Target object determination method, device, equipment and medium based on evidence fusion

Abstract

The present application relates to a method, device, computer equipment and storage medium for determining a target object based on evidence fusion. The present application can process the full amount of data of business information and enterprise public information within the scope of the program, and improve the computational efficiency of determining the target object based on business information within the scope of the program. The method includes: obtaining first information of the target enterprise, and obtaining second information for the target enterprise; obtaining multiple pieces of evidence based on the correlation characteristics of public information and logistics information; calculating the above multiple pieces of evidence based on evidence fusion to obtain the confidence of each piece of evidence; obtaining the comprehensive confidence of the target object of the target enterprise that multiple related person objects carried in the second information belong to the target object of the target enterprise according to the confidence of each piece of evidence; taking the above related person objects whose comprehensive confidence is higher than the preset confidence threshold as the target object of the target enterprise.

Description

Technical Field

[0001] The present application relates to the technical field of data analysis, and in particular to a method, apparatus, computer device and storage medium for determining a target object based on evidence fusion. Background Art

[0002] With the booming development of supply chain finance business, identifying corporate customers in the customer base is becoming increasingly important for financial companies. Only by exploring the relationship between customers and the companies they represent can we refine business data based on the company dimension and gain insight into the company's capital needs and risks.

[0003] At present, there are two main methods to obtain the relationship between a customer and the company he represents. One is through the company membership information form, and the other is to verify whether the logistics information of the mobile phone number in the company's public annual report is the company representative information.

[0004] Although the corporate member information table is easy to obtain, its size is too small to quickly expand the business scale; the second method has the problem of inaccurate results due to the large amount of calculation and lack of cross-validation. Summary of the invention

[0005] Based on this, it is necessary to provide a target object determination method, device, computer equipment and storage medium based on evidence fusion to address the above technical problems.

[0006] A target object determination method based on evidence fusion, the method comprising:

[0007] Acquire first information of a target enterprise, and acquire second information for the target enterprise;

[0008] Based on the correlation features between the first information and the second information, obtain multiple pieces of evidence;

[0009] Calculating the plurality of evidences based on evidence fusion to obtain the confidence level of each piece of evidence;

[0010] According to the confidence of each piece of evidence, a comprehensive confidence that the multiple related person objects carried in the second information belong to the target object of the target enterprise is obtained;

[0011] The related person objects whose comprehensive confidence is higher than a preset confidence threshold are taken as target objects of the target enterprise.

[0012] In one embodiment, each of the evidences includes a plurality of focal elements; the calculation of the plurality of evidences based on evidence fusion to obtain the confidence of each evidence includes:

[0013] Based on the DS evidence fusion theory, a basic probability distribution function corresponding to each of the evidences is assigned to each of the evidences;

[0014] Using the basic probability distribution function to calculate the multiple focal elements in each of the evidence to obtain a basic probability distribution value corresponding to each of the focal elements;

[0015] Taking the multiple focal elements contained in each of the evidence as a focal element combination, and taking the basic probability distribution value corresponding to each focal element in the focal element combination as the probability combination corresponding to the focal element combination;

[0016] The probabilities are combined as the confidence of the evidence.

[0017] In one embodiment, obtaining, based on the confidence of each piece of evidence, a comprehensive confidence that the multiple related person objects carried in the second information belong to the target object of the target enterprise includes:

[0018] Constructing a confidence function comprehensive model based on DS evidence fusion theory and each of the basic probability distribution functions;

[0019] The confidences of the multiple pieces of evidence corresponding to the associated features are input into the confidence function combination model to obtain the comprehensive confidence that the multiple related person objects carried in the logistics information belong to the target object of the target enterprise.

[0020] In one embodiment, after the step of taking the plurality of focal elements contained in each of the evidence as a focal element combination and taking the basic probability distribution value corresponding to each focal element in the focal element combination as the probability combination corresponding to the focal element combination, the step further includes:

[0021] Constructing an activation function corresponding to each of the evidences using the prior probability of each of the evidences;

[0022] Using the activation function, screening the probability combination in the evidence corresponding to the activation function to obtain a screened probability combination;

[0023] The combining of the probabilities as the confidence of the evidence comprises:

[0024] The screened probability combination is used as the confidence of the evidence.

