A customs tax collection risk screening method based on a rule engine

Abstract

The application belongs to the technical field of data processing in the aspect of customs tax revenue risk prevention and control, and proposes a customs tax revenue risk identification method based on a rule engine. The application obtains high-quality customs basic data by data cleaning and format conversion on customs internal data and customs external data; obtains the type label of each standard customs data by field mapping of the standard customs data through a standard customs tax revenue risk data model, and eliminates data understanding deviation caused by expression difference; and performs customs tax revenue risk identification on the customs declaration form through the rule engine and the complete customs tax revenue data entity, so that the accuracy of the obtained customs tax revenue risk identification result is higher. Data comparison on the three risk dimensions of the customs declaration form, i.e. price risk dimension, classification risk dimension and origin risk dimension, realizes comprehensive identification of the customs tax revenue risk existing in the customs declaration form.

Description

Technical Field

[0001] This invention relates to the field of data processing technology for customs tax risk prevention and control, and in particular to a method for identifying customs tax risks based on a rule engine. Background Technology

[0002] Customs data processing systems involve a vast amount of complex data processing tasks. Among these, data processing related to customs tax risks refers to whether the declared information data in the import and export goods clearance process is consistent with the actual customs tax collection and management regulations. Customs tax risks may lead to the loss of customs duties, import value-added tax, consumption tax, and other taxes, or to omissions and errors in the customs tax collection and management process, posing a serious challenge to the rigor and effectiveness of customs tax collection and management. Therefore, only by accurately identifying customs tax risks can we proactively plug loopholes in customs tax revenue loss and ensure that all taxes due are collected.

[0003] The current mainstream methods for identifying customs tax risks rely solely on customs declaration data. This means that customs tax risk identification is based on a single source of data, ignoring the diversity and interrelationship of factors that cause customs tax risks. Consequently, it is impossible to accurately and comprehensively identify all existing customs tax risks. Summary of the Invention

[0004] To address the problem that existing technologies rely on single-source data for customs tax risk identification, resulting in low accuracy and an inability to comprehensively identify all existing customs tax risks, a rule engine-based method for customs tax risk identification is proposed, which solves the above problems.

[0005] The technical solution of this invention is as follows:

[0006] A method for identifying customs tax risks based on a rules engine, comprising the following specific steps:

[0007] S1: Perform data cleaning and format conversion on each piece of data in the internal and external customs data to obtain all the standard customs data corresponding to each piece of data; map the fields of each standard customs data to the preset customs tax risk standard data model to obtain the business type label of each standard customs data; save all the standard customs data and corresponding type labels corresponding to each piece of data into the customs database in the form of customs-related data groups.

[0008] S2: Extract the key related field set from the customs declaration form, and perform matching and retrieval in the customs database based on the key related field set to obtain the complete customs tax data entity of the customs declaration form; the key related field set includes: customs declaration number, commodity code, consignor / consignee code, and specifications;

[0009] S3: By comparing the customs declaration with data from multiple risk dimensions using the rule engine and the complete customs tax data entity, customs tax risk identification results for the customs declaration in each risk dimension are obtained; the multiple risk dimensions include: price risk dimension, classification risk dimension and country of origin risk dimension.

[0010] S4: Generate structured customs tax risk warning information based on the customs tax risk identification results of each risk dimension.

[0011] Preferably, the internal customs data refers to data generated or collected by the customs business system and data stored in the customs data resource pool; the external customs data refers to data provided by third-party institutions or public channels.

[0012] Preferably, the data cleaning method described in S1 is as follows: deduplication, null value filling, and format standardization processing are performed on internal customs data and external customs data to obtain preprocessed customs data; the format standardization processing includes: unifying character and encoding standards, unifying the expression format of dates and times, unifying numerical values, currencies, and units of measurement, and unifying data type constraints.

[0013] Preferably, the format conversion method described in S1 is as follows:

[0014] A1: Convert the structured data in the preprocessed customs data into a unified standard format to obtain the standard customs data, and save it; the unified standard format is: [field name, field type, field value];

[0015] A2: Using natural language processing technology, text extraction and information parsing are performed on the unstructured data in the preprocessed customs data to obtain customs tax information; the customs tax information is then converted into a unified standard format to obtain the standard customs data, which is then saved.

