Relationship-graph-based decision processing method, and device

Abstract

The present application can be used in the technical field of artificial intelligence. Provided are a relationship-graph-based decision processing method and a device. The method comprises: determining initial decision results respectively corresponding to multiple decision objects; determining association information between the multiple decision objects, the association information comprising the similarity between any two decision objects in terms of a decision factor; constructing a first relationship graph on the basis of the association information between the multiple decision objects, each node in the first relationship graph corresponding to one decision object, and when there is a first connection edge between the two decision objects, the edge weight of the first connection edge being positively correlated with the association information; and by means of the first relationship graph, updating the initial decision results respectively corresponding to the multiple decision objects, so as to obtain a target decision result of each decision object. In the present application, on the basis of the initial decision results, the association information between the decision objects can be fully utilized, and a dynamic damping coefficient can be constructed for decision processing, so that the decision accuracy can be effectively improved.

Description

Decision-making processing method and device based on relationship graph

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on December 28, 2023, with application number 202311845743.4 and application name “Decision-making processing method and device based on relationship graph”, all contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of artificial intelligence technology, and in particular to a decision-making processing method and device based on a relationship graph. Background Art

[0003] In the field of artificial intelligence, intelligent decision-making can be performed based on one or more factors to produce a decision outcome. In this decision-making scenario, the factors needed for decision-making are various factors that influence the decision outcome, which can be referred to as decision factors. For example, decision-making can be performed based on various factors of a decision-making object, such as a merchant or bank, to determine the status of the decision-making object.

[0004] In the prior art, the main process of decision processing may include: first, for a decision object, obtaining one or more decision factors of the decision object; then, inputting all decision factors of the decision object into the decision algorithm to obtain the decision result.

[0005] However, the above scheme has the problem of low decision-making accuracy. Summary of the Invention

[0006] The present application provides a decision-making processing method and device based on a relationship graph, which can improve decision-making accuracy.

[0007] In a first aspect, the present application provides a decision-making method based on a relationship graph, the method comprising:

[0008] Determine the initial decision results corresponding to multiple decision objects;

[0009] Determining association information between the multiple decision objects, the association information including similarity between any two decision objects in decision factors;

[0010] Constructing a first relationship graph based on the association information between the multiple decision objects, wherein each node in the first relationship graph corresponds to one of the decision objects, and when there is a first connecting edge between two of the decision objects, the edge weight of the first connecting edge is positively correlated with the association information;

[0011] The initial decision results respectively corresponding to the multiple decision objects are updated through the first relationship graph to obtain a target decision result for each decision object.

[0012] In a second aspect, the present application provides a decision processing device based on a relationship graph, comprising:

[0013] An initial decision module, used to determine initial decision results corresponding to multiple decision objects;

[0014] an association determination module, configured to determine association information between the plurality of decision objects, wherein the association information includes similarity between any two decision objects in terms of decision factors;

[0015] A first graph construction module is configured to construct a first relationship graph based on the association information between the multiple decision objects, wherein each node in the first relationship graph corresponds to one of the decision objects, and when there is a first connecting edge between two of the decision objects, the edge weight of the first connecting edge is positively correlated with the association information;

[0016] The target decision module is used to update the initial decision results corresponding to the multiple decision objects respectively through the first relationship map to obtain a target decision result for each decision object.

[0017] In a third aspect, the present application provides an electronic device, comprising a memory and at least one processor;

[0018] wherein the memory stores computer-executable instructions;

[0019] At least one processor executes the computer-executable instructions stored in the memory, so that the electronic device implements the method of the first aspect described above.

[0020] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement the method described in the first aspect.

[0021] In a fifth aspect, the present application provides a computer program product for implementing the method of the first aspect.

[0022] In combination with the above technical solutions, the decision processing method and device based on the relationship graph provided by the present application can determine the initial decision results corresponding to multiple decision objects respectively; determine the association information between multiple decision objects, including the similarity between any two decision objects in the decision factors; construct a first relationship graph based on the association information between multiple decision objects, each node in the first relationship graph corresponds to a decision object, and when there is a first connecting edge between two decision objects, the edge weight of the first connecting edge is positively correlated with the association information; update the initial decision results corresponding to multiple decision objects respectively through the first relationship graph to obtain the target decision result for each decision object. Based on the initial decision results, the present application can make full use of the association information between decision objects and construct a dynamic damping coefficient for decision processing, which can effectively improve the accuracy of decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] FIG1 is a schematic diagram of a decision-making process provided by the prior art;

[0024] FIG2 is a flowchart of a decision-making method based on a relationship graph provided in an embodiment of the present application;

[0025] FIG3 is a schematic structural diagram of a second relationship map provided in an embodiment of the present application;

[0026] FIG4 is a schematic structural diagram of a first relationship map provided in an embodiment of the present application;

[0027] FIG5 is a flowchart of another decision-making method based on a relationship graph provided in an embodiment of the present application;

[0028] FIG6 is a structural block diagram of a decision processing device based on a relationship graph provided in an embodiment of the present application;

[0029] FIG7 is a structural block diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0030] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0031] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0032] It should be noted that the decision-making processing method and device based on the relationship graph of the present application can be used in the field of artificial intelligence technology, and can also be used in any field other than the field of artificial intelligence technology. The application field of the decision-making processing method and device based on the relationship graph of the present application is not limited.

[0033] The decision-making process of the embodiments of this application has a variety of application scenarios, including but not limited to: financial technology application scenarios and engineering application scenarios. For example, in financial technology application scenarios, it is possible to conduct capability assessment, value assessment, demand assessment, supplier selection, strategic decision-making, and manufacturing system decision-making on financial entities to obtain decision results.

[0034] It should be noted that the decision-making object in the embodiments of the present application can be any object in any scenario in any field, including physical objects and virtual objects. For example, it can be a physical object such as a merchant, a bank, or a manufacturing system, or it can be a virtual object such as a deployment strategy or a manufacturing method. When the decision-making object or the decision-making scenario is different, the decision result will be different. For example, the decision result can be the capability assessment result, operating status, etc. of the physical object, or the result of whether the deployment strategy or manufacturing method is executed.

