Automatic bill reimbursement method and device and storage medium

Abstract

The invention relates to the field of bill reimbursement, and discloses an automatic bill reimbursement method and device and a storage medium. The method comprises the following steps: receiving a bill image to be reimbursed; based on a preset OCR algorithm and a preset semantic recognition algorithm, carrying out image recognition processing on the bill image to obtain bill data; when the group verification value is an unqualified value, performing exception elimination processing on the bill data according to preset exception process processing; when the group verification value is a qualified value, obtaining a priority value corresponding to the bill data according to a preset priority calculation formula; and writing the marked bill data into a preset reimbursement processing pipeline, and performing sorting reimbursement processing on the marked bill data based on the sequence of the priority values in the reimbursement processing pipeline to generate a reimbursement result. In the embodiment of the invention, the emergency degree of the bill is intelligently analyzed, and the bill reimbursement priority is utilized to perform sorting processing, so that the emergency bill reimbursement speed is improved, and the bill reimbursement timeliness is ensured.

Description

Technical Field

[0001] This invention relates to the field of expense reimbursement, and more particularly to an automatic expense reimbursement method, device, and storage medium for expense receipts. Background Technology

[0002] The front-end of the current invoice reimbursement system uses Vue 2.7. Vue.js is a progressive framework for building user interfaces, and Vue 2.7, as its latest version, offers improvements in performance, stability, and functionality. Vue integrates slots, providing developers with greater flexibility and extensibility, facilitating the invocation and customization of components. The back-end system of the invoice reimbursement system uses a framework built on Spring Boot. Spring Boot supports rapid application creation and deployment and includes commonly used web servers such as Tomcat and Jetty, simplifying the deployment process. Based on Vue 2.7 and Spring Boot, this framework eliminates the need for developers to write SQL statements. By integrating Vue's slots and Spring Boot's auto-configuration features, it achieves seamless integration and efficient collaboration between the front-end and back-end.

[0003] While frameworks developed using Vue 2.7 and Spring Boot aim to simplify expense reimbursement processes, their current level of intelligence remains limited. The system relies on historical data and individual reporting habits to generate the necessary documents, which restricts its flexibility and adaptability. If employee reporting habits change or new expense types need to be processed, the system may require reconfiguration and adjustments. Therefore, a new technology is needed to address the technical problem of insufficient intelligence in expense reimbursement, leading to poor timeliness. Summary of the Invention

[0004] The main objective of this invention is to solve the technical problem of poor timeliness in current invoice reimbursement due to insufficient intelligence in the process.

[0005] The first aspect of this invention provides an automatic reimbursement method for receipts, comprising the steps of: Receive images of receipts pending reimbursement; Based on a pre-set OCR algorithm and a pre-set semantic recognition algorithm, the invoice image is processed to obtain invoice data, wherein the invoice data includes: invoice number, user ID, invoice type, invoice amount, invoice time, merchant ID, and department information; According to the preset type weight formula, the type weight of the invoice type is calculated to obtain the type weight; The invoice amount is weighted according to a preset amount weighting formula to obtain the amount weight; The attenuation weight is calculated based on the preset time attenuation formula. Based on the preset risk factor calculation formula, risk calculation is performed on the invoice number and the invoice amount to obtain risk parameters; According to the preset group anomaly verification algorithm, the user ID, the merchant ID, and the department information are subjected to group anomaly verification processing to obtain the group verification value; When the group verification value is unqualified, the ticket data is processed according to the preset abnormality process to remove the abnormality. When the group verification value is qualified, priority values ​​are calculated for the type weight, the amount weight, the attenuation weight, the risk parameter, and the group verification value according to the preset priority calculation formula to obtain the priority value corresponding to the bill data. The priority value is then combined with the bill data to obtain the marked bill data. The marked invoice data is written into a pre-set reimbursement processing pipeline, and the marked invoice data is sorted and processed for reimbursement based on priority values ​​in the reimbursement processing pipeline to generate reimbursement results.