[0016] Preferably, the method for determining the business type label of the standard customs data in S1 is as follows:

[0017] B1: Construct a standard data model for customs tax risk; the standard data model for customs tax risk includes six categories of standard fields: entity information, commodity information, price information, country of origin information, logistics information, and enterprise qualification information;

[0018] B2: Determine the meaning of the field names in each standard customs data and the standard fields in the standard field set according to the customs business specifications and industry terminology definitions; map and bind the field names in each standard customs data and the standard fields in the standard field set according to the meaning of the field names in the standard customs data and the standard fields in the standard field set to obtain the business type labels of each standard customs data.

[0019] Preferably, the data types included in the customs-related data group in S1 are: commodity codes and full-dimensional customs-related data of the commodity codes; the full-dimensional customs-related data includes: commodity price, certificate of origin, material specifications, and industry classification.

[0020] Preferably, S2 specifically includes:

[0021] C1: Extract the customs declaration number, commodity code, consignor / consignee code, and specifications from the customs declaration form; use the customs declaration number, commodity code, consignor / consignee code, and specifications as matching indices, and perform equivalent matching and fuzzy search with each standard customs data in the customs database to obtain the matching data corresponding to each matching index;

[0022] C2: Construct the complete customs tax data entity of the customs declaration form using the matching data and the corresponding customs-related data group in the customs database.

[0023] Preferably, the method for identifying customs tax risks by comparing multi-risk dimension data of the customs declaration form through a rule engine and the complete customs tax data entity, as described in S3, is as follows:

[0024] D1: Business personnel compile customs tax risk rules in different dimensions based on expert experience and knowledge, and build a customs tax risk rule library based on these different dimensions; the different dimensions include: price risk dimension, classification risk dimension, and country of origin risk dimension; the customs tax risk rules are specific data comparison logic;

[0025] D2: Compare and match all fields of the customs declaration form with all customs tax risk rules in the customs tax risk rule base one by one to obtain the customs tax risk rule set corresponding to the customs declaration form;

[0026] D3: Extract standard customs data from the complete customs tax data entity of the customs declaration based on the business type tag. The standard customs data is required by each customs tax risk rule in the customs tax risk rule set. Based on each customs tax risk rule and the required standard customs data, perform customs tax risk identification on the customs declaration by comparing data from multiple risk dimensions, including price risk dimension, classification risk dimension, and country of origin risk dimension, and obtain the customs tax risk identification results for each risk dimension.

[0027] Preferably, the customs tax risk rule base classifies all customs tax risk rules by commodity type. Each commodity category includes customs tax risk rules corresponding to price risk dimension, classification risk dimension, and country of origin risk dimension. That is, each commodity category includes corresponding price comparison rules, classification comparison rules, and country of origin comparison rules. The customs tax risk rules include: rule name, usage scenario, triggering conditions, calculation logic (i.e., comparison logic), and handling suggestions.

[0028] Preferably, the entire process of information regarding the formulation, testing, approval, activation, operation monitoring, iterative updates, and repeal of the customs tax risk rules should be uniformly archived.

[0029] Beneficial effects:

[0030] This invention belongs to the field of data processing technology in customs tax risk prevention and control, and proposes a customs tax risk identification method based on a rule engine. This invention does not only target single-source data such as customs declarations, but also utilizes multi-source data from both internal and external customs data, providing a data source guarantee for enhancing the accuracy of customs tax risk identification results. By cleaning the internal and external customs data and removing invalid and abnormal data, the quality of basic customs data can be guaranteed. Furthermore, by converting the formats of the internal and external customs data, structured and unstructured data from both sources are unified into a standard customs data format, effectively improving the efficiency and accuracy of subsequent data matching and retrieval. By mapping fields of the standard customs data using a customs tax risk standard data model, type labels are obtained for each standard customs data set. This unifies the descriptions of standard fields with the same meaning but different expressions across different standard customs data sets, thus eliminating data comprehension biases caused by differences in expression in external data. It ensures interoperability and comparison between standard customs data from different sources, making the customs data related to customs declarations in the complete customs tax data entity obtained through matching retrieval more comprehensive and complete. This provides more reference information for subsequent multi-dimensional data comparisons of customs declarations using rule engines and complete customs tax data entities, significantly improving the accuracy of customs tax risk identification results and achieving high-precision customs tax risk identification. In customs tax risk, price is a direct factor affecting customs tax revenue; discrepancies between the declared commodity code and the actual commodity attributes are indirect factors; and falsification or misrepresentation of the country of origin are also indirect factors. Therefore, by comparing data across three risk dimensions—price risk, classification risk, and country of origin—customs declarations, a comprehensive identification of customs tax risks can be achieved. Attached Figure Description

[0031] Figure 1This is a flowchart of a customs tax risk identification method based on a rules engine.

