Supply chain risk assessment system based on multi-source data fusion

Abstract

The invention relates to the technical field of supply chain risk assessment, in particular to a supply chain risk assessment system based on multi-source data fusion, which comprises a data acquisition module, a configuration multi-source data acquisition end and establishment of a binding relationship between the acquisition end and each node in a database, and is characterized in that the database is used for storing multi-source data related to supply chain risk assessment; the data fusion module is used for constructing a knowledge graph based on the multi-source data, nodes of the knowledge graph correspond to risk influence factors, and edges of the knowledge graph represent the incidence relation between the factors; the model construction module is used for constructing a risk assessment model based on the knowledge graph; the risk assessment module performs risk assessment on the supply chain based on the risk assessment model to obtain a risk assessment result; by fusing the multi-source data, the change condition of the multi-source data is monitored in real time, the dynamic change in the supply chain is responded in real time, the evaluation strategy is adjusted in time, and the accuracy and timeliness of the evaluation result are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of supply chain risk management, and particularly to a supply chain risk assessment system based on multi-source data fusion. Background Technique

[0002] A supply chain refers to a functional network chain structure organization that connects raw material suppliers, product manufacturers, logistics transportation enterprises, warehousing managers, product distributors, and end consumers into a whole around a core enterprise. Supply chain financial risk refers to the possibility of losses to participating entities caused by unpredictable internal uncertainties or external interferences in the supply chain financial financing process. In supply chain management, risk assessment is a key link to ensure the stable operation of the supply chain and reduce potential risks. As the supply chain becomes increasingly complex, the involved data sources are more diverse, including supplier information, logistics data, market demand data, etc. Traditional supply chain risk assessment methods often rely on single data sources or simple statistical models, making it difficult to comprehensively and accurately reflect the actual risk situation of the supply chain, thus affecting the efficiency and accuracy of the assessment.

[0003] Therefore, we propose a supply chain risk assessment system based on multi-source data fusion to solve the above problems. Summary of the Invention

[0004] The purpose of the present invention is to provide a supply chain risk assessment system based on multi-source data fusion to solve the problems raised in the above background technique.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A supply chain risk assessment system based on multi-source data fusion, including: A data acquisition module, which configures multi-source data acquisition terminals, establishes a binding relationship between the acquisition terminals and each node in the database. Among them, the database is used to store multi-source data related to supply chain risk assessment; classifies the multi-source data and stores them in a database composed of multiple sub-databases according to categories; binds the sub-databases to the acquisition terminals of the corresponding category of basic data; A data fusion module, which obtains the classified basic data and the corresponding sub-databases from the database; extracts the basic data features as risk impact factors, obtains the association relationships as the association relationships between the factors, sets the knowledge graph nodes and edges, and constructs a knowledge graph based on the nodes and edges; A model construction module, which takes the knowledge graph nodes as the center, defines an association area including a main association area and a secondary association area; sets a virtual boundary layer, establishes a communication channel between the acquisition terminal and the node, and establishes a driving relationship between the sensing point and the risk assessment model; determines the change nodes and dynamically adjusts the association area; evaluates the association area, and trains the evaluation model with the comprehensive results; The risk assessment module obtains the knowledge graph formed by the fusion of multi-source data; determines the associated areas, conducts risk assessment on the supply chain based on the risk assessment model, and obtains the risk assessment results.

[0006] Preferably, the data acquisition module includes: Collection unit: Set up collection points at the data sources of multi-source data, register collection terminals at the corresponding collection points, collect basic data for supply chain risk assessment based on the collection terminals, and obtain multi-source data for supply chain risk assessment; Storage unit: classifies multi-source data to obtain multiple categories of basic data, and stores the multiple categories of basic data in a database by category, wherein the database includes multiple sub-databases, each of which corresponds to a category of basic data; Establishing unit: Binding the sub-database with the basic data collection end of the corresponding category.

[0007] Preferably, the data fusion module includes: The first acquisition unit is used to acquire and extract the classified basic data and the sub-database corresponding to the basic data from the database; The data fusion unit extracts the data features of a type of basic data and the corresponding sub-database as a risk impact factor, and obtains multiple risk impact factors based on multiple types of basic data; obtains the correlation between each type of basic data, and uses the correlation between each type of basic data as the correlation between factors; The graph construction unit sets multiple nodes of the knowledge graph corresponding to multiple risk influencing factors, sets multiple edges of the knowledge graph corresponding to the association relationship between the factors, and constructs the knowledge graph based on the multiple nodes and the multiple edges.

