Vulnerability :

[0082] (8)

[0083] in, Indicates region In the region The surplus that failed to be absorbed by internal reserves due to reduced exports is also known as regional... In the region Regional supply exposure under export shocks For the region The initial import volume. For the region The risk index itself, where the numerator in the formula represents the risk spread from other regions to the region. The amount, For the region The number of simulations affected by the decrease in exports.

[0084] For regions whose vulnerability exceeds a set threshold, it can be suggested that they be targeted for strengthening their own trade infrastructure, that is, strengthening the infrastructure in these regions as an auxiliary decision-making recommendation for experts.

[0085] In summary, this embodiment provides a data processing method for international grain trade based on an improved cascading failure model. Through effective data mining, it offers more useful information and richer auxiliary decision-making data for the simulation and analysis of the cascading failure model. Specifically, this is reflected in the following aspects:

[0086] (1) A multi-dimensional trade flow redistribution mechanism was constructed through data mining to enhance the model's realism: Existing technologies have relatively simple mechanisms for redistributing surplus trade volume, failing to systematically integrate the influence of multiple complex factors, and neglecting the unstable effects brought about by inertial and random factors. This embodiment quantifies the impact of multi-dimensional factors on trade flow through a multiple linear regression model, extracting important data to reflect the degree of trade preference. And this data is compared with the impact assignment weight parameters. This mechanism, applied in conjunction with cascading failure models, enables trade flow redistribution based on multi-dimensional stable driving factors. It transforms post-shock trade adjustments from simple proportional allocations to optimized allocations that better align with objective data patterns. This enhances the realism of model simulations and the accuracy of predictions, providing more effective guidance for the construction and adjustment of trade infrastructure.

[0087] (2) Improving prediction accuracy by comprehensively simulating dynamic shock propagation based on mined data: Existing technologies for simulating trade shocks generally simplify the shock propagation mechanism, focusing primarily on the static structure analysis of the network, making it difficult to capture the dynamic cascading effects of trade disruptions. This embodiment introduces an initial export reduction rate based on mined data. Impact transmission rate Reserve absorption rate and impact allocation weight parameters These four key parameters construct a more dynamic and realistic cascading failure model. This parameterized dynamic modeling approach allows this embodiment to meticulously depict the complex process of initial supply reduction in exporting regions propagating through trade networks, and effectively simulate the flow, attenuation, and secondary propagation of shocks between nodes. For example, through parameters... This allows for the quantification of the retransmission of shocks from importing regions to their trading partners, rather than simply terminating at the current point. This comprehensive modeling of dynamic propagation mechanisms significantly improves the predictive accuracy and explanatory power of dynamic responses in trade networks, providing more effective guidance for the construction and adjustment of trade infrastructure, thereby enabling more effective responses to real-world trade disruption scenarios.

[0088] (3) Unifying data measurement standards to improve data comparability and analytical accuracy: Existing technologies for processing grain trade data typically use indicators such as weight or value directly, failing to uniformly convert them to kilocalorie equivalents. This results in insufficient cross-regional data comparability and makes it difficult to accurately assess the nutritional value and actual supply capacity of different grain commodities. This embodiment innovatively converts all grain production, trade, and reserve data into kilocalorie equivalents. This standardized measurement method effectively solves the problem of inconsistent units of measurement caused by differences in grain types, making the overall analysis of grain supply and demand balance, food security risk assessment, and comparison of nutritional contributions among different grains more scientific and accurate. It eliminates the analytical complexity caused by differences in crop types and substitution effects, ensuring global data comparability, thereby improving the accuracy and applicability of data analysis.

[0089] Figure 2 This is a schematic diagram of the structure of a grain trade data mining system based on a cascading failure model provided in an embodiment of the present invention. Figure 2 As shown, the system includes:

[0090] The data acquisition module is used to acquire grain trade data from multiple regions.

[0091] The data modeling module is used to convert the grain trade data into kilocalories and perform multiple linear regression analysis based on the converted data to obtain a linear regression model with the logarithm of inter-regional grain trade volume as the dependent variable and non-inertial driving factors of trade volume as independent variables. These non-inertial driving factors include whether regions are adjacent, geographical distance, grain yield, and grain reserves. The module removes regression coefficients and driving factors with low significance from the linear regression model and takes the antilogarithm of the model to obtain a first data model that reflects the degree of inter-regional trade preference, excluding inertial and random factors. Using this first data model, the original second data model, which reflects the distribution ratio of export volume among related importing regions, is modified, and the degree of modification is adjusted using exponential weight parameters to obtain a third data model that reflects the distribution ratio.