[0035] FIG1 is a schematic diagram of a decision-making process provided by the prior art. Referring to FIG1 , for the decision-making process of decision object D1, the decision factors M11, M12, and M13 of decision object D1 can be input into the decision algorithm to obtain the decision result of decision object D1. If the same decision-making process needs to be performed on another decision object D2, then the decision factors M21, M22, and M23 of decision pair D2 need to be input into the decision algorithm to obtain the decision result of decision object D2. For the same decision-making process, the decision factors required are the same, so the decision factors M11 and M21 are the same decision factors, but the values ​​may be different. Similarly, M12 and M22 are the same decision factors, but the values ​​are different, and M13 and M23 are the same decision factors, but the values ​​are different.

[0036] However, when making a decision for a decision object, the above decision processing process only considers the decision factors of one decision object, which will result in low decision accuracy.

[0037] To address the aforementioned technical issues, embodiments of the present application can first determine an initial decision result for each decision object, and then combine this initial decision result with the association information between multiple decision objects to determine a target decision result for each decision object. This allows decision processing to be fully utilized based on the initial decision result, effectively improving decision accuracy.

[0038] For example, in a risk analysis scenario, through the above-mentioned decision-making processing scheme of multiple decision factors, it is determined that the risk of decision object D1 is greater, while the risk of decision object D2, which is strongly associated with decision object D1, is smaller. In this scenario, the decision accuracy of decision object D1 or decision object D2 may be wrong. However, the embodiment of the present application combines the association information of the decision object and the initial decision result for decision processing, and can identify this strong association relationship between decision object D1 and decision object D2. Therefore, the same target decision result will be obtained for decision object D1 and decision object D2, which can ensure the consistency of the decision results of the two and help improve the decision accuracy.

[0039] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0040] FIG2 is a flowchart of a decision-making method based on a relationship graph provided by an embodiment of the present application. Referring to FIG2 , the decision-making method based on a relationship graph of the present application may include:

[0041] S201: Determine initial decision results corresponding to multiple decision objects.

[0042] Each decision object corresponds to an initial decision result, which can be a decision result obtained by any decision algorithm. For example, it can be obtained by making a decision based on multiple decision factors, or by making a decision based on a set of decision factors.

[0043] The above decision factors are related to specific decision scenarios. For example, for operating entities, decision factors may include but are not limited to: marketing dependence, proportion of online operations, proportion of card-based operations, proportion of bank and merchant operations, proportion of non-financial operations, etc., to determine the status of the operating entity.

[0044] In some embodiments, the initial decision result can be obtained through the following process: first, obtaining the decision weight corresponding to at least one decision factor, where the decision factor is used to perform decision processing on multiple decision objects, and the decision weight of a decision factor is used to indicate the degree of influence of the decision factor on the decision processing; then, based on the decision weight of the decision factor and the decision factor corresponding to the decision object, a decision processing is performed to obtain the initial decision result of the decision object. In this way, the influence of the decision factor can be flexibly adjusted in combination with the decision weight, thereby achieving the purpose of flexibly adjusting the decision result and improving the flexibility of decision making.

[0045] The decision weights of the above decision factors are positively correlated with the degree of influence of the decision factors on the decision process. The greater the decision weight of a decision factor, the greater its influence on the decision process. Conversely, the smaller the decision weight of the decision factor, the smaller its influence on the decision process.

[0046] It is understandable that the above initial decision results are positively correlated with the above decision weights and decision factors. In one embodiment, the above initial decision results can be calculated using the following formula (1):

[0047] Where DR(init) is the initial decision result of a decision object, n is the number of decision factors required for the decision process, Mi is the i-th decision factor of the decision object, or the result of the i-th decision factor normalized, and Wi is the decision weight corresponding to the i-th decision factor Mi of the decision object. i is a positive integer greater than or equal to 1 and less than or equal to n.

[0048] It can be understood that the unitization processing of the i-th decision factor can be performed by dividing the i-th decision factor by the sum of all decision factors of the decision object.

[0049] The decision weights here can be flexibly set manually according to the application scenario and requirements, or can be determined through S2011 to S2013.

[0050] S2011. Construct a comparison matrix, where the element in the i-th row and j-th column of the comparison matrix is ​​used to represent the relative importance of the i-th decision factor compared to the j-th decision factor, and i and j are both integers greater than or equal to 1.

[0051] S2012. Determine a priority matrix based on the comparison matrix, where the priority matrix includes the priorities corresponding to the respective decision factors in the comparison matrix.

[0052] S2013: Verify the priority matrix, and when the verification passes, use the priorities included in the priority matrix as decision weights of the decision factors.

[0053] The comparison matrix may be an n×n matrix, i and j are integers greater than or equal to 1 and less than or equal to n, and n is the number of decision factors.

[0054] The comparison matrix is ​​determined by the relative importance of multiple decision factors. To construct the comparison matrix, the set of decision factors can be denoted as MS = {M1, M2, …Mn}, where M1, M2, …Mn are the first, second, through nth decision factors, respectively. The relative importance of any i-th decision factor compared to any j-th decision factor can be determined through expert scoring or automated determination.

[0055] The greater the relative importance of any i-th decision factor compared to any j-th decision factor, the greater the importance of the i-th decision factor is, and the greater the impact of the i-th decision factor on the decision result is. Conversely, the smaller the relative importance of any i-th decision factor compared to any j-th decision factor, the smaller the importance of the i-th decision factor is, and the smaller the impact of the i-th decision factor on the decision result is.

[0056] For example, when it is determined based on expert experience that the i-th decision factor and the j-th decision factor are equally important, the relative importance of the i-th decision factor compared to the j-th decision factor can be set to 1. When it is determined based on expert experience that the i-th decision factor is more important than the j-th decision factor, the relative importance of the i-th decision factor compared to the j-th decision factor can be set to be greater than 1. When it is determined based on expert experience that the j-th decision factor is more important than the i-th factor, the relative importance of the i-th decision factor compared to the j-th decision factor can be set to be less than 1.

[0057] The contrast matrix A constructed according to the relative importance of multiple decision factors can be as follows:

[0058] In the comparison matrix A, Aii=1, which means that the relative importance of the i-th decision factor compared to the i-th decision factor is 1. Aij=1 / Aji, which means that the relative importance of the i-th decision factor compared to the j-th decision factor is the reciprocal of the relative importance of the j-th decision factor compared to the i-th decision factor.

[0059] After obtaining the comparison matrix A, a priority matrix P can be determined based on the comparison matrix A. The priority matrix P includes the priorities corresponding to the decision factors in the comparison matrix A. The priority corresponding to each decision factor can be determined in two ways.