[0025] In one embodiment, the focal element combination includes conflicting focal elements; the method of using the multiple focal elements included in each of the evidence as a focal element combination, and using the basic probability distribution value corresponding to each focal element in the focal element combination as the probability combination corresponding to the focal element combination, further includes:

[0026] Set the proportional coefficient corresponding to the conflict focal element;

[0027] Optimizing the basic probability distribution values ​​of the conflicting focal elements according to the proportional coefficients to obtain an optimized probability combination corresponding to the focal element combination;

[0028] The combining of the probabilities as the confidence of the evidence comprises:

[0029] The optimized probability combination corresponding to the focus element combination is used as the confidence of the evidence.

[0030] In one embodiment, each of the evidences contains the same focal element; and obtaining, based on the confidence of each of the evidences, a comprehensive confidence that the multiple related person objects carried in the second information belong to the target object of the target enterprise, further includes:

[0031] The confidence level of the evidence containing the same focal element is obtained by numerical calculation.

[0032] In one embodiment, after combining the probabilities as the confidence of the evidence, the method further includes:

[0033] Calling a stroke processing function to calculate the strokes of the name corresponding to the evidence to obtain an adjustment factor;

[0034] Using the adjustment factor to adjust the comprehensive confidence to obtain an adjusted comprehensive confidence;

[0035] A preset truncation function is used to perform truncation processing on the adjusted comprehensive confidence to obtain a truncated comprehensive confidence.

[0036] A device for determining an enterprise owner based on evidence fusion, the device comprising:

[0037] An information acquisition module, used to acquire first information of a target enterprise and to acquire second information for the target enterprise;

[0038] an evidence acquisition module, configured to obtain a plurality of pieces of evidence based on correlation features between the first information and the second information;

[0039] A confidence calculation module is used to calculate multiple pieces of evidence based on evidence fusion to obtain the confidence of each piece of evidence;

[0040] A comprehensive confidence calculation module, used to obtain a comprehensive confidence that the multiple related person objects carried in the second information belong to the target object of the target enterprise according to the confidence of each piece of evidence;

[0041] The business owner determination module is used to take the related person objects whose comprehensive confidence is higher than a preset confidence threshold as the target objects of the target enterprise.

[0042] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned method for determining an enterprise owner based on evidence fusion are implemented.

[0043] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for determining a business owner based on evidence fusion.

[0044] The above-mentioned target object determination method, device, computer equipment and storage medium based on evidence fusion obtain the first information of the target enterprise and the second information for the target enterprise; obtain multiple evidences based on the correlation characteristics of the first information and the second information; calculate the above-mentioned multiple evidences based on evidence fusion to obtain the confidence of each evidence; obtain the comprehensive confidence that the multiple related person objects carried in the second information belong to the target object of the target enterprise according to the confidence of each evidence; and use the above-mentioned related person objects whose comprehensive confidence is higher than the preset confidence threshold as the target object of the target enterprise. This method can realize the processing of the full amount of business information and social public information, and improve the calculation efficiency of determining the target object based on business information. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A schematic diagram of a flow chart of a target object determination method based on evidence fusion in one embodiment;

[0046] Figure 2 A reasoning framework diagram of multiple pieces of evidence in one embodiment;

[0047] Figure 3 A schematic diagram of a flow chart for calculating the confidence level of each piece of evidence in one embodiment;

[0048] Figure 4 A schematic flow chart of a target object determination method based on evidence fusion in another embodiment;

[0049] Figure 5 is a flow chart of a numerical calculation method in another embodiment;

[0050] Figure 6 It is a structural block diagram of a target object determination device based on evidence fusion in one embodiment;

[0051] Figure 7 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. In addition, the data and information involved in the present disclosure may be data and information authorized by the user or fully authorized by all parties.