[0032] Figure 2 This is a flowchart illustrating the entire lifecycle management process of customs tax risk rules. Detailed Implementation

[0033] The following description is based on specific embodiments and accompanying drawings.

[0034] like Figure 1 As shown, this invention provides a customs tax risk identification method based on a rule engine, the specific steps of which include:

[0035] S1: Perform data cleaning and format conversion on each piece of data in the internal and external customs data to obtain all the standard customs data corresponding to each piece of data; map the fields of each standard customs data to the preset customs tax risk standard data model to obtain the business type label of each standard customs data; save all the standard customs data and corresponding type labels corresponding to each piece of data into the customs database in the form of customs-related data groups.

[0036] S2: Extract the key related field set from the customs declaration form, and perform matching and retrieval in the customs database based on the key related field set to obtain the complete customs tax data entity of the customs declaration form; the key related field set includes: customs declaration number, commodity code, consignor / consignee code, and specifications;

[0037] S3: By comparing the customs declaration with data from multiple risk dimensions using the rule engine and the complete customs tax data entity, customs tax risk identification results for the customs declaration in each risk dimension are obtained; the multiple risk dimensions include: price risk dimension, classification risk dimension and country of origin risk dimension.

[0038] S4: Generate structured customs tax risk warning information based on the customs tax risk identification results of each risk dimension.

[0039] Preferably, the internal customs data refers to data generated or collected by the customs business system and data stored in the customs data resource pool; the external customs data refers to data provided by third-party institutions or public channels.

[0040] Specifically, internal customs data includes, but is not limited to: customs declaration data (including commodity name, specifications, declared price, quantity, country of origin, country of trade, etc.), tax data (including actual tax paid, dutiable value, tax rate application records, etc.), inspection data (including inspection results, inspection instructions, inspection methods, etc.), commodity barcode data, historical document review conclusion data, and enterprise historical declaration behavior data.

[0041] Customs external data includes but is not limited to: Internet commodity price quotation data (such as the prices of commodities of the same specification on e-commerce platforms), reference price data (such as industry benchmark prices and international market price indices), ship trajectory and demurrage data, price quotation data of bulk commodities such as non-ferrous metals, import and export medical device registration data, enterprise qualification and credit data (such as industrial and commercial registration information and import and export qualification certificates), logistics and transportation data (such as bill of lading information and route information), and trade quotation data.

[0042] Preferably, the data cleaning method is: performing deduplication, null value filling and format standardization processing on customs internal data and customs external data to obtain preprocessed customs data; the format standardization processing includes: unifying character and coding standards, unifying the expression format of date and time, unifying numerical values, currencies and measurement units, and unifying data type constraints.

[0043] Specifically, the unifying character and coding standards refers to:

[0044] Unify all data into UTF-8 encoding, clean full-width numbers / symbols, redundant spaces, line breaks, and invalid special characters, and eliminate garbled characters in enterprise declarations and external quotation data.

[0045] Example: Change the full-width number in "1000.5 Yuan" to the half-width standard format of "1000.5".

[0046] The unifying the expression format of date and time refers to:

[0047] All dates are forced to adopt the unified YYYY-MM-DD format, and the storage rules of hours, minutes and seconds are unified; compatible with various original writing formats (2026 / 6 / 26, 20260626, June 2026 of the Chinese lunar calendar) and perform batch conversion.

[0048] The unifying numerical values, currencies and measurement units refers to:

[0049] The price field is uniformly reserved with 2 decimal places; the currency identifiers are unified (Renminbi CNY, US dollar USD); weight and quantity are uniformly converted to statutory units such as kilogram and piece, to eliminate the mixing of ten thousand, thousand, kg and kilogram; redundant characters such as thousandth separators and currency symbols are removed.