[0008] Preferably, the model building module includes: Region determination unit: Take each node in the knowledge graph as the center point and the node as the center point as the target node; define all nodes and edges directly associated with the target node as the associated region of the target node, where the associated region includes the primary associated region and the secondary associated region; Boundary construction unit: set a virtual boundary layer with sensing points in the knowledge graph, where the sensing points correspond to the nodes one by one; establish a communication channel between the acquisition end and the corresponding node, where each communication channel corresponds to a sensing point; establish a driving relationship between the sensing point and the risk assessment model; Area adjustment unit: determines the changed nodes to obtain the changed nodes, and dynamically adjusts the associated areas based on the changed nodes; Model construction unit: Conduct separate risk assessments for each associated area, identify and eliminate the associated areas with duplicate calculations, integrate the assessment results of each associated area to obtain the risk assessment result corresponding to the knowledge graph, and train the assessment model based on the risk assessment results corresponding to multiple historical knowledge graphs.

[0009] Preferably, the step of defining all the nodes and edges directly associated with the target node as the associated area of the target node, where the associated area includes the main associated area and the secondary associated area, includes: Set influence thresholds for the corresponding risk impact factors respectively; Take the risk impact factors exceeding the influence threshold as the core risk factors; Take the risk impact factors not exceeding the influence threshold as the auxiliary risk factors; Take the nodes corresponding to the core risk factors as the main nodes, and take the nodes corresponding to the auxiliary risk factors as the secondary nodes; Obtain the associated areas corresponding to multiple main nodes and secondary nodes respectively to obtain multiple main associated areas and secondary associated areas before adjustment.

[0010] Preferably, the step of determining the changed nodes to obtain the changed nodes and dynamically adjusting the main associated area and the secondary associated area based on the changed nodes includes: Collect the changed nodes based on the sensing points to obtain the active nodes, and record the time interval between two consecutive changes of each active node; For each active node, judge whether the latest change interval exceeds the preset duration threshold; Obtain the active nodes corresponding to the latest change interval exceeding the preset duration threshold, mark them as the changed nodes, and take the data at the last change as the basic data corresponding to the nodes; Take the associated area corresponding to the changed nodes as the main associated area after adjustment; Take the nodes that have not changed as the non-changed nodes, and take the associated area corresponding to the non-changed nodes as the secondary associated area after adjustment.

[0011] Preferably, the risk assessment module includes: Second acquisition unit: Obtain multi-source data for supply chain risk assessment from the database and integrate them into a knowledge graph; Risk assessment unit: Determine the associated areas in the knowledge graph, and conduct risk assessment on the associated areas based on the assessment model to obtain the risk assessment result, where the associated areas in the knowledge graph include the associated areas before adjustment and the associated areas after adjustment.

[0012] Preferably, the risk assessment unit includes: Comparison subunit: used to count the number of changed nodes, calculate the proportion of the changed nodes in the knowledge graph based on the number of changed nodes to obtain the proportion of changed nodes; compare the proportion of changed nodes with a preset proportion threshold. Initial evaluation subunit: regard the number of changed nodes corresponding to the proportion of changed nodes exceeding the preset proportion threshold as the number of abnormal changed nodes, where the changed nodes corresponding to the number of abnormal changed nodes are recorded as abnormal changed nodes; for abnormal changed nodes, conduct risk assessment based on the main associated area and the secondary associated area before adjustment. Secondary evaluation subunit: regard the number of changed nodes corresponding to the proportion of changed nodes not exceeding the preset proportion threshold as the number of normal changed nodes, where the changed nodes corresponding to the number of normal changed nodes are recorded as normal changed nodes; for normal changed nodes, conduct risk assessment based on the main associated area after adjustment and the secondary associated area before adjustment.