[0006] Optionally, in a first implementation of the first aspect of the present invention, the step of performing group anomaly verification processing on the user ID, the merchant ID, and the department information according to a preset group anomaly verification algorithm to obtain a group verification value includes: Determine whether the number of reimbursement invoices submitted by the user ID within the first time window exceeds a preset frequency threshold; When the preset comment threshold is exceeded, the preset unqualified value is assigned to the group verification value to generate the group verification value; If the preset frequency threshold is not exceeded, then based on the department information, it is determined whether the total number of invoices for the merchant ID in more than two departments exceeds the abnormal threshold. When the abnormal threshold is exceeded, the preset unqualified value is assigned to the group verification value to generate the group verification value. If the abnormal threshold is not exceeded, the preset qualified value is copied to the group verification value to generate the group verification value.

[0007] Optionally, in a second implementation of the first aspect of the present invention, the step of calculating the priority value corresponding to the bill data by performing priority value calculation on the type weight, the amount weight, the attenuation weight, the risk parameter, and the group verification value according to a preset priority calculation formula includes: S=[Φ(W t )+η(W a )+ζ(△d)]×R+G, where, Φ(W t ) represents the type weight, η(W)a ) represents the amount weight, ζ(△d) represents the attenuation weight, R represents the risk parameter, G represents the group verification value, S represents the priority value corresponding to the invoice data, and △d represents the number of days between the invoice issuance time and the preset reimbursement deadline.

[0008] Optionally, in a third implementation of the first aspect of the present invention, the step of performing anomaly removal processing on the invoice data according to a preset anomaly handling process includes: When the group verification value is less than the preset first dynamic threshold, the ticket data is frozen. When the group verification value is greater than or equal to a preset first dynamic threshold and less than a preset second dynamic threshold, the ticket data is sent to a preset manual port for review by the manual port.

[0009] Optionally, in a fourth implementation of the first aspect of the present invention, the step of performing risk calculation on the invoice number and the invoice amount according to a preset risk factor calculation formula to obtain risk parameters includes: R=1+e -（α*F+β*D） Where F is the reimbursement frequency corresponding to the invoice number, D is the discrete value of the invoice amount, R is the risk parameter, and α and β are trainable factor parameters.

[0010] Optionally, in a fifth implementation of the first aspect of the present invention, the step of calculating the type weight of the invoice type according to a preset type weight formula to obtain the type weight includes: Φ(W t )=m1+log2(W t ), where W t For the mapping value of the invoice type, Φ(W) t ) represents the type weight, and m1 is the first constant.

[0011] Optionally, in a sixth implementation of the first aspect of the present invention, the step of calculating the amount weight of the invoice amount according to a preset amount weight formula to obtain the amount weight includes: η(W a )=m2*(m3-e -m4*Wa ), where w a For the invoice amount, η(W) a m1 is the monetary weight, m2 is the second constant, m3 is the third constant, and m4 is the fourth constant.

[0012] Optionally, in the seventh implementation of the first aspect of the present invention, the step of calculating the attenuation weight of the invoice issuance time according to a preset time attenuation formula to obtain the attenuation weight includes: ζ(△d)=m5*e -m6*△dWhere △d is the number of days between the invoice issuance date and the preset reimbursement deadline, ζ(△d) is the attenuation weight, m5 is the fifth constant, and m6 is the sixth constant.

[0013] A second aspect of the present invention provides an automatic reimbursement device for invoices, comprising: a memory and at least one processor, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a circuit; the at least one processor invokes the instructions in the memory to cause the automatic reimbursement device for invoices to execute the above-described automatic reimbursement method for invoices.

[0014] A third aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described method for automatic expense reimbursement of invoices.