[0050] Example: The external price "1,200USD" is standardized to the numerical value 1200.

[0051] The unifying data type constraints refers to:

[0052] Prices and quantities are forcibly converted to numeric types, dates are uniformly set to time types, and commodity codes and customs declaration numbers are uniformly set to fixed-length character types, so as to avoid conflicts caused by storing numbers as text or mixing text into numeric values.

[0053] Preferably, the format conversion method is as follows:

[0054] A1: Convert the structured data in the preprocessed customs data into a unified standard format to obtain standard customs data, and save it; the unified standard format is: [field name, field type, field value]; the structured fields include database tables, standard reports, etc.

[0055] Specifically, the unified standard format adopts a structured format of "field name + field type + field value", for example: commodity code (character type): 85171200; declared unit price (numeric type): 256.80; country of origin (character type): Germany.

[0056] A2: Using natural language processing technology, text extraction and information parsing are performed on the unstructured data in the preprocessed customs data to obtain customs tax information; the customs tax information is converted into a unified standard format to obtain standard customs data, which is then saved; the unstructured data includes: web page text, logistics document image recognition text, contract documents, etc.

[0057] Preferably, the method for determining the business type label of the standard customs data is as follows:

[0058] B1: Construct a standard data model for customs tax risk; the standard data model for customs tax risk includes six categories of standard fields: subject information, commodity information, price information, country of origin information, logistics information, and enterprise qualification information, covering all data items required for customs duty verification.

[0059] B2: Determine the meaning of the field names in each standard customs data and the standard fields in the standard field set according to the customs business specifications and industry terminology definitions; map and bind the field names in each standard customs data and the standard fields in the standard field set according to the meaning of the field names in the standard customs data and the standard fields in the standard field set to obtain the business type labels of each standard customs data.

[0060] Specifically, the standard field bound to the field name in each standard customs data is the business type label of each standard customs data.

[0061] Determining the business type labels for standard customs data enables semantic alignment. Regardless of the naming or description methods used by the original data source, information with the same business meaning is ultimately grouped under the same standard field, eliminating data comprehension biases caused by differences in field names and descriptions, and ensuring interoperability and comparison between different data sources.

[0062] After data cleaning, format conversion, and the aforementioned field mapping processes, the original internal and external data sources each generate independent standardized structured datasets. To address the issue that relying solely on key field equivalence / fuzzy matching cannot automatically attach supporting data such as prices and origins, basic data associations are pre-defined during the field mapping stage: using commodity codes as the basic link, the binding and association between commodity codes and corresponding commodity price trends, origin documentation, material specifications, industry classification standards, and other supporting data are pre-established, thus constructing a customs-related data group. The customs-related data group ensures that a single standardized structured record is no longer an isolated set of fields; each commodity code can be pre-attached with its supporting full-dimensional attribute data, providing the underlying association foundation for aggregating complete data entities in subsequent feature association stages.

[0063] Specifically, the customs-related data group includes: commodity codes and full-dimensional customs-related data of the commodity codes; the full-dimensional customs-related data includes, but is not limited to: commodity prices, country of origin documentation, material specifications, and industry classification.

[0064] Preferably, S2 specifically includes:

[0065] C1: Extract the customs declaration number, commodity code, consignor / consignee code, and specifications from the customs declaration form; use the customs declaration number, commodity code, consignor / consignee code, and specifications as matching indexes, and perform equivalent matching and fuzzy retrieval with each standard customs data in the customs database to obtain the matching data corresponding to each matching index; the equivalent matching and fuzzy retrieval realize the aggregation of internal and external data corresponding to the same declaration business, the same commodity, and the same business entity.

[0066] Specifically, the reasons for choosing customs declaration number, commodity code, consignor / consignee code, and specifications are as follows: the customs declaration number is the unique identifier for a single customs declaration transaction; the commodity code is the legal classification code for import and export commodities; the consignor / consignee code is the unique number of the business entity; and the specifications are used to accurately distinguish different categories of commodities under the same code. These four fields are unique, universal, and strongly related to business, and are the core identifiers that connect the entire data chain.

[0067] C2: Construct the complete customs tax data entity of the customs declaration form using the matching data and the corresponding customs-related data group in the customs database.