[0013] Compared with the prior art, the beneficial effects of the present invention are: 1. By establishing the binding relationship between the acquisition end and each node in the database, it is possible to comprehensively and accurately collect multi-source data related to supply chain risk assessment, integrate the multi-source data, visually present the risk impact factors and the correlation relationships between the factors, dynamically adjust the associated area according to the changed nodes, be able to respond in real time to the dynamic changes in the supply chain, timely adjust the assessment strategy, and improve the accuracy and timeliness of the assessment results. 2. During the definition process of the associated area, set the impact threshold for each corresponding risk impact factor respectively, regard the risk impact factor exceeding the impact threshold as the core risk factor, and the one not exceeding as the auxiliary risk factor, highlighting the key risks, making the risk assessment more targeted, and improving the assessment efficiency and accuracy; conduct assessments based on the associated area before and after adjustment respectively to ensure the comprehensiveness and accuracy of the assessment results. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0015] Figure 1 It is the system structure block diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0017] Embodiment. Please refer to Figure 1 , the present invention provides a technical solution for a supply chain risk assessment system based on multi-source data fusion: A supply chain risk assessment system based on multi-source data fusion includes: A data collection module, which configures multi-source data collection terminals, establishes a binding relationship between the collection terminals and each node in the database. Among them, the database is used to store multi-source data related to supply chain risk assessment; classifies the multi-source data and stores it in a database composed of multiple sub-databases according to categories; binds the sub-databases to the collection terminals of the corresponding category of basic data; The data collection module includes: A collection unit: Arranges collection points at the data sources of multi-source data, registers collection terminals corresponding to the collection points, and respectively collects basic data for supply chain risk assessment based on the collection terminals to obtain multi-source data for supply chain risk assessment; A storage unit: Classifies the multi-source data to obtain multiple categories of basic data, and stores the multiple categories of basic data in the database according to categories. Among them, the database includes multiple sub-databases, and each sub-database corresponds to one category of basic data; An establishment unit: Binds between the sub-databases and the collection terminals of the corresponding category of basic data; Specifically, multi-source data can include internal supply chain data (such as operation data, financial data), external data (such as market data, policy data), supplier data (such as supplier performance, financial status), logistics data (such as transportation status, warehousing data), etc. Acquisition terminals are arranged separately according to the sources of the basic data, and various types of basic data are collected respectively through the acquisition terminals, and the basic data is preprocessed. The preprocessing includes normalizing it. For example, if different systems use different names for the same data, a mapping table can be established to convert the aliases into standard field names. The normalization process at least includes processing of naming, unit, format, value range, etc. By analyzing the differences in field naming, unit, and format of each data source, corresponding mapping rules are established and batch conversion is performed to ensure the consistency of multi-source data, thus facilitating the subsequent use of the data. After the normalization process, the processed multi-source data is classified, and each type corresponds to a sub-database for separate storage, which can facilitate the risk impact factors corresponding to the nodes when constructing a knowledge graph later; binding the acquisition terminal to the corresponding sub-data and establishing a communication channel between the two is convenient for subsequent updating of the basic data in the sub-database when the data collected by the acquisition terminal changes, and judging whether there is an update of the basic data through the activation of the communication channel, and then judging whether the nodes of the knowledge graph have changed; A data fusion module, which obtains the classified basic data and the corresponding sub-databases from the database; extracts the basic data features as risk impact factors, obtains the association relationships as the association relationships between the factors, sets the nodes and edges of the knowledge graph, and constructs the knowledge graph based on the nodes and edges; The data fusion module includes: A first acquisition unit: obtains and extracts the classified basic data and the sub-databases corresponding to the basic data from the database; A data fusion unit, which extracts the data features of a type of basic data and the corresponding sub-database as a risk impact factor, and obtains multiple risk impact factors based on multiple types of basic data; obtains the association relationships between each type of basic data, and takes the association relationships between each type of basic data as the association relationships between the factors; A graph construction unit, which sets multiple nodes of the knowledge graph corresponding to multiple risk impact factors, sets multiple edges of the knowledge graph corresponding to the association relationships between the factors, and constructs the knowledge graph based on the multiple nodes and multiple edges; Specifically, classify the preprocessed multi-source data, and each type corresponds to a sub-database for separate storage (such as an operation data sub-database, a financial data sub-database, etc.). Bind the acquisition end to the corresponding sub-data and establish a communication channel between the two for subsequent data updates; obtain and extract the classified basic data and the corresponding sub-database from the database. Extract the data features of a class of basic data (such as the "order volume" feature in operation data) and the corresponding sub-database (such as the operation data sub-database) as a risk impact factor. Based on multiple classes of basic data, obtain multiple risk impact factors (such as "order volume", "sales amount", "inventory turnover rate", etc.). Analyze the correlation relationships between each class of basic data (such as the positive correlation between "order volume" and "sales amount"). Take the correlation relationships between each class of basic data as the correlation relationships between the factors. Set multiple nodes of the knowledge graph corresponding to multiple risk impact factors (such as "high order volume", "growing sales amount", "low inventory turnover rate", etc.). Set multiple edges of the knowledge graph corresponding to the correlation relationships between the factors (such as a positive correlation edge between "high order volume" and "growing sales amount"). Construct a knowledge graph based on multiple nodes and multiple edges to display different risk impact factors and their correlation relationships.