[0053] In one embodiment, Figure 1 As shown, a method for determining business owners based on evidence fusion is provided, which can be used on a terminal or server with a processor, wherein the terminal can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, and portable wearable devices, and the server can be implemented as an independent server or a server cluster composed of multiple servers. The above method includes the following steps:

[0054] Step S101, obtaining first information of a target enterprise, and obtaining second information for the target enterprise.

[0055] Among them, the first information of the target enterprise refers to the information that the enterprise voluntarily discloses to the public or is certified by a legal organization and disclosed to the public for legal or legitimate reasons, such as taxpayer number, annual report information, registration information of the enterprise with the industrial and commercial management agency, or corporate credit information reported by the bank, etc. The second information on the target enterprise refers to the business data related to the target enterprise obtained by the user of this patent within its own business scope or the scope of program license and can be used within the legal scope, or the data related to the target enterprise obtained by the user of this patent after being authorized by the target enterprise for legal purposes, such as logistics information related to the target enterprise collected by a logistics company or express company and authorized by the target enterprise for legal purposes, or customer information obtained by a takeaway company, etc.; the logistics information may be a courier note containing the name of the enterprise or the name of a person related to the enterprise, etc., or the logistics information collected by the logistics company is only used for internal analysis, such as only used for precision marketing or service analysis within the scope of the business license.

[0056] Specifically, first information disclosed by the enterprise to the public is collected, such as the name of the enterprise can be obtained through the annual report disclosed by the enterprise, and the relevant information of the people therein is stored in the database; second information related to the target enterprise is obtained, such as the logistics information obtained by the logistics company within the scope of program permission, such as express delivery information containing the company name or the name of the person related to the company, which is stored in the database.

[0057] Step S102, obtaining multiple pieces of evidence based on the correlation features between the first information and the second information.

[0058] Evidence refers to clues used to determine the relevance of the above-mentioned second information to the target enterprise, such as the mobile phone number in the enterprise annual report and the related or identical mobile phone number in the logistics information, the registered address of the enterprise and the related or identical address in the logistics information, etc. Figure 2 shown.

[0059] Specifically, by searching for clues in the first information of the target enterprise, such as public information, which are related or identical to the second information, such as logistics information, multiple pieces of evidence can be obtained.

[0060] Combined with the factors that affect the quality of the results and the actual business scenarios, the following characteristics need to be considered to determine the association:

[0061] Evidence 1: The name of the sender or recipient corresponding to the mobile phone number in the annual report is equal to the name of the company representative; the value is 0.9 for "yes", 0.05 for "no", 0.05 for "not sure"; or the value is 0.6 for "yes", 0.3 for "no", 0.1 for "not sure".

[0062] Evidence 2: The name of the sender and recipient corresponding to the mobile phone number extracted from the annual report mailbox is equal to the name of the company representative; the value of "yes": 0.9, "no": 0.05, "unclear": 0.05; or the value of "yes": 0,6, "no" 0.2, "unclear" 0,2.

[0063] Evidence 3: GPS distance: The Euclidean distance between the GPS location of the courier when it receives or delivers the item and the GPS location of the company’s registered place of business. The Euclidean distance formula is as follows:

[0064]

[0065] Where (x1, y1) is the latitude and longitude value of the GPS positioning of the express delivery address; (x2, y2) represents the latitude and longitude value of the GPS positioning of the delivery address.

[0066] Evidence 4: Similarity between the express delivery address and the registered address: After extracting the province, city, and district information from both, the similarity is calculated using the edit distance between the two, and an adjustment factor is made based on the consecutive letters and numbers in the string.

[0067] Evidence 5: Similarity between express delivery address and company name: After extracting the province, city and district information for both, we then extract the common suffixes of the company name, and finally calculate the similarity based on the edit distance between the two.

[0068] Evidence 6: Name complexity: Nonlinear mapping is performed based on the number of strokes x in the name. The mapping function is:

[0069]

[0070] Among them, R is the average number of strokes in the name.