[0068] Specifically, a concrete implementation of the feature association is as follows: taking a customs declaration as the core entity, associating the internal tax collection records and inspection records corresponding to the document, and matching external price data of goods of the same specifications, qualification data of corresponding consignor and consignee enterprises, and logistics bill of lading data of the goods, to form a complete data entity with the customs declaration as the core.

[0069] Preferably, the method for identifying customs tax risks by comparing multi-risk dimension data of the customs declaration through a rule engine and the complete customs tax data entity is as follows:

[0070] D1: Business personnel compile customs tax risk rules in different dimensions based on expert experience and knowledge, and build a customs tax risk rule library based on these different dimensions; the different dimensions include: price risk dimension, classification risk dimension, and country of origin risk dimension; the customs tax risk rules are specific data comparison logic;

[0071] D2: Compare and match all fields of the customs declaration form with all customs tax risk rules in the customs tax risk rule base one by one to obtain the customs tax risk rule set corresponding to the customs declaration form;

[0072] D3: Extract standard customs data from the complete customs tax data entity of the customs declaration based on the business type tag. The standard customs data is required by each customs tax risk rule in the customs tax risk rule set. Based on each customs tax risk rule and the required standard customs data, perform customs tax risk identification on the customs declaration by comparing data from multiple risk dimensions, including price risk dimension, classification risk dimension, and country of origin risk dimension, and obtain the customs tax risk identification results for each risk dimension.

[0073] Specifically, the price risk dimension involves comparing the declared price in the customs declaration with external market price data, reference price data, and industry benchmark price data for the same specifications of goods during the same period. Combined with fields such as quantity, currency, and transaction method, it identifies risks such as underreporting, overreporting, and omission of elements of the dutiable value.

[0074] The classification risk dimension involves verifying the consistency between the declared commodity code and internal fields such as commodity name, specifications, material, use, and functional characteristics, and comparing it with external commodity classification databases and industry standard classification information to identify risks such as classification errors, incorrect tariff code entries, and concealment of specifications.

[0075] The origin risk dimension involves comparing the declared origin with qualification certificate data, trade route data, port of shipment information, and origin rules data under preferential trade agreements to identify risks such as false origin and non-compliance with origin rules.

[0076] Specifically, after performing the comparisons across the above dimensions, an example of automatically generating structured risk warning information is as follows:

[0077] 1) Compare price risk dimensions to obtain risk warning information for each price risk dimension:

[0078] Calculate the difference between the declared data and the reference data in each customs declaration, and compare it with the threshold of this dimension to determine whether there is any risk.

[0079] Example (price dimension): The external reference unit price of a certain medical device is 1,000 yuan, and the unit price declared by the enterprise is 600 yuan. The system risk rule presets a lower limit threshold of 70% of the reference price. The declared price is lower than the threshold, so it is determined that there is a price risk.

[0080] 2) Compare the risk dimensions of country of origin to obtain risk warning information for each dimension:

[0081] The consistency between the declared country of origin and the certificate of origin, shipping route, port of departure, and other reference data in the customs declaration is compared one by one, and the risk is judged by comparing the country of origin information with the threshold.

[0082] Example (origin dimension): The company declares that the country of origin of the goods is South Korea, but the port of shipment on the bill of lading provided is Vietnam, and there is no China-South Korea preferential certificate of origin. The system's preset origin information matching threshold is 100%. The information cannot be completely matched, and it is determined that there is a risk of false origin.

[0083] 3) Compare the risk categories to obtain risk warning information for each category:

[0084] Each customs declaration form is checked for its matching degree with the reference information such as commodity code, commodity name, specifications, material, and function. The classification deviation threshold is used to determine whether there is any risk.

[0085] Example (Classification Dimension): A company declared commodity code 9018 (medical devices), stating that the material of the product is ordinary plastic and it only has a simple storage function. However, the product corresponding to this tariff code needs to have professional diagnosis and treatment functions. The system's preset classification feature matching threshold is 80%. The matching degree does not meet the standard, and it is determined that there is a risk of classification error.

[0086] Preferably, the customs tax risk rule base classifies all customs tax risk rules by commodity type. Each commodity category includes customs tax risk rules corresponding to price risk dimension, classification risk dimension, and country of origin risk dimension. That is, each commodity category includes corresponding price comparison rules, classification comparison rules, and country of origin comparison rules. The customs tax risk rules include: rule name, usage scenario, triggering conditions, calculation logic (i.e., comparison logic), and handling suggestions.

