Search

CN-122023034-A - Reverse billing full-flow management system based on multi-system docking

CN122023034ACN 122023034 ACN122023034 ACN 122023034ACN-122023034-A

Abstract

The invention relates to the technical field of tax informatization, in particular to a reverse billing whole flow management system based on multi-system docking, the standardized safe butt joint of the tax system, the business system, the enterprise internal management system and the natural person transaction end system is realized by constructing a multi-system butt joint module, a data integration and processing module, a qualification authentication and risk management and control module, an automatic billing and tax declaration module and a blockchain evidence storage and data tracing module. The method adopts the cryptofactor model to carry out intelligent qualification authentication, identifies abnormal transaction modes by calculating intra-class divergence matrix, ensures that data cannot be tampered by using a blockchain technology, solves the problem of data island caused by incompatibility of interfaces among multiple systems and different data standards in the renewable resource recycling service, realizes full-flow automatic closed-loop management from qualification audit, transaction verification to automatic billing and tax declaration of reverse billing service, remarkably improves service processing efficiency, and reduces enterprise operation cost.

Inventors

  • XU PENG
  • LV FENGHUI
  • YAN JIAGUO

Assignees

  • 资环链金再生资源有限公司

Dates

Publication Date
20260512
Application Date
20260127

Claims (8)

  1. 1. A reverse billing full flow management system based on multi-system docking, comprising: the multi-system docking module is used for carrying out data interaction with the tax system, the business system, the enterprise internal management system and the natural person transaction end system, adopting a standardized interface adaptation protocol, and integrating SSL/TLS encryption transmission, single sign-on authentication and API key dynamic authorization management technology; The data integration and processing module is used for automatically collecting core data from a plurality of heterogeneous systems, denoising, redundancy removing and format standardization processing are carried out on the multi-source heterogeneous data by utilizing a data cleaning algorithm, and real-time synchronization and integration of the data are carried out by an ETL tool, so that a unified central data warehouse is constructed; The qualification authentication and risk management and control module is used for intelligently evaluating an entity applying qualification, analyzing transaction data in real time based on a rule engine and a machine learning model, and identifying potential tax compliance risks and false transaction risks; the automatic billing and tax payment reporting module is used for automatically triggering billing application after settlement is completed, automatically billing reverse invoices on the receipt number electronic billing platform and the tax system, automatically summarizing billing data and tax details, and generating a standardized tax payment reporting form; And the block chain certification and data tracing module is used for uploading the key business data to the alliance chain for certification, generating a unique hash value, providing a data tracing engine and supporting tracing the circulation track and operation record of the data in the whole flow according to the key field.
  2. 2. The reverse billing process management system based on multi-system docking according to claim 1, wherein the qualification certification in the qualification certification and risk management module adopts a hidden factor model, and by finding a user hidden factor matrix P and an article hidden factor matrix Q, p×q ζ is approximately fitted to a known original scoring matrix R, and the mathematical expression is: the fit for a single association is expressed as: Wherein r ui represents the true score of the user u on the object i, p u represents the hidden vector of the user u, represents the preference degree of the user on k hidden factors, q i represents the hidden vector of the object i, represents the attribution degree of the object on k hidden factors, k represents the dimension of the hidden factors, is a super parameter, and T represents transposition. The objective function is a minimization loss function: Wherein, the Is the square error term of the error, For regularization terms, where λ is a regularization coefficient for controlling the intensity of regularization, p u q i 2 are the L2 norms of vectors p u and q i , respectively, and the optimization process uses a random gradient descent algorithm for iterative update.
  3. 3. The reverse billing process management system based on multi-system docking according to claim 2, wherein the update formula of the random gradient descent algorithm is: Wherein, gamma is the learning rate, and the step length of each update is controlled, and e ui is:
  4. 4. A reverse billing process management system based on multi-system interfacing according to claim 3, wherein risk management in said qualification and risk management module comprises computing intra-class divergence matrices assuming C classes, for class i data, the number of samples is ni, the sample set is: the divergence matrix for that category is: where the parameter x is the single sample data vector belonging to class i and mu i is the mean vector of all samples in class i. The overall intra-class divergence matrix is: wherein the formula of S i is:
  5. 5. the reverse billing process management system based on multi-system docking according to claim 4, wherein the multi-system docking module further comprises an interface access frequency limiting and exception interception mechanism for monitoring interface call frequency in real time, automatically triggering interception when an exception access behavior is detected, and recording an exception log for alerting.
  6. 6. The system of claim 5, wherein the workflow of the automatic billing and tax declaration module comprises automatically generating an electronic contract and pushing and signing after the transaction is completed, automatically generating a settlement bill and triggering payment after the receipt quality inspection is passed, automatically triggering billing application after the settlement is completed, automatically filling billing information, check limit and compliance into the digital billing platform, completing automatic billing of the reverse invoice, and returning the billing result to the related service system in real time.
  7. 7. The reverse billing process management system based on multi-system docking according to claim 6, wherein the blockchain certification and data tracing module uploads key business data such as qualification certification records, transaction certificates, billing information, risk early warning logs and operation traces to the alliance chain for certification, generates a unique hash value, and ensures non-falsification and traceability of the data.
  8. 8. The reverse billing process management system of claim 7 wherein the data integration and processing module is further configured to correlate and map critical data such as merchant stream, funds stream, ticket stream and information stream and generate a non-tamperable distributed certificate.