[0015] In this embodiment of the invention, by intelligently identifying and processing various invoice types, key information on the invoices is automatically extracted and matched with preset intelligent filling settings. The corresponding data of invoice number, user ID, invoice type, invoice amount, invoice time, merchant ID, and department information are analyzed. The abnormal risks in reimbursement are intelligently controlled, and the priority of reimbursement processing is generated based on the reimbursement content. The priority of invoice reimbursement is used for sorting and processing, which improves the speed of reimbursement of urgent invoices and ensures the timeliness of invoice reimbursement. This solves the technical problem of poor timeliness of reimbursement due to insufficient intelligence in current invoice reimbursement. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of one embodiment of the automatic reimbursement method for invoices in this invention; Figure 2 is a schematic diagram of a specific embodiment of the automatic reimbursement method for invoices in an embodiment of the present invention, which involves 102 steps. Figure 3 This is a schematic diagram of a specific embodiment of the automatic reimbursement method for invoices in this invention. Figure 4 This is a schematic diagram of a specific embodiment of the 104 steps of the automatic reimbursement method for invoices in this invention. Figure 5 This is a schematic diagram of one embodiment of the automatic reimbursement device for receipts in this invention. Detailed Implementation

[0017] This invention provides an automatic reimbursement method, device, and storage medium for invoices.

[0018] The embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. While some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the accompanying drawings and embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0019] In the description of the embodiments disclosed in this invention, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc., may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0020] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 An embodiment of the automatic reimbursement method for invoices in this invention includes the following steps: 101. Receive images of receipts pending reimbursement; In this embodiment, please refer to Figure 2a , Figure 2a This is a schematic diagram illustrating an implementation of the automatic reimbursement method for invoices in this invention. Figure 2a The component that allows users to drag and drop invoices and attachments to upload them is located here, allowing users to drag and drop images of invoices to be reimbursed to upload to the system.

[0021] 102. Based on the preset OCR algorithm and preset semantic recognition algorithm, perform image recognition processing on the invoice image to obtain invoice data, wherein the invoice data includes: invoice number, user ID, invoice type, invoice amount, invoice time, merchant ID, and department information; In this embodiment, OCR technology is used to extract key information from the invoice image, and a semantic recognition algorithm is used to analyze the invoice content and automatically match the corresponding invoice data. The invoice data includes: invoice number, user ID, invoice type, invoice amount, invoice time, merchant ID, and department information.

[0022] Specifically, the data to be verified mainly includes the following aspects: 1. Basic invoice information: invoice code, invoice number, invoice date, amount, tax rate, tax amount, etc.

[0023] 2. Reimbursement application information: reimbursement item, reimbursement amount, reimbursement date, applicant information, etc.

[0024] 3. Attachment information: Flight information, travel dates, and travelers in the itinerary; contract number, other party's company name, and contract amount in the contract.

[0025] 4. Preset rules and database information: company reimbursement policies, tax regulations, historical reimbursement records, etc.

[0026] The identified invoice type will be matched with the preset invoice types in the smart data entry settings to ensure that the uploaded invoice type complies with the company's reimbursement regulations.

[0027] Please see Figure 2b Please see Figure 2b , Figure 2b This is another schematic diagram illustrating the automatic expense reimbursement method for invoices in this invention. When an employee uploads an invoice, the system automatically checks the validity of the invoice according to the invoice verification rules in the intelligent data entry settings. The intelligent reimbursement page automatically generates a reimbursement form filling interface based on the template configured in the intelligent data entry settings. For example, if the employee selects the travel expense template, the system will automatically display the relevant fields for travel expenses (such as transportation costs, accommodation costs, etc.). The intelligent reimbursement page automatically fills in some reimbursement information according to the automatic fill rules in the intelligent data entry settings. For example, based on the employee's historical data, it automatically fills in commonly used expense types, amount ranges, etc.

[0028] 103. Calculate the type weight of the invoice type according to the preset type weight formula to obtain the type weight; In this embodiment, the configuration of invoice types includes: electronic invoices (ordinary invoices), VAT electronic ordinary invoices, VAT ordinary invoices, VAT electronic special invoices, invoices, reimbursement invoices, taxi receipts, air tickets, train tickets, OFD invoices, and VAT special invoices. Different invoice types correspond to different mapping values. These different mapping values ​​are based on the type weights set according to actual needs. By calculating the weights of the invoice types, the type weights of the invoice data are obtained. For example: travel expenses (high priority), catering expenses (medium priority), and office supplies expenses (low priority).

[0029] Specifically, step 103 includes the following specific implementation methods: 1031、Φ(W t )=m1+log2(W t ), where W t For the mapping value of the invoice type, Φ(W) t ) represents the type weight, and m1 is the first constant.