[0018] Specifically, for example, assume there is a supply chain management system that needs to construct the following knowledge graph: Nodes: high order volume, low order volume, growing sales amount, declining sales amount, high inventory turnover rate, low inventory turnover rate, excellent supplier performance, poor supplier performance; Edges: high order volume —— growing sales amount (indicating that when the order volume is high, the sales amount often grows), low inventory turnover rate —— low order volume (indicating that when the inventory turnover rate is low, the order volume often is also low), excellent supplier performance —— high order volume (indicating that when the supplier performance is excellent, the order volume often is high); Construct a graph that includes the above nodes and edges to display different risk impact factors and their correlation relationships. Through the knowledge graph, potential risk points in the supply chain can be intuitively identified (such as "low inventory turnover rate" may lead to "low order volume"). When the data collected by the acquisition end changes, it can automatically update the basic data in the sub-database and determine whether the nodes of the knowledge graph have changed through the communication channel to achieve real-time monitoring and early warning; The model construction module, centered on the knowledge graph nodes, defines the association area including the main association area and the secondary association area; sets the virtual boundary layer, establishes the communication channel between the acquisition end and the nodes, and establishes the driving relationship between the induction points and the risk assessment model; determines the changed nodes and dynamically adjusts the association area; evaluates the association area and trains and evaluates the model with the comprehensive results. The model construction module includes: Region determination unit: Taking each node in the knowledge graph as the center point, the node serving as the center point is used as the target node; all nodes and edges directly associated with the target node are defined as the associated region of the target node, where the associated region includes the primary associated region and the secondary associated region; Boundary construction unit: Set a virtual boundary layer with induction points in the knowledge graph, where the induction points correspond to the nodes one by one; establish a communication channel between the acquisition end and the corresponding node, where each communication channel corresponds to an induction point; establish a driving relationship between the induction point and the risk assessment model; Region adjustment unit: Determine the changed nodes to obtain the changed nodes, and dynamically adjust the associated region based on the changed nodes; Model construction unit: Conduct separate risk assessments on each associated region, identify and eliminate the repeatedly calculated associated regions, synthesize the assessment results of each associated region to obtain the risk assessment result corresponding to the knowledge graph, and train the assessment model based on the risk assessment results corresponding to multiple historical knowledge graphs; It should be noted that for each associated region, a separate risk assessment is carried out, duplicate - calculated associated regions are identified and excluded, and the assessment results of each associated region are integrated to obtain the specific content of the risk assessment result corresponding to the knowledge graph: The assessment results of each associated region are integrated to obtain the risk assessment result of the entire knowledge graph. The integration method can be simple averaging, weighted averaging, or a more complex fusion algorithm, depending specifically on the assessment objective and data characteristics; Based on the integrated assessment results, a first - assessment model is constructed. This model can be a mathematical formula, a decision tree, a neural network, or other forms of models, depending specifically on the complexity and requirements of the assessment task. For example: Suppose there is a supply - chain risk knowledge graph, which contains the following nodes and edges: Nodes: A (Supplier performance), B (Order volume), C (Inventory turnover rate), D (Sales volume); Edges: A - B (Supplier performance affects order volume), B - C (Order volume affects inventory turnover rate), C - D (Inventory turnover rate affects sales volume); The associated region centered on node A: {A, B, A - B}; The associated region centered on node B: {B, A, C, A - B, B - C} (note the overlap with the associated region of A); The associated region centered on node C: {C, B, D, B - C, C - D}; The associated region centered on node D: {D, C, C - D}; Use machine - learning algorithms to conduct risk assessments on each associated region to obtain the assessment scores of each region. When integrating the assessment results, remove the duplicate - calculated associated regions (such as the A - B edge is calculated once in the associated regions of both A and B). The assessment scores of each associated region are weighted - averaged to obtain the risk assessment result of the entire knowledge graph. Based on the integrated assessment results, a neural - network model is constructed as the first - assessment model, which can input new data and output risk assessment scores. It can achieve real - time early warning and monitoring of supply - chain risks.

[0019] The specific content of training an assessment model based on the risk assessment results corresponding to multiple historical knowledge graphs: Definition and extraction of associated regions: The knowledge graph , where is the set of nodes, is the set of edges; Set the associated depth k (such as 1 - order neighbors); The set of associated regions , where is the sub - graph centered on the node ; For the node , its associated region is defined as: , where, represents a specific node in the knowledge graph, which is the central node for defining the associated region. represents the associated region centered on the node center, and this region contains the nodes related to All nodes within a distance not exceeding and the edges between these nodes. represents any node in the knowledge graph, used to compare with the central node to determine whether it belongs to the associated region. : represents the shortest path length between nodes and , used to measure the distance between two nodes. represents the association depth, which is a preset threshold used to determine which nodes belong to the associated region. Only nodes with a shortest path length not exceeding to will be included in . represents the set of all nodes in the knowledge graph. represents the edge connecting nodes and . If and both belong to , then this edge will also be included in . represents the set of all edges in the knowledge graph; the meaning of the formula is: for any node in the knowledge graph, its associated region is a subgraph composed of all nodes with a shortest path length not exceeding to and the edges between these nodes; Independently evaluate the risk of the associated region: Input the associated region , feature vector (including node attributes, edge weights, etc.). Output the risk score . Formula: , where is the risk assessment function (such as a machine learning model), is the feature weight (which can be determined by the entropy weight method or expert weighting); Deduplication processing: For the set of associated regions , the deduplicated set .