Description

Reverse billing full-flow management system based on multi-system docking Technical Field The invention relates to the technical field of tax informatization, in particular to a reverse billing whole flow management system based on multi-system docking. Background In the reverse billing business of renewable resources, the existing management system has the problem of insufficient multi-system cooperation. Interfaces among the tax system, the business system, the recovery enterprise internal management system and the natural person transaction end system are incompatible, so that a data island is formed, manual secondary input is needed for data transmission, the efficiency is low, and errors are easy to occur. Meanwhile, the multi-system data verification relies on manual verification, the compliance risk is high, the full-flow closed-loop management of qualification audit, transaction verification, billing application and tax declaration cannot be realized, the grounding efficiency of a reverse billing policy is seriously influenced, and the operation cost and the supervision difficulty of enterprises are increased. Disclosure of Invention The invention aims to provide a reverse billing whole flow management system based on multi-system butt joint, which realizes standardized safe butt joint of tax system, business system, enterprise internal management system and natural person transaction end system by constructing a multi-system butt joint module, a data integration and processing module, a qualification authentication and risk management and control module, an automatic billing and tax declaration module and a blockchain certification and data traceability module, and solves the problems in the background technology. In order to achieve the above purpose, the present invention provides the following technical solutions: A reverse billing process management system based on multi-system docking, comprising: the multi-system docking module is used for carrying out data interaction with the tax system, the business system, the enterprise internal management system and the natural person transaction end system, adopting a standardized interface adaptation protocol, and integrating SSL/TLS encryption transmission, single sign-on authentication and API key dynamic authorization management technology; The data integration and processing module is used for automatically collecting core data from a plurality of heterogeneous systems, denoising, redundancy removing and format standardization processing are carried out on the multi-source heterogeneous data by utilizing a data cleaning algorithm, and real-time synchronization and integration of the data are carried out by an ETL tool, so that a unified central data warehouse is constructed; The qualification authentication and risk management and control module is used for intelligently evaluating an entity applying qualification, analyzing transaction data in real time based on a rule engine and a machine learning model, and identifying potential tax compliance risks and false transaction risks; the automatic billing and tax payment reporting module is used for automatically triggering billing application after settlement is completed, automatically billing reverse invoices on the receipt number electronic billing platform and the tax system, automatically summarizing billing data and tax details, and generating a standardized tax payment reporting form; And the block chain certification and data tracing module is used for uploading the key business data to the alliance chain for certification, generating a unique hash value, providing a data tracing engine and supporting tracing the circulation track and operation record of the data in the whole flow according to the key field. Preferably, the qualification authentication and risk management module adopts a hidden factor model, and by finding the user hidden factor matrix P and the article hidden factor matrix Q, the P×Q≡t is approximately fitted with the known original scoring matrix R, and the mathematical expression is as follows: the fit for a single association is expressed as: Wherein r ui represents the true score of the user u on the object i, p u represents the hidden vector of the user u, represents the preference degree of the user on k hidden factors, q i represents the hidden vector of the object i, represents the attribution degree of the object on k hidden factors, k represents the dimension of the hidden factors, is a super parameter, and T represents transposition. The objective function is a minimization loss function: Wherein, the Is the square error term of the error,For regularization terms, where λ is a regularization coefficient for controlling the intensity of regularization, p u q i 2 are the L2 norms of vectors p u and q i, respectively, and the optimization process uses a random gradient descent algorithm for iterative update. Preferably, the update formula of the random gradient des