[0030] In step 1031, the type weight formula can be set as Φ(W) t )=1+log2(W t ), mapping the invoice type value Wt Substitute these values ​​into the formula to obtain the relevant type weight values.

[0031] 104. Calculate the amount weight of the invoice based on the preset amount weight formula. In this embodiment, different invoice reimbursement amounts have different weights. Generally, the larger the amount, the higher the urgency, and the higher the weight of the amount. The amount weights are generated based on the amount weight formula. For example, the invoice amount weights are set as follows: amount > 1000 yuan (weight = 2), amount ≤ 1000 yuan (weight = 1).

[0032] Specifically, step 104 includes the following specific implementation methods: 1041、η(W a )=m2*(m3-e -m4*Wa ), where w a For the invoice amount, η(W) a m1 is the monetary weight, m2 is the second constant, m3 is the third constant, and m4 is the fourth constant.

[0033] In step 1041, the amount weighting formula has a saturation characteristic, and the amount weighting formula can be η(W) a ) = 2 * (1 - e -0.001*Wa In the process of substituting amounts, the relevant weight gradually decreases as the invoice amount increases.

[0034] 105. Calculate the attenuation weight of the invoice issuance time according to the preset time attenuation formula to obtain the attenuation weight; In this embodiment, the time decay formula is based on the principle that the smaller the number of days between the invoice date and the reimbursement deadline, the greater the weight, indicating that the reimbursement is more urgent.

[0035] Specifically, step 105 includes the following specific implementation methods: 1051, ζ(△d)=m5*e -m6*△d Where △d is the number of days between the invoice issuance date and the preset reimbursement deadline, ζ(△d) is the attenuation weight, m5 is the fifth constant, and m6 is the sixth constant.

[0036] In step 1051, the time decay formula can be ζ(Δd) = 3*e -0.5*△d The weighting is increased sharply as the invoice issuance date approaches the preset reimbursement deadline, based on the number of days between the invoice issuance date and the deadline.

[0037] 106. Based on the preset risk factor calculation formula, perform risk calculation on the invoice number and the invoice amount to obtain risk parameters; In this embodiment, the risk factor calculation formula calculates the risk of the dispersion of reimbursement frequency and reimbursement amount, assesses the risk probability of invoices or reimbursement behavior, and obtains risk parameters.

[0038] Specifically, the 106 steps include the following specific implementation methods: 1061. R = 1 + e -（α*F+β*D） Where F is the reimbursement frequency corresponding to the invoice number, D is the discrete value of the invoice amount, R is the risk parameter, and α and β are trainable factor parameters.

[0039] In step 1061, F represents the reimbursement frequency corresponding to the invoice number. It can be calculated by dividing the average number of submissions of the same type of invoices in the past 7 days by the number of submissions of the same type of invoices in the past 24 hours and then subtracting 1. If the number of submissions in the past 24 hours is much higher than the average number of submissions in the past 7 days, the value of F will be larger, indicating that there is an abnormal frequency.

[0040] D is the discrete value of the invoice amount. N is the total number of invoices of the same type and amount, W ~ a The average of all invoices of the same type, max(W) a W represents the maximum amount of invoices of the same type. a，i Let be the amount of the i-th invoice. For each invoice, calculate the difference between its amount and the maximum amount, divide it by the average amount, and then square it. Finally, take the average of all these values ​​as D. If the invoice amount fluctuates greatly, the value of D will be large, indicating a high degree of dispersion in the amount. Analyzing the frequency and dispersion of reimbursements can help identify sudden increases in reimbursement amounts.

[0041] 107. According to the preset group anomaly verification algorithm, perform group anomaly verification processing on the user ID, the merchant ID, and the department information to obtain the group verification value; In this embodiment, the group verification value can be set to -1, 0, 1, etc. It analyzes whether the same user submits more than 3 invoices of the same type within a set 1-hour period. If this is the case, the group verification value is set to 1, indicating that the invoice is abnormal, thus suppressing high-frequency transactions. Furthermore, if different departments issue more than 50 invoices to the same merchant, the group verification value is set to 1, indicating that the invoice is abnormal, thus preventing collaborative fraud.