[0071] Step S103, calculating the above multiple evidences based on evidence fusion to obtain the confidence of each evidence;

[0072] Among them, DS (Dempster-Shafer) evidence fusion theory is an important method of uncertain reasoning. It originated in the 1960s, when American mathematician Dempster began to use upper and lower limit probabilities to solve multi-valued mapping problems, and published many papers since 1967, marking the birth of evidence theory. Dempster's student Shafer introduced the concept of trust function and made further development of evidence theory.

[0073] Specifically, the above-mentioned evidences are probability-calculated through the DS (Dempster-Shafer) evidence fusion theory to obtain the credibility (confidence) of each piece of evidence on whether the recipient or sender is the owner of the target enterprise, and the confidence of each piece of evidence is calculated.

[0074] Step S104, based on the confidence of each of the above evidences, obtain the comprehensive confidence that the multiple related person objects carried in the above logistics information belong to the target object of the above target enterprise.

[0075] Specifically, based on the DS evidence fusion theory, according to the confidence of each of the above-mentioned evidences, the comprehensive confidence of the combination of multiple evidences is calculated, which is used to determine whether each relevant person object carried in the second information, such as the logistics information, such as the logistics consignor, belongs to the target object of the target enterprise (the target object may be the owner of the enterprise). The calculation method may, for example, set a corresponding weight for the confidence of each evidence, and weight the confidence of all evidence to obtain a comprehensive confidence; for example, a suitable comprehensive confidence function may be set to calculate the comprehensive confidence of the combination of multiple evidences.

[0076] Step S105, the logistics consignment and collection entity whose comprehensive confidence is higher than the preset confidence threshold is taken as the owner of the target enterprise.

[0077] Specifically, a confidence threshold is set, for example 80%, and the reasoning result corresponding to the comprehensive confidence that is higher than the confidence threshold of 80% is taken as the final conclusion, that is, the relevant person object such as the recipient or sender is determined to be the target object of the target enterprise (such as the business owner), and the association result is imported into the named entity link library.

[0078] In the above embodiment, the first information of the target enterprise and the second information for the target enterprise are obtained; multiple pieces of evidence are obtained based on the correlation characteristics of the first information and the second information; the confidence of each piece of evidence is calculated based on the evidence fusion; the comprehensive confidence that multiple related person objects carried in the logistics information belong to the target object of the target enterprise is obtained according to the confidence of each piece of evidence; the related person objects whose comprehensive confidence is higher than the preset confidence threshold are used as the target objects of the target enterprise. This method can realize the processing of the full amount of user data and improve the calculation efficiency of determining the target object based on business information.

[0079] In one embodiment, each of the above evidences includes multiple focal elements, such as Figure 3 As shown, the above step S103 includes:

[0080] Step S301, based on the DS evidence fusion theory, assigning a basic probability distribution function corresponding to each piece of evidence one by one to each piece of evidence;

[0081] Specifically, a basic probability distribution function is assigned to each piece of evidence. For example, for the above-mentioned evidence 1: the name of the sender and recipient corresponding to the mobile phone number in the annual report is equal to the name of the corporate representative, a corresponding basic probability distribution function m1 is assigned.

[0082] Evidence 2: The name of the sender and recipient corresponding to the mobile phone number extracted from the annual report mailbox is equal to the name of the company representative, and the corresponding basic probability distribution function m2 is assigned.

[0083] Step S302, using the basic probability distribution function to calculate the multiple focal elements in each piece of evidence to obtain the basic probability distribution value corresponding to each focal element;

[0084] Specifically, for example, Evidence 1: the name of the sender and recipient corresponding to the mobile phone number in the annual report is equal to the name of the company representative; the probability of the focal element in Evidence 1 is allocated through the basic probability distribution function m1 of Evidence 1, such as the value of "yes": 0.9, "no": 0.05.

[0085] Evidence 2: The name of the sender and recipient corresponding to the mobile phone number extracted from the annual report mailbox is equal to the name of the company representative; the probability of the focal element in Evidence 2 is allocated through the basic probability distribution function m2 of Evidence 2, such as the value of "yes": 0.9, "no": 0.05.