[0092] The data simulation module is used to run a cascading failure model based on the third data model and the directed graph of grain trade between the multiple regions, to simulate changes in grain supply in each region under different parameters;

[0093] The data visualization module is used to display the diffusion process and key data of export shocks in the directional graph of grain trade in real time during the simulation.

[0094] This embodiment has the same inventive concept as the above method embodiments, and any limitations in the above methods are applicable to this embodiment and can achieve the technical effects of any of the above embodiments.

[0095] Specifically, the data visualization model can display regions that have participated in the calculation, are currently participating in the calculation, and are not participating in the calculation using different colors during each simulation. When an export shock spreads to a certain region, that region is displayed in the corresponding color, thus dynamically demonstrating the spread process of the export shock, intuitively showing the spread path and interactions, and providing decision-makers with a more intuitive simulation effect. Simultaneously, it can also display important data of concern to decision-makers in real time. For example, it can be integrated with a GIS system to categorize the impact and vulnerability in the above simulation calculations and display them in different colors on the corresponding geographical locations, highlighting regions with impact and vulnerability exceeding a set threshold, and so on.

[0096] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A method for mining grain trade data based on a cascading failure model, characterized in that, include: Obtain grain trade data from multiple regions; The grain trade data were uniformly converted to kilocalorie equivalents, and multiple linear regression analysis was performed based on the converted data to obtain a linear regression model with the logarithm of inter-regional grain trade volume as the dependent variable and non-inertial driving factors of trade volume as independent variables: ,in, Indicates region To the region Grain trade volume This indicates taking the natural logarithm. This indicates taking the logarithm of a constant; Indicates with region and / or region The relevant first A type of non-inertial driving factor that does not require taking a constant logarithm. express Quantity, express The corresponding regression coefficients; Indicates with region and / or region The relevant first Non-inertial driving factors that require taking the constant logarithm express Quantity, express The corresponding regression coefficients; The intercept is indicated; the non-inertial driving factors include whether the regions are adjacent, geographical distance, grain yield, and grain reserves. By removing the regression coefficients and driving factors with low significance from the linear regression model, and taking the antilogarithm of the linear regression model, a first data model reflecting the degree of trade preference between regions is obtained, excluding inertial and random factors. ,in, Indicates region To the region The degree of trade preference Indicates Exponential operations with base 0; Indicates all The first one retained in the set formed Non-inertial driving factors, express Quantity, express The corresponding regression coefficients in the linear regression model; Indicates all The non-inertial driving factors retained in the formed set, express Quantity, express The corresponding regression coefficients in the linear regression model; Using the first data model, the original second data model, which reflects the distribution of export volume among various related importing regions, is modified, and the degree of modification is adjusted by an exponential weighting parameter to obtain a third data model that reflects the distribution ratio; wherein, the second model is... The third data model is , Indicates the region before the correction. Export volume in related import regions The allocation ratio in Indicates the corrected region Export volume in related import regions The allocation ratio in; Based on the third data model and the directed graph of grain trade among the multiple regions, a cascade failure model is run to simulate the changes in grain supply in each region under different parameters. During the simulation, the diffusion process of export shocks in the directed graph of grain trade and important data are displayed in real time.

2. The method according to claim 1, characterized in that, The acquisition of grain trade data from multiple regions includes: acquiring the trade volume of various grains in each of the multiple regions during the same time period; Accordingly, the process of uniformly converting the grain trade data into kilocalorie equivalents and performing multiple linear regression analysis based on the converted data to obtain a linear regression model with the logarithm of the inter-regional grain trade volume as the dependent variable and the non-inertial driving factors of the trade volume as the independent variables includes: converting the trade volume of various grains into data in kilocalories according to FAO standards; and for each pair of regions, adding the converted trade volumes of the various grains between the two regions to obtain the final grain trade volume between the two regions.

3. The method according to claim 1, characterized in that, The process involves using the first data model to modify the original second data model, which reflects the distribution of export volume among various related importing regions, and adjusting the degree of modification through exponential weight parameters to obtain a third data model that reflects the distribution ratio. This includes: In the first data model In the second data model Weighted multiplication, and through exponential weight parameters Adjustment Weights are used to predict regional distributions. To various related import regions Non-inertial trade volume ; According to each related importing region This yields a third data model that reflects the allocation ratio.

4. The method according to claim 1, characterized in that, The parameters in the simulation include the export reduction rate in exporting regions, the shock propagation rate and reserve absorption rate in importing regions, and the exponential weighting parameters in the third data model. Accordingly, the cascading failure model based on the third data model and the directed graph of grain trade among multiple regions is used to simulate changes in grain supply in various regions under different parameters, including: Based on the third data model and the directed graph of grain trade among the multiple regions, a cascading failure model is run to simulate the changes in grain supply in each region caused by different export regions under different parameter combinations.