[0042] In another implementation, the group verification value can be set to -2, +1.5, -0.5, and 0, representing highly abnormal, normal, slightly abnormal, and risk-free behavior, respectively. A counter G is set up, and the following verification methods are used for integral calculation.

[0043] When three different users of the same merchant submit invoices or expense reports within a short period (e.g., within 1 hour), increment counter G by -2. Expense reports submitted within the same city increment counter G by -0.5. Invoices or expense reports related to emergency disaster response projects increment counter G by 0. A large number of invoices or expense reports submitted within a short period (e.g., within 1 hour) outside of working hours (e.g., 10 PM to 6 AM) increment counter G by -2. Expense reports submitted to a designated special account increment counter G by 1.5.

[0044] Finally, the integral of all satisfied behaviors is counted to obtain a group verification value G.

[0045] Please see Figure 3 , Figure 3 This is a specific embodiment of step 107 of the automatic reimbursement method for invoices in this invention. Step 107 includes the following specific implementation methods: 1071. Determine whether the number of reimbursement invoices submitted by the user ID within the first time window exceeds a preset frequency threshold; 1072. When the preset comment threshold is exceeded, the preset unqualified value is assigned to the group verification value to generate the group verification value; 1073. When the preset frequency threshold is not exceeded, the total number of invoices for the merchant ID in more than two departments is determined based on the department information to see if the abnormal threshold is exceeded. 1074. When the abnormal threshold is exceeded, the preset non-compliance value is assigned to the group verification value to generate the group verification value; 1075. When the abnormal threshold is not exceeded, the preset qualified value is copied to the group verification value to generate the group verification value.

[0046] In steps 1071-1075, the verification value is defined as follows: -1 indicates an abnormal invoice, suggesting high-risk fraud. 0 indicates a normal invoice, meaning no abnormal behavior was detected. 1 indicates a suspicious invoice, requiring further review.

[0047] Check if the same user submitted more than 3 invoices of the same type within 1 hour. If the same user submitted 3 or more invoices of the same type within 1 hour, the verification value is 1 (suspicious invoice); otherwise, the verification value is 0 (normal invoice).

[0048] To determine if there are more than 5 invoices issued by the same merchant across departments, the checksum is 1 (suspicious invoice) if the condition is met, and 0 (normal invoice) if the condition is not met.

[0049] If the invoice is not marked as suspicious after the above logic is followed, set the group verification value to 1.

[0050] 108. When the group verification value is unqualified, the ticket data shall be processed according to the preset abnormality process to perform abnormality removal. In this embodiment, when the group verification value is found to be unqualified, it is directly determined that there are some abnormalities in the invoice reimbursement, and manual intervention is required to confirm the true situation. Based on the abnormality process, the invoice data is processed to eliminate abnormalities.

[0051] For details, please refer to Figure 4 , Figure 4 This is a specific embodiment of the 108 steps of the automatic reimbursement method for invoices in this invention. The 108 steps include the following specific implementation methods: 1081. When the group verification value is less than the preset first dynamic threshold, the ticket data is frozen. 1082. When the group verification value is greater than or equal to the preset first dynamic threshold and less than the preset second dynamic threshold, the ticket data is sent to the preset manual port for review by the manual port.

[0052] In steps 1081-1082, the first dynamic threshold is set to -1.5. If the group verification value G integral is less than -1.5, the bill data is directly identified as abnormal and the bill data is frozen.

[0053] The second dynamic threshold is set to 1. If the group verification value G is between -1.5 and 1, the ticket data will be sent to the preset manual port so that the manual port can verify the relevant facts.

[0054] 109. When the group verification value is qualified, priority values ​​are calculated for the type weight, the amount weight, the attenuation weight, the risk parameter, and the group verification value according to the preset priority calculation formula to obtain the priority value corresponding to the bill data. The priority value is then combined with the bill data to obtain the marked bill data. In this embodiment, a priority value is calculated based on the priority calculation formula, combined with type weight, amount weight, attenuation weight, risk parameter, and group verification value. This priority value determines the priority of the invoice data for reimbursement. When the number of reimbursements is large, invoice data with higher priority values ​​are processed earlier. The priority value is then combined with the invoice data to obtain the marked invoice data.