[0086] Step S303, taking the multiple focal elements contained in each piece of evidence as a focal element combination, and taking the basic probability distribution value corresponding to each focal element in the focal element combination as the probability combination corresponding to the focal element combination;

[0087] Specifically, for example, the focus element combination of evidence 1 is {(yes), (no)}, and its corresponding probability combination is {0.9, 0.05}.

[0088] Step S304: Combining the probabilities as the confidence of the evidence.

[0089] Specifically, for example, the above-mentioned probability combination is used as the confidence of evidence 1.

[0090] In the above embodiment, the confidence of each piece of evidence is obtained by specifying a basic probability distribution function for each piece of evidence, thereby providing a data basis for the subsequent determination of the business owner.

[0091] In one embodiment, if Figure 4 As shown, Figure 4 The flowchart for calculating the comprehensive confidence level is shown in step S104, which includes:

[0092] Step S401, constructing a confidence function comprehensive model based on DS evidence fusion theory and each basic probability distribution function.

[0093] Specifically, the basic probability distribution function is the probability distribution function corresponding to each piece of evidence, also called the confidence function of evidence i. The confidence function comprehensive model is:

[0094]

[0095]

[0096] Among them, A is the target event, M is the comprehensive confidence function, and m i is the confidence function (basic probability distribution function) of evidence i, is the evidence fusion operator (or orthogonal sum), A i It represents a set of evidence i that contains the target event, that is, the i-th focal element, N is the number of evidence, and K is the normalization coefficient.

[0097] For example, Evidence 1: The name of the sender and recipient corresponding to the mobile phone number in the annual report is equal to the name of the company representative; the basic probability distribution function m1 of Evidence 1 is used to distribute the probability of the focal element in Evidence 1, such as the value of "yes": 0.9, "no": 0.05.

[0098] Evidence 2: The name of the sender and recipient corresponding to the mobile phone number extracted from the annual report mailbox is equal to the name of the company representative; the basic probability distribution function m2 of Evidence 2 is used to distribute the probability of the focal element in Evidence 2, such as the value of "yes": 0.9, "no": 0.05, "unclear": 0.05.

[0099] Step S402, input the confidences of the multiple pieces of evidence corresponding to the associated features into the confidence function combination model to obtain the comprehensive confidence that the multiple related person objects carried in the logistics information belong to the target object of the target enterprise.

[0100] Specifically, the confidence of each focal element of the above-mentioned multiple pieces of evidence is input into the above-mentioned confidence function combination model to obtain the comprehensive confidence of whether each recipient or sender in the target event is the owner of the target enterprise.

[0101] The above embodiment establishes a basic probability distribution function for multiple pieces of evidence and constructs a confidence function combination model based on the basic probability distribution function, thereby ultimately obtaining multiple confidence levels for the target event and improving the computational efficiency of determining business owners based on logistics information.

[0102] In one embodiment, after the above step S303, the method further includes: constructing an activation function corresponding to each of the evidences using the prior probability of each of the evidences; using the activation function to filter the probability combination in the evidences corresponding to the activation function to obtain the filtered probability combination;

[0103] Specifically, based on the original feature correlation, a ReLU-like activation function is superimposed:

[0104] f(x)=max(b,x)

[0105] Among them, b is the prior probability of the application scenario.

[0106] The above step S304 includes: using the screened probability combination as the confidence of the evidence.

[0107] Specifically, the filtered probability combination is used as the confidence of each piece of evidence.

[0108] In one embodiment, the focal element combination includes a conflicting focal element; the step S303 further includes: setting a proportional coefficient corresponding to the conflicting focal element; optimizing the basic probability distribution value of the conflicting focal element according to the proportional coefficient to obtain an optimized probability combination corresponding to the focal element combination;

[0109] Specifically, the focal elements in each body of evidence are defined as "yes", "no", and "unclear", where "yes" and "no" are completely mutually exclusive. The correlation of the features in the application scenario corresponds to the focal element "yes", and then the complementary part of the correlation of the features is allocated to "no" and "unclear" according to the specified ratio w. The default value of w is 0.2, which can also be manually adjusted.

[0110] The above step S304 includes: taking the optimized probability combination corresponding to the focus element combination as the confidence of the evidence.