[0020] Structural deduplication: Compare subgraph isomorphism (such as using the VF2 algorithm) to eliminate structurally similar associated regions. Node deduplication: Merge duplicate nodes (such as deduplicating by node ID). Formula (example: similarity threshold based on the node set ): For any two associated regions and , calculate ; If , then one of the associated regions is removed, represents the set of associated regions, which is the initial set before the duplicate removal operation and contains multiple associated regions. represents the set of associated regions after duplicate removal, which is obtained from through a series of duplicate removal operations (such as structural duplicate removal, node duplicate removal, etc.). represents the th associated region in the set of associated regions, which is a specific subgraph and contains all the nodes and edges directly associated with a certain central point. represents the th associated region in the set of associated regions, which is also a specific subgraph, similar to but may have different nodes and edges. : represents the set of nodes in the associated region , that is, the set of all nodes contained in . represents the set of nodes in the associated region , that is, the set of all nodes contained in . represents the and between similarity, which is an index used to measure the similarity between two sets. Here, it is used to measure the similarity between the node sets of two associated regions (i.e., two subgraphs). represents similarity threshold, which is a preset similarity threshold used to determine whether two associated regions are similar enough to require the removal of one of them. If , then it is considered that and have too high similarity and one of them needs to be removed to avoid duplicate calculation. The meaning of the entire formula is: for any two associated regions in the set of associated regions and , calculate their Jaccard similarity. If the similarity exceeds the preset similarity threshold , then it is considered that these two associated regions have too high similarity and one of them needs to be removed to obtain the set of associated regions after duplicate removal; Comprehensive evaluation result: the set of risk scores after duplicate removal , output the global risk score , formula (weighted average method): , where is the associated region The weight (which can be determined based on node degree, centrality, etc.), represents the set of deduplicated risk scores, that is, the remaining set of risk scores after removing the risk scores of overlapping calculation or highly similar associated regions. represents the set of deduplicated risk scores the th risk score in it, corresponding to the risk assessment result of a specific associated region (after removing duplicates). represents the set of deduplicated risk scores the total number of risk scores in it, that is, the number of associated regions (after removing duplicates). represents the global risk score, which is an overall risk assessment result obtained by integrating the risk scores of each deduplicated associated region through a certain fusion method (such as weighted average). represents the weight of the th associated region in the set of deduplicated risk scores (corresponding to ). This weight can be determined based on various factors, such as the node degree of the associated region (i.e., the number of connections of nodes within the associated region), centrality (i.e., the importance of the associated region in the overall knowledge graph), etc. The weight is used to adjust the influence degree of each associated region on the global risk score when integrating the risk scores of each associated region. The meaning of the whole formula is: the global risk score is obtained by integrating the risk scores of each deduplicated associated region through the weighted average method. The risk score of each associated region is multiplied by its corresponding weight , and then all the products are added together to obtain the global risk score . This method allows us to adjust the influence of each associated region on the global risk score according to the importance of each associated region, collect the risk assessment results corresponding to multiple historical knowledge graphs, including the risk scores and global risk scores of each associated region. Select a suitable machine learning model (such as neural network, decision tree, etc.) as the evaluation model. Use historical data to train the evaluation model and adjust the model parameters to minimize the prediction error. Use the validation set to evaluate the model performance to ensure the accuracy and generalization ability of the model. An effective evaluation model can be trained based on the risk assessment results corresponding to multiple historical knowledge graphs to achieve the risk assessment of the new knowledge graph; Specifically, the associated area here refers to taking the node as the center point, extracting other nodes related to the center point, and taking the area composed of the other nodes associated with the center point and the node as the associated area corresponding to the center point. Each node is used as the center point once, and the risks of the associated areas corresponding to the nodes are calculated respectively. Finally, the risks of multiple associated areas are comprehensively calculated. During the comprehensive calculation, the repeatedly calculated associated areas need to be excluded, so as to complete the risk assessment of the knowledge graph, which can facilitate the subsequent direct extraction of the risk assessment of the secondary part of the associated area, reduce the complexity of the overall assessment calculation, and thus improve the efficiency of the supply chain risk assessment; Defining all the nodes and edges directly associated with the target node as the associated area of the target node, wherein the steps of the associated area including the primary associated area and the secondary associated area are as follows: respectively setting influence thresholds for the corresponding risk impact factors, and taking the risk impact factors exceeding the influence threshold as the core risk factors; taking the risk impact factors not exceeding the influence threshold as the auxiliary risk factors; taking the nodes corresponding to the core risk factors as the primary nodes, and taking the nodes corresponding to the auxiliary risk factors as the secondary nodes; obtaining the associated areas corresponding to multiple primary nodes and secondary nodes respectively, and obtaining multiple primary associated areas and secondary associated areas before adjustment; Determining the changed nodes to obtain the changed nodes, and the steps of dynamically adjusting the primary associated area and the secondary associated area based on the changed nodes are as follows: collecting the changed nodes based on the sensing points to obtain the active nodes, and recording the time interval between two consecutive changes of each active node; for each active node, judging whether the latest change interval exceeds the preset duration threshold; obtaining the active nodes corresponding to the latest change intervals exceeding the preset duration threshold and marking them as the changed nodes, and taking the data at the last change of the changed nodes as the basic data corresponding to the nodes; taking the associated area corresponding to the changed nodes as the adjusted primary associated area; taking the nodes that have not changed as the non-changed nodes, and taking the associated area corresponding to the non-changed nodes as the adjusted secondary associated area; Specifically, active node: a node with recent changes. Change interval duration: the time difference between two consecutive changes. Preset duration threshold: a time judgment criterion defined by the user (such as 12 days). Changed node: an active node with a change interval exceeding the threshold. Representative data: the data snapshot at the last change of the changed node. By monitoring the change interval and comparing it with the threshold, the nodes that have not been updated for a long time or are abnormally active (i.e., the changed nodes) can be quickly located, avoiding the non-discriminatory processing of all nodes, and improving the analysis efficiency. Taking the data at the last change as the representative data ensures that the latest state of the node is used, avoiding decision-making biases caused by outdated data. The preset duration threshold can be flexibly adjusted according to business requirements (such as in supply chain risk monitoring, the threshold can be set as "supplier delivery cycle + safety buffer period") to adapt to different scenarios; Specifically, nodes correspond to sub-databases, and sensing points correspond to nodes. Therefore, when the basic data collected at the collection end is transmitted to the corresponding sub-database through the communication channel, it is equivalent to being transmitted to the corresponding node. When the collection end collects the changed basic data and transmits it to the corresponding node, at this time, the sensing point will find that new basic data is transmitted in the communication channel, indicating that the node on the knowledge graph has changed. Therefore, the evaluation result will also change, and it is necessary to re-evaluate the risk in the knowledge graph. There are two types of re-risk evaluations. One is to divide the risk impact factors corresponding to the node into core risk factors and auxiliary risk factors. The core risk factors are used as the main part, and the auxiliary risk factors are used as the secondary part. The two parts are separately evaluated according to the first evaluation model and then integrated. The other is based on the first evaluation model. When subsequent nodes change, the main part and the secondary part are adjusted, and the evaluation is carried out according to the adjusted main part and secondary part. Before performing the supply chain risk assessment, the change frequency of the node is judged preferentially. When the change frequency is within the time threshold, the risk assessment model is not started. When it exceeds the time threshold, the assessment is carried out. For example, it is set that when there is no change within a certain time, a re-evaluation is carried out.