[0055] Specifically, the "priority calculation formula" for step 109 includes the following specific implementation methods: S=[Φ(W t )+η(W a )+ζ(△d)]×R+G, where, Φ(Wt ) represents the type weight, η(W) a ) represents the amount weight, ζ(△d) represents the attenuation weight, R represents the risk parameter, G represents the group verification value, S represents the priority value corresponding to the invoice data, and △d represents the number of days between the invoice issuance time and the preset reimbursement deadline.

[0056] 110. Write the marked invoice data into a preset reimbursement processing pipeline, and in the reimbursement processing pipeline, sort the marked invoice data according to the priority value to generate reimbursement results.

[0057] In this embodiment, the marked receipt data is written into a preset reimbursement processing pipeline. The reimbursement processing pipeline sorts the marked receipt data according to the priority value in the marked receipt data, generates a reimbursement result, and then returns the reimbursement result to the user who uploaded the receipt image.

[0058] In this embodiment of the invention, by intelligently identifying and processing various invoice types, key information on the invoices is automatically extracted and matched with preset intelligent filling settings. The corresponding data of invoice number, user ID, invoice type, invoice amount, invoice time, merchant ID, and department information are analyzed. The abnormal risks in reimbursement are intelligently controlled, and the priority of reimbursement processing is generated based on the reimbursement content. The priority of invoice reimbursement is used for sorting and processing, which improves the speed of reimbursement of urgent invoices and ensures the timeliness of invoice reimbursement. This solves the technical problem of poor timeliness of reimbursement due to insufficient intelligence in current invoice reimbursement.

[0059] Figure 5 This is a schematic diagram of the structure of an automatic reimbursement device for receipts provided in an embodiment of the present invention. The automatic reimbursement device 500 can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 510 (e.g., one or more processors) and a memory 520, and one or more storage media 530 (e.g., one or more mass storage devices) for storing application programs 533 or data 532. The memory 520 and storage media 530 can be temporary or persistent storage. The program stored in the storage media 530 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the automatic reimbursement device 500. Furthermore, the processor 510 may be configured to communicate with the storage media 530 and execute the series of instruction operations in the storage media 530 on the automatic reimbursement device 500.

[0060] The receipt-based automated reimbursement device 500 may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input / output interfaces 560, and / or one or more operating systems 531, such as Windows Server, Mac OS X, Unix, Linux, Free BSD, etc. Those skilled in the art will understand that... Figure 5 The illustrated structure of the automated expense reimbursement device does not constitute a limitation on the automated expense reimbursement device based on expense receipts. It may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0061] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the automatic reimbursement method for the invoice.

[0062] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0063] Furthermore, although the operations are described in a specific order, this should be understood as requiring that such operations be performed in the specific order shown or in sequential order, or requiring that all illustrated operations be performed to achieve the desired result. In certain environments, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single implementation. Conversely, various features described in the context of a single implementation may also be implemented individually or in any suitable sub-combination in multiple implementations.

[0064] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. An automatic reimbursement method for receipts, characterized in that, Including the following steps: Receive images of receipts pending reimbursement; Based on a pre-set OCR algorithm and a pre-set semantic recognition algorithm, the invoice image is processed to obtain invoice data, wherein the invoice data includes: invoice number, user ID, invoice type, invoice amount, invoice time, merchant ID, and department information; According to the preset type weight formula, the type weight of the invoice type is calculated to obtain the type weight; The invoice amount is weighted according to a preset amount weighting formula to obtain the amount weight; The attenuation weight is calculated based on the preset time attenuation formula. Based on the preset risk factor calculation formula, risk calculation is performed on the invoice number and the invoice amount to obtain risk parameters; According to the preset group anomaly verification algorithm, the user ID, the merchant ID, and the department information are subjected to group anomaly verification processing to obtain the group verification value; When the group verification value is unqualified, the ticket data is processed according to the preset abnormality process to remove the abnormality. When the group verification value is qualified, priority values ​​are calculated for the type weight, the amount weight, the attenuation weight, the risk parameter, and the group verification value according to the preset priority calculation formula to obtain the priority value corresponding to the bill data. The priority value is then combined with the bill data to obtain the marked bill data. The marked invoice data is written into a pre-set reimbursement processing pipeline, and the marked invoice data is sorted and processed for reimbursement based on priority values ​​in the reimbursement processing pipeline to generate reimbursement results.