[0111] Specifically, the optimized probability combination after the above tuning is used as the confidence of each piece of evidence.

[0112] In the above embodiment, the probability of conflicting focal elements is optimized by setting a proportionality coefficient, thereby improving the accuracy of calculation.

[0113] In one embodiment, the above-mentioned evidences contain the same focal element. The above-mentioned step S402 includes: obtaining the confidence of the evidences containing the same focal element by numerical calculation.

[0114] Specifically, the calculation process of the DS synthesis rule is numerically optimized, and the calculation of the common focus set of all arbitrary evidence bodies is converted into direct access to arrays a and k.

[0115]

[0116]

[0117] Among them, i represents the i-th evidence, j is an array in array a or array k, and arrays a and k represent the probability of taking the i-th focal element in each evidence.

[0118] For example, there are 2 pieces of evidence, each piece of evidence has 3 focal elements (situations), and each focal element has a probability. Then a=[(0,0),(0,2),(2,0)], then when j=(0,2),i=1, A_(j_i)=A_0=“yes”, m_i(A_(j_i))=m_1(A_0) represents the probability of taking “yes” in the first piece of evidence. Algorithm for arrays a and k Figure 5 As shown, r = [0, 1, 2] represents three focal elements, "yes", "no", and "unclear", respectively. The number of times the array [0, 1, 2] is iterated depends on the number of pieces of evidence. Then, using the greedy algorithm + backtracking, the data that meet rule 1 (including 0 and not including 1) are added to array a, and the data that meet rule 2 (not including 0 and 1 at the same time) are added to array k.

[0119] In the above embodiment, the accuracy of calculation is improved by using a numerical optimization algorithm.

[0120] In one embodiment, after the above step S304, it also includes: calling a stroke processing function to calculate the strokes of the name corresponding to the evidence to obtain an adjustment factor; using the adjustment factor to adjust the comprehensive confidence to obtain an adjusted comprehensive confidence; using a preset truncation function to truncate the adjusted comprehensive confidence to obtain a truncated comprehensive confidence.

[0121] Specifically, the confidence after fusion is adjusted according to the complexity of the name. The stroke processing function is called to calculate the strokes of the name corresponding to the evidence to obtain the adjustment factor

[0122] Secondly, use the multiplication model to adjust the comprehensive confidence. The most convenient model for engineering practice is the multiplication model, that is, comprehensive confidence * adjustment factor. However, this will cause the adjusted comprehensive confidence to exceed 100%, so these confidences are truncated. According to business rules, the truncation function is set as:

[0123] f(x)=max(0,min(0.99,x))

[0124] Where x is the comprehensive confidence. Note: The range of the upper and lower bounds can be set according to the actual business scenario and the impact of each indicator.

[0125] In the above embodiment, the outliers are optimized by setting the truncation function, invalid data is removed, and the accuracy of the calculation results is further improved.

[0126] It should be understood that although Figures 1 to 5 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figures 1 to 5 At least part of the steps may include multiple steps or multiple stages. These steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.

[0127] In one embodiment, Figure 6 As shown, a target object determination device 600 based on evidence fusion is provided, comprising: an information acquisition module 601, an evidence acquisition module 602, a confidence calculation module 603, a comprehensive confidence calculation module 604 and a target object determination module 605, wherein:

[0128] The information acquisition module is used to acquire first information of a target enterprise and acquire second information for the target enterprise.

[0129] The evidence acquisition module is used to obtain multiple pieces of evidence based on the correlation features between the first information and the second information.

[0130] The confidence calculation module is used to calculate multiple pieces of evidence based on evidence fusion to obtain the confidence of each piece of evidence.

[0131] The comprehensive confidence calculation module is used to obtain the comprehensive confidence that the multiple related person objects carried in the second information belong to the target object of the target enterprise according to the confidence of each piece of evidence.

[0132] The target object determination module is used to take the relevant person objects whose comprehensive confidence is higher than a preset confidence threshold as the target objects of the target enterprise.