[0021] A risk assessment module that obtains a knowledge graph formed by fusing multi-source data; determines an associated area, and performs a risk assessment on the supply chain based on a risk assessment model to obtain a risk assessment result; The risk assessment module includes: A second acquisition unit: acquires multi-source data for supply chain risk assessment from a database and fuses it into a knowledge graph; A risk assessment unit: determines the associated area in the knowledge graph, and performs a risk assessment on the corresponding associated area based on the assessment model to obtain a risk assessment result. Among them, the associated area in the knowledge graph includes the associated area before adjustment and the associated area after adjustment; Specifically, the associated area in the knowledge graph includes the associated area before adjustment and the associated area after adjustment. Among them, the associated area before adjustment refers to the main associated area and the secondary associated area divided according to the initial core risk factors and auxiliary risk factors, belonging to the main associated area before adjustment and the secondary associated area before adjustment. The associated area after adjustment refers to the area with the node corresponding to the changed node that does not exceed the preset condition as the center point, and the center point is used as the core risk factor to adjust the position of the main associated area, belonging to the main associated area after adjustment and the secondary associated area after adjustment; The risk assessment unit includes: A comparison subunit: used to count the number of changed nodes, calculate the proportion of the changed nodes in the knowledge graph based on the number of changed nodes to obtain the proportion of changed nodes; compare the proportion of changed nodes with a preset proportion threshold; Initial evaluation subunit: The number of change nodes corresponding to the proportion of change nodes exceeding the preset proportion threshold is used as the number of abnormal change nodes. Among them, the change nodes corresponding to the number of abnormal change nodes are recorded as abnormal change nodes. For abnormal change nodes, risk assessment is carried out based on the primary associated area before adjustment and the secondary associated area before adjustment; Secondary evaluation subunit: The number of change nodes corresponding to the proportion of change nodes not exceeding the preset proportion threshold is used as the number of normal change nodes. Among them, the change nodes corresponding to the number of normal change nodes are recorded as normal change nodes. For normal change nodes, risk assessment is carried out based on the primary associated area after adjustment and the secondary associated area before adjustment; Specifically, when the proportion of the number of change nodes exceeds the preset proportion range, a global re-evaluation is required. At this time, the primary associated area and the secondary associated area before adjustment are used. And when performing the global re-evaluation, the data after the node change is used, that is, the data collected most recently by the acquisition end is used as the data corresponding to each node to complete the global re-evaluation of the entire knowledge graph; when the proportion of the number of change nodes does not exceed the preset proportion threshold, a separate risk assessment is required. At this time, the primary associated area and the secondary associated area after adjustment are used. For example, if there are 10 nodes corresponding to the overall knowledge graph and the data of one change node has changed and the proportion does not exceed the preset proportion threshold, the primary associated area and the secondary associated area after adjustment are used. At this time, the primary associated area becomes the associated area centered on the change node, and the associated areas corresponding to other nodes belong to the secondary associated areas. Therefore, only the primary associated area needs to be re-evaluated, and the initial evaluation results can be used for the secondary associated areas. By combining the initial evaluation results with the evaluation results of the primary associated area after adjustment, the new risk assessment after the data of the node has changed can be obtained, which can improve the efficiency of supply chain risk assessment; In the present invention, by establishing the binding relationship between the acquisition end and each node in the database, multi-source data related to supply chain risk assessment can be comprehensively and accurately collected, the multi-source data can be integrated, the risk impact factors and the correlation relationships between the factors can be visually presented, the correlation area can be dynamically adjusted according to the changed nodes, the dynamic changes in the supply chain can be responded to in real time, the assessment strategy can be adjusted in a timely manner, and the accuracy and timeliness of the assessment results are improved; in the process of defining the correlation area, impact thresholds are respectively set for the corresponding risk impact factors, the risk impact factors exceeding the impact threshold are used as core risk factors, and those not exceeding are used as auxiliary risk factors, highlighting the key risks, making the risk assessment more targeted, and improving the assessment efficiency and accuracy; the assessment is respectively carried out based on the correlation area before and after adjustment to ensure the comprehensiveness and accuracy of the assessment results; various data sources such as internal data, external data, supplier data, and logistics data of the supply chain can be integrated, and through data fusion, the supply chain risk can be comprehensively and accurately assessed.