5. The method according to claim 1, characterized in that, The grain trade directed graph uses each region as a node, and if there is trade volume between two regions, an edge is constructed between the nodes of the two regions pointing from the exporting region to the importing region. Accordingly, the cascading failure model based on the third data model and the directed graph of grain trade among multiple regions is used to simulate changes in grain supply in various regions under different parameters, including: In each simulation, based on the directed graph of grain trade among the multiple regions and the cascading failure model, the following operations are performed sequentially: Based on the export reduction ratio parameter, calculate the export volume of a certain export region after the reduction, while the initial trade volume of other regions remains unchanged. In the directed graph of grain trade, each import region pointed to by each edge starting from a certain export region is traversed to form an associated import region. Based on the third data model, calculate the distribution of the reduced export volume among the relevant import regions; Based on the shock propagation ratio parameter, the allocation of each related import region is propagated to other regions pointed to by the edge starting from each related import region in the grain trade directed graph, and the unpropagated allocation of each related import region is divided into the part that can be absorbed by internal reserves and the remaining part. If the remaining portion of a certain import-related region is greater than 0, it is determined that the related import region is affected by the decrease in exports from the certain export region. The allocated amounts distributed outwards are used as the export reduction amounts for each associated importing region, and the above operations continue to be performed; this process continues until the import reduction amount of the propagated region falls below the set proportion of the initial import amount, or reaches the boundary node of the grain trade directed graph, at which point the propagation terminates.

6. The method according to claim 1, characterized in that, Following the simulation of grain supply changes in various regions under different parameters, the following is also included: After each simulation, the region is quantified according to the following formula. Influence : in, Indicates the region The number of regions whose supply is affected when exports decrease. For the region The number of regions to which it directly exports. For the region The decrease in total trade volume in a directed trade graph caused by a decrease in exports. This represents the initial total trade volume in the directed trade graph; For each simulation Take the average to get the region Ultimate impact; and / or The region is quantified using the following formula. Vulnerability : in, Indicates region In the region The remaining portion that could not be absorbed by internal reserves due to reduced exports For the region Its own risk index, For the region The initial import volume. For the region The number of simulations affected by the decrease in exports.

7. The method according to claim 6, characterized in that, Also includes: For regions whose ultimate influence exceeds a set threshold, suggestions are made regarding strengthening trade infrastructure and adjusting external objectives, for expert decision-making. and / or For regions where vulnerability exceeds a set threshold, suggestions are made regarding strengthening and adjusting their own trade infrastructure, for expert decision-making. The construction and adjustment of trade infrastructure includes at least one of the following: building transport roads, optimizing transport routes, and improving transport distances.

8. A grain trade data mining system based on a cascading failure model, characterized in that, include: The data acquisition module is used to acquire grain trade data from multiple regions; The data modeling module is used to convert the grain trade data into kilocalories and perform multiple linear regression analysis based on the converted data to obtain a linear regression model with the logarithm of inter-regional grain trade volume as the dependent variable and non-inertial driving factors of trade volume as independent variables. ,in, Indicates region To the region Grain trade volume This indicates taking the natural logarithm. This indicates taking the logarithm of a constant; Indicates with region and / or region The relevant first A type of non-inertial driving factor that does not require taking a constant logarithm. express Quantity, express The corresponding regression coefficients; Indicates with region and / or region The relevant first Non-inertial driving factors that require taking the constant logarithm express Quantity, express The corresponding regression coefficients; The intercept is represented; the non-inertial driving factors include whether regions are adjacent, geographical distance, grain yield, and grain reserves; low-significance regression coefficients and driving factors are removed from the linear regression model, and the inverse coefficient of the linear regression model is taken to obtain a first data model that excludes inertial and random factors and reflects the degree of trade preference between regions: ,in, Indicates region To the region The degree of trade preference Indicates Exponential operations with base 0; Indicates all The first one retained in the set formed Non-inertial driving factors, express Quantity, express The corresponding regression coefficients in the linear regression model; Indicates all The non-inertial driving factors retained in the formed set, express Quantity, express The regression coefficients in the linear regression model; using the first data model, the original second data model reflecting the distribution ratio of export volume among various related import regions is modified, and the degree of modification is adjusted by an exponential weight parameter to obtain a third data model reflecting the distribution ratio; wherein, the second model is The third data model is , Indicates the region before the correction. Export volume in related import regions The allocation ratio in Indicates the corrected region Export volume in related import regions The allocation ratio in; The data simulation module is used to run a cascade failure model based on the third data model and the directed graph of grain trade between the multiple regions to simulate changes in grain supply in each region under different parameters. The data visualization module is used to display the diffusion process and key data of export shocks in the directional graph of grain trade in real time during the simulation.

Images