[0133] In one embodiment, the above-mentioned confidence calculation module 603 is also used to: based on the DS evidence fusion theory, assign a basic probability distribution function corresponding to each of the evidences; use the basic probability distribution function to calculate the multiple focal elements in each of the evidences to obtain the basic probability distribution value corresponding to each of the focal elements; use the multiple focal elements contained in each of the evidences as a focal element combination, and use the basic probability distribution value corresponding to each focal element in the focal element combination as the probability combination corresponding to the focal element combination; use the probability combination as the confidence of the evidence.

[0134] In one embodiment, the above-mentioned comprehensive confidence calculation module 604 is further used to: construct a comprehensive confidence function model based on the DS evidence fusion theory and each of the basic probability distribution functions; input the confidence of the multiple evidences corresponding to the associated features into the confidence function combination model to obtain the comprehensive confidence that the multiple logistics consignment and collection entities carried in the logistics information belong to the owners of the target enterprise.

[0135] In one embodiment, the confidence calculation module 603 is also used to construct an activation function corresponding to each of the evidences using the prior probabilities of the evidences; use the activation function to filter the probability combinations in the evidences corresponding to the activation function to obtain the filtered probability combinations; and use the optimized probability combination corresponding to the focal element combination as the confidence of the evidence.

[0136] In one embodiment, the above-mentioned comprehensive confidence calculation module 604 is further used to: obtain the confidence of the evidence containing the same focal element through numerical calculation.

[0137] In one embodiment, the above-mentioned comprehensive confidence calculation module 604 is also used to: use the complexity of the name to adjust the comprehensive confidence to obtain the adjusted comprehensive confidence; use a preset truncation function to truncate the adjusted comprehensive confidence to obtain the truncated comprehensive confidence.

[0138] For the specific limitations of the business owner determination device based on evidence fusion, please refer to the limitations of the business owner determination method based on evidence fusion above, which will not be repeated here. Each module in the above-mentioned business owner determination device based on evidence fusion can be implemented in whole or in part through software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0139] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 7 As shown. The computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store public information of enterprises and logistics information. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for determining an enterprise owner based on evidence fusion is implemented.

[0140] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0141] In one embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the steps in the above-mentioned embodiment of the method for determining a business owner based on evidence fusion are implemented.

[0142] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps in the above-mentioned embodiment of the method for determining the business owner based on evidence fusion are implemented. A person of ordinary skill in the art can understand that the implementation of all or part of the process in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the process of the embodiment of each method as described above. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0143] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0144] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A target object determination method based on evidence fusion, characterized in that: The method comprises: Acquire first information of a target enterprise, and acquire second information about the target enterprise; the first information includes public information; the second information includes logistics information; Based on the correlation characteristics of the first information and the second information, a plurality of evidences are obtained; the obtaining of the plurality of evidences based on the correlation characteristics of the first information and the second information includes: querying the social public information of the target enterprise for clues related to or identical to the logistics information as evidence to obtain the plurality of evidences; The multiple evidences are calculated based on evidence fusion to obtain the confidence of each evidence; the confidence of each evidence includes the confidence of each evidence on the conclusion of whether the recipient or sender is the target object of the target enterprise; each evidence contains multiple focal elements; the multiple evidences are calculated based on evidence fusion to obtain the confidence of each evidence, including: based on DS evidence fusion theory, each evidence is assigned a basic probability distribution function corresponding to each evidence; the basic probability distribution function is used to calculate the multiple focal elements in each evidence to obtain the basic probability distribution value corresponding to each focal element; the multiple focal elements contained in each evidence are used as focal element combinations, and the basic probability distribution value corresponding to each focal element in the focal element combination is used as the probability combination corresponding to the focal element combination; the probability combination is used as the confidence of the evidence; According to the confidence of each piece of evidence, a comprehensive confidence that the multiple related person objects carried in the second information belong to the target object of the target enterprise is obtained; the related person objects include a recipient or a sender; The related person objects whose comprehensive confidence is higher than a preset confidence threshold are taken as target objects of the target enterprise.