[0087] Preferably, the entire process of information regarding the formulation, testing, approval, activation, operation monitoring, iterative updates, and repeal of the customs tax risk rules should be uniformly archived.

[0088] Specifically, such as Figure 2 As shown, the steps for implementing full lifecycle management of customs tax risk rules include:

[0089] Step S201 Rule creation and configuration

[0090] Business personnel develop risk rules based on an expert experience knowledge base. Specifically:

[0091] (1) Basis for rule formulation:

[0092] Expert Experience Knowledge Base: The risk identification experience accumulated by customs document review experts, inspection experts, classification experts, and pricing experts over a long period of work is structured and organized into configurable expert rule templates.

[0093] Business personnel add or modify risk rules in the system based on the expert knowledge base. Tables 1 and 2 show examples of customs tax risk rules, each of which requires setting the rule name, applicable scenarios, triggering conditions, calculation logic, and handling suggestions. The system automatically assigns a unique version number to each rule and saves a snapshot of the current rule configuration.

[0094] Table 1. Example 1 of Customs Tax Risk Rules

[0095]

[0096] Table 2 Example 2 of Customs Tax Risk Rules

[0097]

[0098] Step S202 Rule Testing and Approval

[0099] 1) Rule Testing

[0100] After the rules are configured, the system supports batch regression testing using historical customs declaration data and outputs indicators such as rule hit rate, accuracy rate, and false positive rate.

[0101] 2) Rule approval; after the test is passed, the rule is submitted for approval, and the approval record is automatically saved (i.e., Figure 2 (Retained approval records in the system).

[0102] Step S203: Rule Activation and Operation Monitoring

[0103] 1) Enabling the rules

[0104] The approved rules have been officially launched and put into use.

[0105] 2) Rule execution monitoring

[0106] The system monitors the rule execution status, hit count, and handling results in real time; rule execution logs, hit logs, and exception logs are all stored in a structured manner (i.e., Figure 2 The structured storage logs can be queried at any time.

[0107] Step S204: Rule Iteration / Deprecation / Archiving

[0108] 1) Rule iteration

[0109] When rules need to be modified, the system generates a new version of the rules (i.e. Figure 2 (New version iterations) while retaining historical versions.

[0110] 2) Repeal of the rules

[0111] When a rule is no longer applicable (e.g., the rule has expired or the rule has reached its hit limit), the rule will be deprecated.

[0112] 3) Archiving of rules

[0113] All configurations, logs, and version information are archived uniformly to achieve full traceability, auditability, and backtrackability.

[0114] It should be noted that the specific embodiments described above enable those skilled in the art to more fully understand the present invention, but do not limit the present invention in any way. Therefore, although the present invention has been described in detail with reference to the accompanying drawings and embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the present invention. In short, all technical solutions and improvements that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the present invention patent.

Claims

1. A customs tax risk identification method based on a rule engine, characterized in that, S1: Perform data cleaning and format conversion on each piece of data in the internal and external customs data to obtain all the standard customs data corresponding to each piece of data; map the fields of each standard customs data to the preset customs tax risk standard data model to obtain the business type label of each standard customs data; save all the standard customs data and corresponding type labels corresponding to each piece of data into the customs database in the form of customs-related data groups. S2: Extract the key related field set from the customs declaration form, and perform matching and retrieval in the customs database based on the key related field set to obtain the complete customs tax data entity of the customs declaration form; the key related field set includes: customs declaration number, commodity code, consignor / consignee code, and specifications; S3: By comparing the customs declaration with data from multiple risk dimensions using the rule engine and the complete customs tax data entity, customs tax risk identification results for the customs declaration in each risk dimension are obtained; the multiple risk dimensions include: price risk dimension, classification risk dimension and country of origin risk dimension. S4: Generate structured customs tax risk warning information based on the customs tax risk identification results of each risk dimension.

2. A customs tax risk identification method based on a rule engine as described in claim 1, characterized in that, The internal customs data refers to data generated or collected by the customs business system and data stored in the customs data resource pool; the external customs data refers to data provided by third-party institutions or public channels.