[0060] In the first priority determination method of the above-mentioned decision factors, the priority of the decision factor corresponding to each row can be determined according to the relative importance of each row in the comparison matrix. For example, for the i-th decision factor, the relative importance Ai1 to Ain in the i-th row in the comparison matrix A can be added, multiplied or averaged to obtain the priority of the i-th decision factor.

[0061] In the second priority determination method of the above-mentioned decision factors, first, each column of the comparison matrix is ​​normalized to obtain a unit comparison matrix; then, the average value of each row of the unit comparison matrix is ​​used as the priority of each decision factor in the priority matrix.

[0062] The unitization process can be understood as a normalization process, which is used to process the relative importance of each column in the comparison matrix A to a value greater than or equal to 0 and less than or equal to 1. The specific algorithm can be as follows:

[0063] Wherein, A'ij is the element in the i-th row and j-th column of the unit contrast matrix A', and both the unit contrast matrix A' and the contrast matrix A are n×n matrices.

[0064] Based on the above unit comparison matrix A', the priority of each decision factor can be obtained. Specifically, for the i-th decision factor, its priority can be calculated by the following formula:

[0065] Where Pi is the priority of the i-th decision factor Mi. The larger the priority, the greater the impact of the i-th decision factor Mi on the decision process.

[0066] As can be seen, the first method of determining the priorities of decision factors mentioned above may result in a large priority value, which is not conducive to subsequent calculations during the decision-making process. However, the second method of determining the priorities of decision factors mentioned above, which first normalizes the columns and then determines the priority based on the average value, can ensure that the priority value is within a controllable range as much as possible, helping to reduce the complexity of subsequent calculations.

[0067] After obtaining the priority matrix, the priority matrix can be verified. If the verification is successful, the priorities in the priority matrix are used as decision weights. If the priority matrix verification fails, the relative importance of the decision factors needs to be re-determined to reconstruct the comparison matrix. The priority matrix is ​​then determined through S2011 to S2013 until the constructed priority matrix is ​​successfully verified.

[0068] In some implementations, the comparison matrix and the priority matrix may be combined to first determine the maximum eigenvalue of the comparison matrix. A verification index may then be determined based on the maximum eigenvalue and the number of decision factors in the comparison matrix. Finally, a determination may be made as to whether the verification index satisfies a preset condition. If the verification index satisfies the preset condition, the priority matrix verification is determined to have passed. Otherwise, the priority matrix verification is determined to have failed.

[0069] Specifically, the maximum eigenvalue λmax can be calculated by the following formula:

[0070] Where AP is the product of the n×n contrast matrix A and the n×1 priority matrix P. AP is an n×1 matrix, [AP]i is the i-th element of the matrix AP, and Pi is the i-th element of the priority matrix P. A larger maximum eigenvalue indicates a greater likelihood of inconsistency in the contrast matrix A. In other words, the contrast matrix A is not properly configured, and there are conflicts in relative importance.

[0071] For example, if A12 is greater than 1 and A21 is less than 1, it means that the first decision factor has a greater impact on the decision process than the second decision factor. At the same time, if A13 is less than 1 and A31 is greater than 1, it means that the third decision factor has a greater impact on the decision process than the first decision factor. Therefore, it can be inferred that the third decision factor has a greater impact on the decision process than the second decision factor. However, if A23 is greater than 1 and A32 is less than 1, it means that the second decision factor has a greater impact on the decision process than the third decision factor. In this way, a conflict occurs. In this case, the corresponding maximum eigenvalue is larger.

[0072] After obtaining the above maximum eigenvalue, the test index can be determined according to the maximum eigenvalue and the number of decision factors. The specific formula is as follows:

[0073] Among them, CR is the test index, which can be understood as the consistency ratio. CI is the consistency index, and RI is the random index, which can be obtained by looking up the table based on the dimension n of the comparison matrix.

[0074] After obtaining the above-mentioned test index CR, if the test index is less than or equal to a preset threshold, it is determined that the priority matrix verification has passed. If the test index is greater than the preset threshold, it is determined that the priority matrix verification has failed. The preset threshold can be set based on experience, for example, it can be 0.1.

[0075] S202: Determine association information between multiple decision objects, where the association information includes similarity between any two decision objects in decision factors.

[0076] It can be understood that the association information is used to indicate the association relationship between any two decision objects. In addition to the similarity between any two decision objects in the decision factors mentioned above, it can also include the first degree of association between any two decision objects. The first degree of association is used to indicate the degree of dependence between the two decision objects.

[0077] Among them, the above similarity can be determined by the following steps: first, all decision factors of each decision object are constructed into a vector to serve as the decision factor vector of the decision object; then, the similarity between the decision factor vectors of the two decision objects is used as the similarity of the decision factors of the above two decision objects.

[0078] The similarity between vectors can be calculated using a variety of similarity algorithms, such as cosine similarity, paradigm distance similarity, etc. Taking cosine similarity as an example, the similarity of the above decision objects in the decision factors can be calculated using the following formula:

[0079] Where αbk is the similarity between the bth and kth decision objects in terms of decision factors. cossimilarity(Vb,Vk) is used to calculate the similarity between Vb and Vk. Vb and Vk are the decision factor vectors corresponding to the bth and kth decision objects, respectively. Vbi and Vki are the values ​​of the i-th dimension in the decision factor vectors corresponding to the bth and kth decision objects, respectively.

[0080] In addition, the degree of dependence between the above two decision objects can be identified based on the interaction behavior between the decision objects, or it can be identified based on the degree of dependence between the decision object and a third-party object. The degree of dependence between the decision object and the third-party object can be determined by the interaction behavior.

[0081] In some embodiments, identifying the degree of dependence between decision objects through the degree of dependence between the decision objects and third-party objects can be achieved through the following process: first, constructing a second relationship graph through multiple decision objects and at least one third-party object, each node in the second relationship graph corresponds to a decision object or a third-party object, and when there is a second connection edge between a decision object and a third-party object, the edge weight of the second connection edge is used to indicate the second degree of association between the decision object and the third-party object; then, for two decision objects connected to at least one same third-party object in the second relationship graph, determining the first degree of association between the two decision objects based on the second degree of association between the two decision objects and the same third-party object.