2. The method according to claim 1, characterized in that The step of obtaining, based on the confidence levels of the respective pieces of evidence, a comprehensive confidence level that the plurality of related person objects carried in the second information belong to the target object of the target enterprise includes: Constructing a confidence function comprehensive model based on DS evidence fusion theory and each of the basic probability distribution functions; The confidences of the multiple pieces of evidence corresponding to the associated features are input into the confidence function combination model to obtain the comprehensive confidence that the multiple related person objects carried in the logistics information belong to the target object of the target enterprise.

3. The method according to claim 2, characterized in that After the step of taking the plurality of focal elements contained in each of the evidence as a focal element combination, and taking the basic probability distribution value corresponding to each focal element in the focal element combination as the probability combination corresponding to the focal element combination, the method further includes: Constructing an activation function corresponding to each of the evidences using the prior probability of each of the evidences; Using the activation function, screening the probability combination in the evidence corresponding to the activation function to obtain a screened probability combination; The combining of the probabilities as the confidence of the evidence comprises: The screened probability combination is used as the confidence of the evidence.

4. The method according to claim 3, characterized in that The focal element combination includes conflicting focal elements; the method of using the multiple focal elements included in each of the evidence as a focal element combination, and using the basic probability distribution value corresponding to each focal element in the focal element combination as the probability combination corresponding to the focal element combination, further includes: Set the proportional coefficient corresponding to the conflict focal element; Optimizing the basic probability distribution values ​​of the conflicting focal elements according to the proportional coefficients to obtain an optimized probability combination corresponding to the focal element combination; The combining of the probabilities as the confidence of the evidence comprises: The optimized probability combination corresponding to the focus element combination is used as the confidence of the evidence.

5. The method according to claim 2, characterized in that: The various pieces of evidence contain the same focal element; and obtaining, based on the confidence of the various pieces of evidence, a comprehensive confidence that the multiple related person objects carried in the second information belong to the target object of the target enterprise, further includes: The confidence level of the evidence containing the same focal element is obtained by numerical calculation.

6. The method according to claim 5, characterized in that After combining the probabilities as the confidence of the evidence, the method further includes: Calling a stroke processing function to calculate the strokes of the name corresponding to the evidence to obtain an adjustment factor; Using the adjustment factor to adjust the comprehensive confidence to obtain an adjusted comprehensive confidence; A preset truncation function is used to perform truncation processing on the adjusted comprehensive confidence to obtain a truncated comprehensive confidence.

7. A target object determination device based on evidence fusion, characterized in that: The device comprises: An information acquisition module, used to acquire first information of a target enterprise and second information about the target enterprise; the first information includes public information; the second information includes logistics information; An evidence acquisition module, configured to obtain multiple pieces of evidence based on the correlation characteristics of the first information and the second information; the obtaining multiple pieces of evidence based on the correlation characteristics of the first information and the second information includes: querying the social public information of the target enterprise for clues related to or identical to the logistics information as evidence to obtain multiple pieces of evidence; A confidence calculation module is used to calculate the multiple evidences based on evidence fusion to obtain the confidence of each evidence; the confidence of each evidence includes the confidence of each piece of evidence for the conclusion of whether the recipient or sender is the target object of the target enterprise; each piece of evidence contains multiple focal elements; the calculation of the multiple evidences based on evidence fusion to obtain the confidence of each evidence includes: based on DS evidence fusion theory, assigning a basic probability distribution function corresponding to each piece of evidence to each piece of evidence; using the basic probability distribution function to calculate the multiple focal elements in each piece of evidence to obtain the basic probability distribution value corresponding to each focal element; using the multiple focal elements contained in each piece of evidence as a focal element combination, and using the basic probability distribution value corresponding to each focal element in the focal element combination as the probability combination corresponding to the focal element combination; using the probability combination as the confidence of the evidence; A comprehensive confidence calculation module, used to obtain, according to the confidence of each piece of evidence, a comprehensive confidence that the multiple related person objects carried in the second information belong to the target object of the target enterprise; the related person objects include a recipient or a sender; The target object determination module is used to take the relevant person objects whose comprehensive confidence is higher than a preset confidence threshold as the target objects of the target enterprise.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images