3. A customs tax risk identification method based on a rule engine as described in claim 2, characterized in that, The data cleaning method described in S1 is as follows: deduplication, null value filling, and format standardization are performed on internal and external customs data to obtain preprocessed customs data; the format standardization process includes: unifying character and encoding standards, unifying the expression format of dates and times, unifying numerical values, currencies, and units of measurement, and unifying data type constraints.

4. A customs tax risk identification method based on a rule engine as described in claim 3, characterized in that, The format conversion method described in S1 is as follows: A1: Convert the structured data in the preprocessed customs data into a unified standard format to obtain the standard customs data, and save it; the unified standard format is: [field name, field type, field value]; A2: Using natural language processing technology, text extraction and information parsing are performed on the unstructured data in the preprocessed customs data to obtain customs tax information; the customs tax information is then converted into a unified standard format to obtain the standard customs data, which is then saved.

5. A customs tax risk identification method based on a rule engine as described in claim 1, characterized in that, The method for determining the business type label of the standard customs data mentioned in S1 is as follows: B1: Construct a standard data model for customs tax risk; the standard data model for customs tax risk includes six categories of standard fields: entity information, commodity information, price information, country of origin information, logistics information, and enterprise qualification information; B2: Determine the meaning of the field names in each standard customs data and the standard fields in the standard field set according to the customs business specifications and industry terminology definitions; map and bind the field names in each standard customs data and the standard fields in the standard field set according to the meaning of the field names in the standard customs data and the standard fields in the standard field set to obtain the business type labels of each standard customs data.

6. A customs tax risk identification method based on a rule engine as described in claim 1, characterized in that, The data types included in the customs-related data group described in S1 are: commodity codes and full-dimensional customs-related data of the commodity codes; the full-dimensional customs-related data includes: commodity price, certificate of origin, material specifications, and industry classification.

7. A customs tax risk identification method based on a rule engine as described in claim 1, characterized in that, Specifically, S2 is: C1: Extract the customs declaration number, commodity code, consignor / consignee code, and specifications from the customs declaration form; use the customs declaration number, commodity code, consignor / consignee code, and specifications as matching indices, and perform equivalent matching and fuzzy search with each standard customs data in the customs database to obtain the matching data corresponding to each matching index; C2: Construct the complete customs tax data entity of the customs declaration form using the matching data and the corresponding customs-related data group in the customs database.

8. A customs tax risk identification method based on a rule engine as described in claim 1, characterized in that, The method for identifying customs tax risks by comparing multi-risk dimension data of the customs declaration form using a rule engine and the complete customs tax data entity, as described in S3, is as follows: D1: Business personnel compile customs tax risk rules in different dimensions based on expert experience and knowledge, and build a customs tax risk rule library based on these different dimensions; the different dimensions include: price risk dimension, classification risk dimension, and country of origin risk dimension; the customs tax risk rules are specific data comparison logic; D2: Compare and match all fields of the customs declaration form with all customs tax risk rules in the customs tax risk rule base one by one to obtain the customs tax risk rule set corresponding to the customs declaration form; D3: Extract standard customs data from the complete customs tax data entity of the customs declaration based on the business type tag. The standard customs data is required by each customs tax risk rule in the customs tax risk rule set. Based on each customs tax risk rule and the required standard customs data, perform customs tax risk identification on the customs declaration by comparing data from multiple risk dimensions, including price risk dimension, classification risk dimension, and country of origin risk dimension, and obtain the customs tax risk identification results for each risk dimension.

9. A customs tax risk identification method based on a rule engine as described in claim 8, characterized in that, The customs tax risk rule base classifies all customs tax risk rules by commodity type. Each commodity category includes customs tax risk rules corresponding to price risk dimension, classification risk dimension, and country of origin risk dimension. That is, each commodity category includes corresponding price comparison rules, classification comparison rules, and country of origin comparison rules. The customs tax risk rules include: rule name, usage scenario, triggering conditions, calculation logic (i.e., comparison logic), and handling suggestions.

10. A customs tax risk identification method based on a rule engine as described in claim 9, characterized in that, Information regarding the entire process of drafting, testing, approval, activation, operation monitoring, iterative updates, and repeal of the aforementioned customs tax risk rules must be uniformly archived.

Images