[0082] The second relationship graph includes all decision-making objects and all third-party objects that interact with these decision-making objects. Each node in the second relationship graph corresponds to a decision-making object or a third-party object. When a decision-making object and a third-party object interact, the decision-making object and the third-party object are connected by an edge, which is called a second connection edge. For example, the decision-making object can be a bank branch, and the third-party object can be a merchant. The second connection edge can be used to indicate the existence of a transaction between the bank branch and the merchant. The weight of the second connection edge can be the transaction amount or transaction frequency, etc. Therefore, a second relationship graph can be constructed based on multiple bank branches and multiple merchants and the transaction behaviors between them to perform decision-making processing on the bank branch and determine its operating status. Of course, decision-making processing can also be performed on merchants. In this case, the merchant serves as the decision-making object and the bank serves as the third-party object.

[0083] The second degree of association corresponding to the second connection edge in the embodiment of the present application is used to represent the degree of association between the decision object connected by the second connection edge and the third-party object. The second degree of association can correspond to different measurement indicators in different scenarios. For example, in a transaction scenario, the second degree of association can be transaction amount, transaction frequency, etc.

[0084] Based on the second relationship graph, two decision objects connected to at least one identical third-party object can be found. This can be referred to as a group of associated decision objects. Multiple groups of such associated decision objects exist in the second relationship graph. The first degree of association between two decision objects is positively correlated with their corresponding second degree of association. In one embodiment, this can be calculated using the following formula:

[0085] Where Ybk is the first degree of association between the bth decision object and the kth decision object. R is the number of third-party objects to which the bth decision object and the kth decision object are jointly connected, and r is an integer greater than or equal to 1 and less than or equal to R. Zbr is the second degree of association between the bth decision object and the rth third-party object, and Zkr is the second degree of association between the kth decision object and the rth third-party object.

[0086] Figure 3 is a schematic diagram of the structure of a second relationship map provided in an embodiment of the present application. As shown in Figure 3, the second relationship map includes three decision objects D1, D2, and D3, and two third-party objects S1 and S2. It will be appreciated that Figure 3 illustrates a relatively small number of decision objects and third-party objects for illustrative purposes. However, in actual applications, the number of decision objects and third-party objects in the second relationship map may exceed the number shown in Figure 3.

[0087] As shown in Figure 3, the second degree of association between the first decision object D1 and the first third-party object S1 is Z11, the second degree of association between the first decision object D1 and the second third-party object S2 is Z12, the second degree of association between the second decision object D2 and the first third-party object S1 is Z21, the second degree of association between the second decision object D2 and the second third-party object S2 is Z22, the second degree of association between the third decision object D3 and the first third-party object S1 is Z31, and the second degree of association between the third decision object D3 and the second third-party object S2 is Z32.

[0088] From the second relationship graph shown in FIG3 , it can be seen that the first decision object D1 and the second decision object D2 are not only connected to the third-party object S1, but also to the third-party object S2. Therefore, according to the above formula (8), the first degree of association between the first decision object D1 and the second decision object D2 can be calculated as Y12 = Z11 × Z21 + Z12 × Z22. Similarly, the first degree of association between the first decision object D1 and the third decision object D3 can be calculated as Y13 = Z11 × Z31 + Z12 × Z32, and the first degree of association between the second decision object D2 and the third decision object D3 can be calculated as Y23 = Z21 × Z31 + Z22 × Z32.

[0089] It can be understood that the construction process of the above-mentioned second relationship graph may include the following steps: first, determine all decision objects and all third-party objects that have interactive behaviors with these decision objects, and then determine whether there is interactive behavior between any decision object and the third-party object, and when there is interactive behavior, determine the corresponding second degree of association based on the interactive behavior between the decision object and the third-party object; then, each decision object and each third-party object is regarded as a node in the second relationship graph; finally, the decision object and the third-party object with interactive behavior are connected through a second connecting edge, and the second degree of association between the decision object and the third-party object is used as the edge weight of the second connecting edge.

[0090] After constructing the second relationship graph, all associated decision objects can be found therein to calculate the first degree of association between the associated decision objects using formula (8).

[0091] It can be seen that the embodiment of the present application can vividly and accurately represent the relationship between the decision object and the third-party object through the second relationship map, and thus can accurately determine the first degree of association based on the second degree of association.

[0092] After obtaining the above-mentioned association information, the target decision result of each decision object can be determined according to the association information between the multiple decision objects and the initial decision results corresponding to the multiple decision objects.

[0093] The target decision result can be understood as updating the initial decision result using the associated information. When the target decision result and the initial decision result are expressed numerically, the target decision result is positively correlated not only with the initial decision result but also with the degree of association represented by the associated information.

[0094] In some embodiments, the association information can be used to indicate whether there is an association between decision objects. The initial decision result of a decision object (referred to as the first decision object) can be used to update the initial decision result of a decision object (referred to as the second decision object) associated with the first decision object to serve as the target decision result for the second decision object. For example, if the initial decision result of the first decision object is larger, the initial decision result of the second decision object can be increased.

[0095] In other embodiments, the association information between the decision objects may further include the degree of association between any two decision objects, namely, the aforementioned first degree of association, so that the initial decision result can be accurately adjusted based on the first degree of association. The target decision result is positively correlated with the first degree of association and the initial decision result, respectively. For example, the target decision result for the second decision object may be the product of the first degree of association between the first and second decision objects and the initial decision result for the second decision object.

[0096] In some embodiments, the process of determining the target decision result is a process of updating the initial decision result in combination with the first relationship graph, specifically including S203 and S204.

[0097] S203. Construct a first relationship graph based on the association information between multiple decision objects, where each node in the first relationship graph corresponds to a decision object. When there is a first connecting edge between two decision objects, the edge weight of the first connecting edge is positively correlated with the association information.

[0098] The first relationship graph can be obtained by removing the third-party object from the second relationship graph, or can be constructed directly from the decision object. The edge weight of the first connecting edge can also be understood as the third degree of association between the two decision objects, where the third degree of association is positively correlated with the association information.

[0099] When deleting third-party objects from the second relationship graph to obtain the first relationship graph, each time a third-party object is deleted, the decision objects connected to the third-party object are connected as the first connection edge in the first relationship graph.

[0100] When the first relationship graph is directly constructed through multiple decision objects, each decision object can be regarded as a node, and the first connection edges between these decision objects can be constructed according to the association information.