[0022] In the description of the present specification, the description with reference to terms such as "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In the present specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0023] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A supply chain risk assessment system based on multi-source data fusion, characterized in that, Including: A data acquisition module, which configures multi-source data acquisition terminals, establishes the binding relationship between the acquisition terminals and each node in the database. The database is used to store multi-source data related to supply chain risk assessment. Classify the multi-source data and store it in a database composed of multiple sub-databases according to categories. Bind the sub-databases to the acquisition terminals of the corresponding category of basic data. A data fusion module, which obtains the classified basic data and the corresponding sub-databases from the database. Extract the characteristics of the basic data as risk impact factors, obtain the association relationships as the association relationships between factors, set the nodes and edges of the knowledge graph, and construct the knowledge graph based on the nodes and edges. A model construction module, which takes the nodes of the knowledge graph as the center, defines the association area including the main association area and the secondary association area. Set the virtual boundary layer, establish the communication channel between the acquisition terminal and the node, and establish the driving relationship between the sensing point and the risk assessment model. Determine the changed nodes and dynamically adjust the association area. Evaluate the association area and train the evaluation model based on the comprehensive results. A risk assessment module, which obtains the knowledge graph formed by the fusion of multi-source data. Determine the association area and conduct a risk assessment on the supply chain based on the risk assessment model to obtain the risk assessment result.

2. The supply chain risk assessment system based on multi-source data fusion according to claim 1, wherein: The data acquisition module includes: An acquisition unit: Arrange acquisition points at the data sources of multi-source data, register the acquisition terminals corresponding to the acquisition points, and respectively acquire the basic data for supply chain risk assessment based on the acquisition terminals to obtain the multi-source data for supply chain risk assessment. A storage unit: Classify the multi-source data to obtain multiple categories of basic data, and store the multiple categories of basic data in the database according to categories. The database includes multiple sub-databases, and each sub-database corresponds to a category of basic data. An establishment unit: Bind between the sub-database and the acquisition terminal of the basic data of the corresponding category.