2. The automatic reimbursement method for receipts according to claim 1, characterized in that, The step of performing group anomaly verification processing on the user ID, the merchant ID, and the department information according to a preset group anomaly verification algorithm to obtain a group verification value includes: Determine whether the number of reimbursement invoices submitted by the user ID within the first time window exceeds a preset frequency threshold; When the preset comment threshold is exceeded, the preset unqualified value is assigned to the group verification value to generate the group verification value; If the preset frequency threshold is not exceeded, then based on the department information, it is determined whether the total number of invoices for the merchant ID in more than two departments exceeds the abnormal threshold. When the abnormal threshold is exceeded, the preset unqualified value is assigned to the group verification value to generate the group verification value. If the abnormal threshold is not exceeded, the preset qualified value is copied to the group verification value to generate the group verification value.

3. The automatic reimbursement method for receipts according to claim 1, characterized in that, The step of calculating the priority value corresponding to the bill data by performing priority value calculation on the type weight, the amount weight, the attenuation weight, the risk parameter, and the group verification value according to the preset priority calculation formula includes: S=[Φ(W t )+η(W a )+ζ(△d)]×R+G, where, Φ(W t ) represents the type weight, η(W) a ) represents the amount weight, ζ(△d) represents the attenuation weight, R represents the risk parameter, G represents the group verification value, S represents the priority value corresponding to the invoice data, and △d represents the number of days between the invoice issuance time and the preset reimbursement deadline.

4. The automatic reimbursement method for receipts according to claim 1, characterized in that, The step of handling the invoice data for anomaly removal according to the preset anomaly process includes: When the group verification value is less than the preset first dynamic threshold, the ticket data is frozen. When the group verification value is greater than or equal to a preset first dynamic threshold and less than a preset second dynamic threshold, the ticket data is sent to a preset manual port for review by the manual port.

5. The automatic reimbursement method for receipts according to claim 1, characterized in that, The step of calculating the risk parameters based on the invoice number and the invoice amount according to the preset risk factor calculation formula includes: R=1+e -（α*F+β*D） Where F is the reimbursement frequency corresponding to the invoice number, D is the discrete value of the invoice amount, R is the risk parameter, and α and β are trainable factor parameters.

6. The automatic reimbursement method for receipts according to claim 1, characterized in that, The step of calculating the type weight of the invoice type according to the preset type weight formula to obtain the type weight includes: Φ(W t )=m1+log2(W t ), where W t For the mapping value of the invoice type, Φ(W) t ) represents the type weight, and m1 is the first constant.

7. The automatic reimbursement method for receipts according to claim 1, characterized in that, The step of calculating the amount weight of the invoice amount according to the preset amount weight formula includes: η(W a )=m2*(m3-e -m4*Wa ), where w a For the invoice amount, η(W) a m1 is the monetary weight, m2 is the second constant, m3 is the third constant, and m4 is the fourth constant.

8. The automatic reimbursement method for receipts according to claim 1, characterized in that, The step of calculating the attenuation weight of the invoice issuance time according to the preset time attenuation formula to obtain the attenuation weight includes: ζ(△d)=m5*e -m6*△d Where △d is the number of days between the invoice issuance date and the preset reimbursement deadline, ζ(△d) is the attenuation weight, m5 is the fifth constant, and m6 is the sixth constant.

9. An automatic reimbursement device for receipts, characterized in that, The automatic reimbursement device for the receipt includes: a memory and at least one processor, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a line; The at least one processor invokes the instructions in the memory to cause the automatic reimbursement device for the invoice to perform the automatic reimbursement method for the invoice as described in any one of claims 1-8.

10. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is executed by the processor, it implements the automatic reimbursement method for invoices as described in any one of claims 1-8.

Images