[0101] It can be seen that the above-mentioned first relationship map is used to represent the association relationship and the third degree of association between all decision objects. The first relationship map includes all decision objects, and each decision object is a node. Among them, the third degree of association between two decision objects in the first relationship map can be positively correlated with the first degree of association between the two decision objects, and positively correlated with the similarity between the two decision objects in the decision factors. For example, the third degree of association between two decision objects in the first relationship map can be the result of multiplying the first degree of association between the two decision objects by the similarity between the two decision objects in the decision factors.

[0102] FIG4 is a schematic diagram of the structure of a first relationship graph provided in an embodiment of the present application. The first relationship graph shown in FIG4 is obtained after deleting the third-party decision object from the second relationship graph shown in FIG3. Referring to FIG4, the first relationship graph includes three decision objects D1, D2, and D3, wherein the third degree of association YA12 between decision objects D1 and D2 = the first degree of association Y12 between decision objects D1 and D2 × the similarity α12 between decision objects D1 and D2 in terms of decision factors; the third degree of association YA13 between decision objects D1 and D3 = the first degree of association Y13 between decision objects D1 and D3 × the similarity α13 between decision objects D1 and D3 in terms of decision factors; and the third degree of association YA23 between decision objects D2 and D3 = the first degree of association Y23 between decision objects D2 and D3 × the similarity α23 between decision objects D2 and D3 in terms of decision factors.

[0103] S204: Update the initial decision results corresponding to the multiple decision objects respectively through the first relationship graph to obtain the target decision result of each decision object.

[0104] In an embodiment of the present application, the initial decision results corresponding to multiple decision objects are updated through the first relationship graph to obtain the target decision result of each decision object, that is, decision processing is achieved based on the first relationship graph through the influence propagation algorithm.

[0105] The embodiment of the present application can accurately determine the association relationship between decision objects through the first relationship map, so as to accurately update the initial decision objects corresponding to each decision object in the first relationship map, and obtain the target decision result of each decision object.

[0106] In addition, the weight of the first connecting edge in the above-mentioned first relationship graph of the embodiment of the present application combines similarity and first association strength to describe the association relationship between decision objects from these two aspects, which can maximize the accuracy of the association between decision objects. Compared with the traditional influence propagation algorithm, the greater the mutual influence of adjacent decision objects, the influence of the present application not only takes into account the adjacency relationship of decision objects, but also takes into account the similarity of decision objects in decision factors. This helps to further improve decision accuracy.

[0107] In some embodiments, S204 may include the following steps: first, for each decision object, the preset damping coefficient is updated to obtain a dynamic damping coefficient based on the average similarity between the decision object and its adjacent decision objects in the decision factors; then, the initial decision results corresponding to the multiple decision objects in the first relationship graph are updated using the dynamic damping coefficient to obtain the target decision result of each decision object. The dynamic damping coefficient is used to control the difficulty of decision propagation between the decision object and the adjacent decision objects.

[0108] The adjacent decision objects are decision objects that are directly connected to the decision object. For example, for the decision object D1 shown in FIG4 , its adjacent decision objects include decision objects D2 and D3.

[0109] The average similarity here is the average of the similarities between the decision object and all adjacent decision objects in terms of decision factors. The similarity between the decision object and the adjacent decision objects in terms of decision factors can be calculated using the aforementioned formula (7).

[0110] The dynamic damping coefficient is positively correlated with the aforementioned average similarity and with the preset damping coefficient. That is, as the preset damping coefficient increases and / or the average similarity increases, the dynamic damping coefficient increases. The preset damping coefficient is a preset fixed value. Compared to the prior art, which uses this preset damping coefficient to control the difficulty of propagation between nodes, the present application uses the dynamic damping coefficient to control the difficulty of decision propagation between decision objects.

[0111] The embodiment of the present application can flexibly control the difficulty of propagation of decision results between different decision objects through the above-mentioned dynamic damping coefficient. When the average similarity between the decision object and each adjacent decision object in terms of decision factors is smaller, the similarity between the representative decision object and the adjacent decision object is lower. At this time, the smaller the dynamic damping coefficient is, the greater the difficulty of propagation of the decision result between the decision object and the adjacent decision object is, that is, the update amplitude of the updated decision result is smaller. Conversely, when the average similarity between the decision object and each adjacent decision object in terms of decision factors is greater, the similarity between the representative decision object and the adjacent decision object is higher. At this time, the larger the dynamic damping coefficient is, the greater the difficulty of propagation of the decision result between the decision object and the adjacent decision object is, that is, the update amplitude of the updated decision result is larger.

[0112] In some embodiments, the process of updating the initial decision result based on the above-mentioned first relationship graph to obtain the target decision result may specifically include multiple rounds of iterations. In each round of iteration, for each decision object, the updated decision result of the decision object in the iteration is determined based on the dynamic damping coefficient and the target information of the decision object in the first relationship graph. The target information includes: the edge weight of the first connecting edge between the decision object and the adjacent decision object of the decision object in the first relationship graph, and the updated decision result of the adjacent decision object in the previous round of iteration. After each round of iteration, if the updated decision results corresponding to multiple decision objects respectively meet the preset convergence conditions, the updated decision results corresponding to the multiple decision objects respectively are used as the corresponding target decision results.

[0113] It can be understood that before the first iteration, the initial decision result of each decision object is used as the updated decision result of that decision object to start the first iteration. Therefore, in the first iteration, the updated decision result of the adjacent decision object in the previous iteration is the initial decision result of that adjacent decision object. In the second and subsequent C-th iterations, the updated decision result of the adjacent decision object in the previous iteration is the updated decision result obtained by the adjacent decision object after the C-1th iteration.

[0114] The updated decision results of the decision object in each iteration are positively correlated with both items in the target information, and can be flexibly adjusted based on this positive correlation. In other words, as the edge weight of the first connecting edge between the decision object and its adjacent decision object in the first relationship graph increases, and / or as the updated decision results of the adjacent decision object in the previous iteration increase, the updated decision result of the decision object in the iteration will also increase accordingly.

[0115] In this way, in each iteration, for each decision object, the updated decision result for that decision object can be determined based on the target information of that decision object in the first relationship graph, thereby achieving an update of the object's decision result. After each iteration, it can be determined whether the updated decision results corresponding to multiple decision objects meet the preset convergence conditions. If not, it means that the iteration has not reached a stable state, and it is necessary to proceed to the next iteration. If the preset convergence conditions are met, it means that the iteration has reached a stable state, and the iteration can be determined to have converged. At this time, the updated decision result of each decision object obtained in this iteration can be used as the target decision result corresponding to that decision object.