3. The supply chain risk assessment system based on multi-source data fusion according to claim 1, wherein: The data fusion module includes: A first acquisition unit: Obtain and extract the classified basic data and the corresponding sub-databases from the database. A data fusion unit, which extracts the data characteristics of a category of basic data and the corresponding sub-database as a risk impact factor, and obtains multiple risk impact factors based on multiple categories of basic data. Obtain the association relationships between each category of basic data, and use the association relationships between each category of basic data as the association relationships between factors. A graph construction unit, which sets multiple nodes of the knowledge graph corresponding to multiple risk impact factors, sets multiple edges of the knowledge graph corresponding to the association relationships between factors, and constructs the knowledge graph based on the multiple nodes and multiple edges.

4. The supply chain risk assessment system based on multi-source data fusion according to claim 1, characterized in that: The model construction module includes: An area determination unit: Take each node in the knowledge graph as the center point, and take the node as the center point as the target node. Define all the nodes and edges directly associated with the target node as the association area of the target node, where the association area includes the main association area and the secondary association area. A boundary construction unit: Set a virtual boundary layer with sensing points in the knowledge graph, where the sensing points correspond to the nodes one by one. Establish the communication channel between the acquisition terminal and the corresponding node, where each communication channel corresponds to a sensing point. Establish the driving relationship between the sensing point and the risk assessment model. Region adjustment unit: Determine the changed nodes to obtain the changed nodes, and dynamically adjust the associated regions based on the changed nodes; Model construction unit: Conduct separate risk assessments on each associated region, identify and eliminate the associated regions with duplicate calculations, integrate the evaluation results of each associated region to obtain the risk assessment result corresponding to the knowledge graph, and train the evaluation model based on the risk assessment results corresponding to multiple historical knowledge graphs.

5. The supply chain risk assessment system based on multi-source data fusion according to claim 4, characterized in that: The step of defining all the nodes and edges directly associated with the target node as the associated region of the target node, where the associated region includes the primary associated region and the secondary associated region includes: Set the influence thresholds for the corresponding risk impact factors respectively; Take the risk impact factors exceeding the influence threshold as the core risk factors; Take the risk impact factors not exceeding the influence threshold as the auxiliary risk factors; Take the nodes corresponding to the core risk factors as the primary nodes, and take the nodes corresponding to the auxiliary risk factors as the secondary nodes; Obtain the associated regions corresponding to multiple primary nodes and secondary nodes respectively to obtain multiple pre-adjustment primary associated regions and pre-adjustment secondary associated regions.

6. The supply chain risk assessment system based on multi-source data fusion according to claim 4, characterized in that: The step of determining the changed nodes to obtain the changed nodes and dynamically adjusting the primary associated region and the secondary associated region based on the changed nodes includes: Collect the changed nodes based on the sensing points to obtain the active nodes, and record the time interval between two consecutive changes of each active node; For each active node, judge whether the latest change interval exceeds the preset duration threshold; Obtain the active nodes corresponding to the latest change interval exceeding the preset duration threshold and mark them as the changed nodes, and take the data at the last change as the basic data corresponding to the node; Take the associated region corresponding to the changed node as the adjusted primary associated region; Take the nodes that have not changed as the non-changed nodes, and take the associated region corresponding to the non-changed nodes as the adjusted secondary associated region.

7. The supply chain risk assessment system based on multi-source data fusion according to claim 1, wherein: The risk assessment module includes: Second acquisition unit: Obtain multi-source data for supply chain risk assessment from the database and integrate them into a knowledge graph; Risk assessment unit: Determine the associated regions in the knowledge graph, and conduct risk assessment on the associated regions based on the evaluation model to obtain the risk assessment result, where the associated regions in the knowledge graph include the pre-adjustment associated regions and the post-adjustment associated regions.

8. The supply chain risk assessment system based on multi-source data fusion according to claim 7, characterized in that: The risk assessment unit includes: Comparison subunit: Used to count the number of changed nodes, calculate the proportion of the changed nodes in the knowledge graph based on the number of changed nodes to obtain the changed node ratio; Compare the changed node ratio with the preset ratio threshold; Initial evaluation subunit: Take the number of changed nodes corresponding to the changed node ratio exceeding the preset ratio threshold as the number of abnormal changed nodes, where the changed nodes corresponding to the number of abnormal changed nodes are recorded as the abnormal changed nodes; For the abnormal changed nodes, conduct risk assessment based on the pre-adjustment primary associated region and the pre-adjustment secondary associated region. Secondary evaluation sub-unit: The number of change nodes corresponding to the proportion of change nodes not exceeding the preset ratio threshold is used as the number of normal change nodes. Among them, the change nodes corresponding to the number of normal change nodes are recorded as normal change nodes; for the normal change nodes, risk assessment is carried out based on the adjusted main associated area and the secondary associated area before adjustment.

Images