[0116] The preset convergence condition can be that the degree of change in the updated decision results for all decision objects is less than or equal to a preset change threshold, for example, less than or equal to 0.0001. This degree of change can be obtained by the following process: first, determining the degree of change in the updated decision results obtained for each decision object in multiple consecutive iterations; then, averaging the degrees of change for all decision objects to obtain the average degree of change in the updated decision results for all decision objects.

[0117] As can be seen from the above process, the embodiment of the present application can continuously update the decision result through continuous iteration, so that when the decision result tends to be stable, the updated decision result tends to be stable as the target decision result. In this way, the accuracy of the target decision result can be guaranteed as much as possible.

[0118] In some embodiments, the updated decision result of the decision object in the above iteration is determined based on the dynamic damping coefficient and the target information of the decision object in the first relationship graph, which can specifically include the following steps: first, based on the edge weight of the first connecting edge between the decision object and each adjacent decision object, and the updated decision result corresponding to the adjacent decision object in the previous round of iteration, the intermediate decision result corresponding to the decision object in the above iteration is determined; then, based on the dynamic damping coefficient and the intermediate decision result, the updated decision result corresponding to the decision object in the above iteration is determined.

[0119] The above-mentioned intermediate decision result may be positively correlated with the edge weight of the above-mentioned first connecting edge and the above-mentioned updated decision result.

[0120] The updated decision result corresponding to the decision object in the above iteration is negatively correlated with the above dynamic damping coefficient, and positively correlated with the above intermediate decision result.

[0121] In some embodiments, the dynamic damping coefficient can adjust not only the propagation difficulty but also the intermediate decision results. Therefore, the updated decision result corresponding to the decision object in the iteration is determined based on the dynamic damping coefficient and the intermediate decision results. Specifically, the intermediate decision result can be updated using the dynamic damping coefficient, and the updated decision result corresponding to the decision object in the iteration is determined based on the updated intermediate decision result and the dynamic damping coefficient.

[0122] Among them, the intermediate decision result after updating is positively correlated with the dynamic damping coefficient, and is also positively correlated with the intermediate decision result before updating.

[0123] The following example shows the calculation formula for updating the decision result of a decision object in iteration:

[0124] Where DRb'(ref) is the updated decision result of the b-th decision object Db in the current iteration, h is a constant that can be 1, and DP is the preset damping coefficient. αb is the average similarity between the b-th decision object Db and all its adjacent decision objects in terms of decision factors. K is the number of adjacent decision objects of the b-th decision object Db, αbk is the similarity between the b-th decision object Db and its k-th adjacent decision object Dk in terms of decision factors, and Ybk is the first degree of association between the b-th decision object Db and its k-th adjacent decision object Dk. DRk(ref) is the updated decision result of the k-th adjacent decision object Dk in the previous iteration.

[0125] It should be noted that DP·αb in the above formula (9) can be understood as the dynamic damping coefficient, It can be understood as the intermediate decision result before the update corresponding to the b-th decision object. It can be understood as the updated intermediate decision result corresponding to the b-th decision object. αbk·Ybk can be understood as the third degree of association between the b-th decision object Db and its k-th adjacent decision object Dk.

[0126] In summary, the embodiment of the present application can represent the association relationship between decision objects and the second association strength through a first relationship map, so as to perform decision processing on multiple decision objects based on the second association map. The association information between decision images can be fully utilized to improve the accuracy of decision processing. When the initial decision result of the decision object is calculated based on the decision factors and their weights, the decision processing is performed in combination with the importance of the decision factors and the association information between the decision factors, which can further improve the accuracy of the decision. In addition, the association information can include two aspects: the similarity of the decision objects in the decision factors and the first degree of association between the decision objects, so as to describe the association relationship between the decision objects from two aspects, which can improve the comprehensiveness of the association relationship and help to further improve the accuracy of the decision.

[0127] Figure 5 is a flowchart of another decision-making method based on a relationship graph provided by an embodiment of the present application. Referring to Figure 5 , the detailed process of the above decision-making process includes the following steps.

[0128] S301. Construct a comparison matrix, where the element in the i-th row and j-th column of the comparison matrix is ​​used to represent the relative importance of the i-th decision factor compared to the j-th decision factor, and i and j are both integers greater than or equal to 1.

[0129] S302: Perform unitization processing on each column of the contrast matrix to obtain a unit contrast matrix.

[0130] S303: Use the average value of each row of the unit comparison matrix as the priority of each decision factor in the priority matrix.

[0131] S304: Determine the maximum eigenvalue of the comparison matrix by combining the comparison matrix and the priority matrix.

[0132] S305. Determine the test index according to the maximum eigenvalue and the number of decision factors in the comparison matrix.

[0133] S306: If the inspection index meets the preset conditions, it is determined that the priority matrix verification has passed.

[0134] S307: When the priority matrix passes verification, each priority included in the priority matrix is ​​used as the decision weight of each decision factor.

[0135] S308: Perform decision processing based on the decision weights of the decision factors and the decision factors corresponding to the decision object to obtain an initial decision result of the decision object.

[0136] S309. Construct a second relationship graph using multiple decision objects and at least one third-party object, where each node in the second relationship graph corresponds to a decision object or a third-party object. When there is a second connecting edge between the decision object and the third-party object, the edge weight of the second connecting edge is used to indicate a second degree of association between the decision object and the third-party object.

[0137] S310. For two decision objects connected to at least one same third-party object in the second relationship graph, determine a first degree of association between the two decision objects according to a second degree of association between the two decision objects and the same third-party object.

[0138] S311. Delete the third-party object from the second relationship graph to obtain a first relationship graph, where each node in the first relationship graph corresponds to a decision object. When there is a first connecting edge between two decision objects, the edge weight of the first connecting edge is used to indicate a third degree of association between the two decision objects. The third degree of association is positively correlated with the first degree of association and the similarity of the decision objects in the decision factors.

[0139] S312. In each round of iteration, for each decision object, the preset damping coefficient is updated according to the average similarity between the decision object and each adjacent decision object in the decision factors to obtain a dynamic damping coefficient.

[0140] S313. Determine the intermediate decision result corresponding to the decision object in the iteration according to the edge weight of the first connecting edge and the updated decision result corresponding to the adjacent decision object in the previous round of iteration.

[0141] S314. Update the intermediate decision result by using the dynamic damping coefficient, so as to determine the updated decision result corresponding to the decision object in the iteration according to the updated intermediate decision result and the dynamic damping coefficient.

[0142] S315. After each round of iteration, if the updated decision results corresponding to the multiple decision objects respectively meet the preset convergence condition, the updated decision results corresponding to the multiple decision objects respectively are used as the corresponding target decision results.

[0143] It should be noted that the above steps S301 to S315 can be flexibly adjusted in order based on mutual independence, and this application does not impose any restrictions on their order. Steps S301 to S315 of this application can refer to the description of the corresponding positions in the above steps S201 to S203, and will not be repeated here.

[0144] FIG6 is a block diagram of a decision processing device based on a relationship graph according to an embodiment of the present application. Referring to FIG6 , the decision processing device 400 based on a relationship graph includes:

[0145] The initial decision module 401 is used to determine initial decision results corresponding to multiple decision objects.

[0146] The association determination module 402 is configured to determine association information between the plurality of decision objects, wherein the association information includes similarity between any two decision objects in terms of decision factors.

[0147] The first graph construction module 403 is used to construct a first relationship graph based on the association information between the multiple decision objects, each node in the first relationship graph corresponds to one decision object, and when there is a first connecting edge between two decision objects, the edge weight of the first connecting edge is positively correlated with the association information.

[0148] The target decision module 404 is configured to update the initial decision results corresponding to the plurality of decision objects respectively through the first relationship graph to obtain a target decision result for each of the decision objects.

[0149] Optionally, the target decision module 404 is further configured to:

[0150] For each of the decision objects, the preset damping coefficient is updated to obtain a dynamic damping coefficient based on the average similarity between the decision object and its adjacent decision objects on the decision factors; the initial decision results corresponding to the multiple decision objects in the first relationship graph are updated using the dynamic damping coefficient to obtain the target decision result of each decision object. The dynamic damping coefficient is used to control the difficulty of decision propagation between the decision object and the adjacent decision objects.

[0151] Optionally, the target decision module 404 is further configured to:

[0152] In each round of iteration of the update, for each decision object, the updated decision result of the decision object in the iteration is determined based on the dynamic damping coefficient and the target information of the decision object in the first relationship graph, and the target information includes: the edge weight of the first connecting edge between the decision object and the adjacent decision object of the decision object in the first relationship graph, and the updated decision result of the adjacent decision object in the previous round of iteration, and the updated decision result is initially the initial decision result; after each round of iteration, if the updated decision results corresponding to the multiple decision objects respectively meet the preset convergence conditions, the updated decision results corresponding to the multiple decision objects respectively are used as the corresponding target decision results.

[0153] Optionally, the target decision module 404 is further configured to:

[0154] According to the edge weight of the first connecting edge and the updated decision result corresponding to the adjacent decision object in the previous round of iteration, the intermediate decision result corresponding to the decision object in the iteration is determined; according to the dynamic damping coefficient and the intermediate decision result, the updated decision result corresponding to the decision object in the iteration is determined.

[0155] Optionally, the target decision module 404 is further configured to:

[0156] The intermediate decision result is updated by the dynamic damping coefficient, so as to determine the updated decision result corresponding to the decision object in the iteration according to the updated intermediate decision result and the dynamic damping coefficient.

[0157] Optionally, the association information further includes: a first association degree between any two decision objects, where the first association degree is used to indicate a degree of dependence between the two decision objects; the association determination module 402 is further configured to:

[0158] A second relationship graph is constructed using the multiple decision objects and at least one third-party object, where each node in the second relationship graph corresponds to one decision object or one third-party object. When there is a second connecting edge between the decision object and the third-party object, the edge weight of the second connecting edge is used to indicate a second degree of association between the decision object and the third-party object. For two decision objects connected to at least one same third-party object in the second relationship graph, the first degree of association between the two decision objects is determined based on the second degree of association between the two decision objects and the same third-party object.

[0159] Optionally, the initial decision module 401 is further configured to:

[0160] Obtain decision weights corresponding to at least one decision factor, where the decision factor is used to perform decision processing on the multiple decision objects, and the decision weight is used to indicate the degree of influence of the corresponding decision factor on the decision processing; perform decision processing based on the decision weight of the decision factor and the decision factor corresponding to the decision object to obtain an initial decision result of the decision object.

[0161] Optionally, the initial decision module 401 is further configured to:

[0162] Construct a comparison matrix, wherein the element in the i-th row and j-th column of the comparison matrix is ​​used to represent the relative importance of the i-th decision factor compared to the j-th decision factor, and both i and j are integers greater than or equal to 1; determine a priority matrix based on the comparison matrix, wherein the priority matrix includes the priorities corresponding to each of the decision factors in the comparison matrix; verify the priority matrix, so that when the verification passes, each priority included in the priority matrix is ​​used as the decision weight of each decision factor.

[0163] Optionally, the initial decision module 401 is further configured to:

[0164] Each column of the comparison matrix is ​​normalized to obtain a unit comparison matrix; and the average value of each row of the unit comparison matrix is ​​used as the priority of each decision factor in the priority matrix.

[0165] Optionally, the initial decision module 401 is further configured to:

[0166] The maximum eigenvalue of the comparison matrix is ​​determined by combining the comparison matrix and the priority matrix; a test index is determined based on the maximum eigenvalue and the number of decision factors in the comparison matrix; if the test index meets a preset condition, it is determined that the priority matrix has been verified.

[0167] The above-mentioned device embodiment corresponds to the above-mentioned method embodiment. For specific descriptions, reference can be made to the descriptions in the above-mentioned method embodiment, and the embodiments of the present application will not be repeated here.

[0168] FIG7 is a block diagram of an electronic device according to an embodiment of the present application. The electronic device 600 includes a memory 602 and at least one processor 601 .

[0169] The memory 602 stores computer-executable instructions.

[0170] At least one processor 601 executes the computer-executable instructions stored in the memory 602, so that the electronic device 600 implements the aforementioned decision-making processing method based on the relationship graph.

[0171] In addition, the electronic device 600 may further include a receiver 603 and a transmitter 604. The receiver 603 is configured to receive information from other devices or equipment and forward it to the processor 601. The transmitter 604 is configured to send the information to the other devices or equipment.

[0172] In an exemplary embodiment, a non-transitory computer-readable storage medium is also provided, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the above-mentioned service grid-based device access method.

[0173] In an exemplary embodiment, a computer program product is also provided for implementing the aforementioned service grid-based device access method.

[0174] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of the present application and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0175] It should be understood that the present application is not limited to the exact structure described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A decision-making processing method based on a relational graph, characterized in that The method includes: Determining initial decision results corresponding to multiple decision objects respectively; Determining the association information among the multiple decision objects, where the association information includes the similarity between any two of the decision objects in terms of decision factors; Constructing a first relationship graph according to the association information among the multiple decision objects, where each node in the first relationship graph corresponds to one of the decision objects, and when there is a first connection edge between two of the decision objects, the edge weight of the first connection edge is positively correlated with the association information; Updating the initial decision results corresponding to the multiple decision objects respectively through the first relationship graph to obtain the target decision result of each decision object.

2. The method according to claim 1, characterized in that The updating the initial decision results corresponding to the multiple decision objects respectively through the first relationship graph to obtain the target decision result of each decision object includes: For each decision object, updating a preset damping coefficient to obtain a dynamic damping coefficient according to the average similarity between the decision object and each adjacent decision object of the decision object in terms of the decision factors; Updating the initial decision results corresponding to the multiple decision objects respectively in the first relationship graph through the dynamic damping coefficient to obtain the target decision result of each decision object, where the dynamic damping coefficient is used to control the difficulty of decision propagation between the decision object and the adjacent decision objects.

3. The method according to claim 2, characterized in that The updating the initial decision results corresponding to the multiple decision objects respectively in the first relationship graph through the dynamic damping coefficient to obtain the target decision result of each decision object includes: In each round of iteration of the updating, for each decision object, determining the updated decision result of the decision object in the iteration according to the dynamic damping coefficient and the target information of the decision object in the first relationship graph, where the target information includes: the edge weight of the first connection edge between the decision object and the adjacent decision object of the decision object in the first relationship graph, and the updated decision result of the adjacent decision object in the previous round of iteration, and the updated decision result is the initial decision result at the beginning; After each round of iteration, if the updated decision results corresponding to the multiple decision objects respectively meet a preset convergence condition, then taking the updated decision results corresponding to the multiple decision objects respectively as the corresponding target decision results.

4. The method according to claim 3, wherein The determining the updated decision result of the decision object in the iteration according to the dynamic damping coefficient and the target information of the decision object in the first relationship graph includes: Determining the intermediate decision result of the decision object in the iteration according to the edge weight of the first connection edge and the updated decision result corresponding to the adjacent decision object in the previous round of iteration; Determining the updated decision result of the decision object in the iteration according to the dynamic damping coefficient and the intermediate decision result.

5. The method according to claim 4, wherein The determining the updated decision result of the decision object in the iteration according to the dynamic damping coefficient and the intermediate decision result includes: Update the intermediate decision result with the dynamic damping coefficient, and determine the updated decision result corresponding to the decision object in the iteration according to the updated intermediate decision result and the dynamic damping coefficient.

6. The method according to any one of claims 1 to 5, characterized in that, The associated information further includes: the first association degree between any two decision objects, and the first association degree is used to indicate the dependence degree between the two decision objects; Determining the associated information between the multiple decision objects includes: Construct a second relationship graph through the multiple decision objects and at least one third-party object. Each node in the second relationship graph corresponds to a decision object or a third-party object. When there is a second connection edge between the decision object and the third-party object, the edge weight of the second connection edge is used to indicate the second association degree between the decision object and the third-party object; For two decision objects connected to at least one same third-party object in the second relationship graph, determine the first association degree between the two decision objects according to the second association degrees between the two decision objects and the same third-party object respectively.

7. The method according to any one of claims 1 to 5, characterized in that Determining the initial decision results respectively corresponding to the multiple decision objects includes: Obtain the decision weights respectively corresponding to at least one decision factor. The decision factor is used to perform decision processing on the multiple decision objects, and the decision weight is used to indicate the influence degree of the corresponding decision factor on the decision processing; Perform decision processing according to the decision weight of the decision factor and the decision factor corresponding to the decision object to obtain the initial decision result of the decision object.

8. The method according to claim 7, characterized in that Obtaining the decision weights respectively corresponding to at least one decision factor includes: Construct a comparison matrix. The element in the i-th row and j-th column of the comparison matrix is used to represent the relative importance degree of the i-th decision factor compared to the j-th decision factor, and both i and j are integers greater than or equal to 1; Determine a priority matrix according to the comparison matrix. The priority matrix includes the priorities respectively corresponding to the decision factors in the comparison matrix; Verify the priority matrix, and when the verification passes, use the priorities included in the priority matrix as the decision weights of the decision factors.

9. The method according to claim 8, wherein Determining the priority matrix according to the comparison matrix includes: Perform unit normalization processing on each column of the comparison matrix to obtain a unit comparison matrix; Use the average value of each row of the unit comparison matrix as the priority of each decision factor in the priority matrix.

10. The method according to claim 8, wherein Verifying the priority matrix includes: Combine the comparison matrix and the priority matrix to determine the maximum eigenvalue of the comparison matrix; Determine a test index according to the maximum eigenvalue and the number of decision factors in the comparison matrix; If the test index meets the preset conditions, determine that the priority matrix passes the verification.

11. A decision-making processing device based on a relational graph, characterized in that Includes: An initial decision module, configured to determine the initial decision results respectively corresponding to the multiple decision objects; An association determination module, configured to determine association information among the multiple decision-making objects, where the association information includes the similarity of any two of the decision-making objects in terms of decision-making factors; A first graph construction module, configured to construct a first relationship graph according to the association information among the multiple decision-making objects, where each node in the first relationship graph corresponds to one of the decision-making objects, and when there is a first connection edge between two of the decision-making objects, the edge weight of the first connection edge is positively correlated with the association information; A target decision-making module, configured to update the initial decision-making results respectively corresponding to the multiple decision-making objects through the first relationship graph to obtain the target decision-making result of each decision-making object.

12. An electronic device, including a memory and at least one processor; Among them, The memory stores computer execution instructions; The at least one processor executes the computer execution instructions stored in the memory, so that the electronic device implements the relationship graph-based decision-making processing method according to any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that, Computer execution instructions are stored in the computer-readable storage medium, and when the computer execution instructions are executed by a processor, they are used to implement the relationship graph-based decision-making processing method according to any one of claims 1 to 10.

Images