KR101947669B1 - System and method for determining hs code based on artificial intelligence learning - Google Patents

Abstract

The present invention relates to an artificial intelligent machine learning-based HS item which can automatically and accurately process the HS item classification, which is a trade customs specialty field, by artificial intelligence technology, so that customs experts and non-specialists can perform trade customs and HS item classification work Classification decision system and decision method. It is composed of system information about Section, Chapter, Heading and Sub-Heading which form HS system classification, A classification system DB storing linked inference rule information; A term information DB for storing a query keyword and an example product name for the item identification of the grade, wherein the example product name is a term information DB including synonyms and synonyms; A feature information DB for storing feature information on the use, functions, and materials of the classified items; A case information DB for storing case information about cases of HS classification of domestic and overseas products or cases of confirmation of unidentified product names; An input / output module for executing a query and a response procedure with a user; An inference module for performing HS item classification through a query according to information of the DBs; And a learning information module for performing query processing and processing according to the information of the DBs in order to resolve contradictions of the HS item classifications, to confirm unclear item names, and to verify the HS item classification results.

Description

TECHNICAL FIELD The present invention relates to a system and method for determining an HS item classification based on artificial intelligence machine learning,

The present invention relates to an artificial intelligent machine learning-based HS item which can automatically and accurately process the HS item classification, which is a trade customs specialty field, by artificial intelligence technology, so that customs experts and non-specialists can perform trade customs and HS item classification work A classification decision system and a decision method.

HS Classification is classified into one HS item classification code in accordance with the Harmonized Commodity Description and Coding System (HS) set by the World Customs Organization (WCO) The Contracting Parties shall carry out HS classification according to the principles set out in the HS System by the International Convention on the Harmonized Commodity Description and Coding System (HS Convention).

Since the customs duty on imported goods is determined in advance for each item classification code, the correct HS item classification must precede the customs value to be paid. Particularly, the functions are similar, but they are different in shape, they are similar in shape, but they have different functions, or different classification codes according to whether they are universal parts or exclusive parts. Therefore, in order to pay correct customs duties, Is important.

The item classification code is a classification code given by the HS Convention for export and import goods. Up to 6 digits are commonly used internationally. Up to 7 digits can be used for up to 10 digits within each country's 6 digits. In Korea, HSK (HS of Korea) is used up to 10 digits.

The structure of the HS Item Classification Code will be described as an example. In the 10 unit HSK of 0102.90-1000

* 01: Ryu (Chapter) -> An animal is classified as a kind, the first two digits.

* 0102: Heading -> Cattle are categorized as the first four digits.

* 0102.90: Subheading -> A subheading in which other cows are categorized.

* 0102.90-9010: A 10-digit code that categorizes "cows". In the case of less than 7 units, each country can use it in a subdivision within the range of 6 units. In Korea, the last 4 digits are subdivided into 10 units.

.

The process of finding an item classification code corresponding to a specific item is called an item classification, and it is practically difficult to classify a specific item in accordance with an item classification code having a hierarchical structure. The classification of HS items is basically based on the HS classification rules of the HS Convention established by the CCC (Customs Cooperation Council), but it is not determined by the type of the goods, but the main characteristics, ingredients, uses, It depends on several factors.

Meanwhile, the recent free trade agreement (FTA) has been concluded with more than 50 countries, and regulations on export and import of products are becoming more strict than ever before. In particular, the HS Classification Schedule established by the HS Convention is a legally binding regulatory item that sets tariffs on exports and imports and further determines the origin of the HS, and is more and more accurate for HS tariff classification due to tariff benefits under the FTA . In the case of export and import of domestic large enterprises and medium-sized companies, most HS categories are carried out by customs or trade specialists. However, SMEs or small business owners have difficulty in receiving support from specialists in terms of cost or local area. to be.

HS classification classified and determined by a person who is a customs specialist is divided into 21 parts, each section is classified into 97 subcategories, and subcategories are classified into 1,224 headings in total And the call consists of 5,205 sub-headings. In addition, each department, class, and issue contains a legally binding Note, which requires a professional giant's ability to understand, memorize, and find vast quantities of item classifications for accurate HS item classifications.

Professional customs brokers also need to remember vast quantities of legal texts to make accurate HS item classification decisions and to know details of changes and understand exactly the uses, functions, ingredients, materials and composition of materials such as accessories and parts of goods. In fact, it was not easy for a trade customs specialist to carry out accurate HS classification of goods for all sectors without making mistakes.

Prior Art Document 1. Patent Publication No. 10-2012-0067062 (published on Jun. 25, 2012)

Accordingly, the present invention was invented to solve the above-mentioned problems. In addition to experts such as customs brokers, it is possible to quickly and easily obtain information on articles even if the manufacturer, the worker, HS item classification system and decision method based on artificial intelligence machine learning which enables to classify and use HS items.

According to an aspect of the present invention,

Classification system that stores system information on Section, Head, and Sub-Heading grades that make up HS item classification, class name, and inference rule information linked by class DB;

A term information DB for storing a query keyword and an example product name for the item identification of the grade, wherein the example product name is a term information DB including synonyms and synonyms;

A feature information DB for storing feature information on the use, functions, and materials of the classified items;

A case information DB for storing case information about cases of HS classification of domestic and overseas products or cases of confirmation of unidentified product names;

An input / output module for executing a query and a response procedure with a user;

An inference module for performing HS item classification through a query according to information of the DBs; And

A learning information module for performing a query and processing according to the information of the DBs in order to resolve contradictions of the HS item classifications, to confirm unclear item names, and to verify the HS item classification results;

Which is based on artificial intelligence machine learning.

According to another aspect of the present invention,

A product name input step of inputting a product name according to a query output to the input / output module;

The reasoning module analyzes the input item name and searches for the class name including the term in the product name in the classification system DB first. If the retrieved class name is plural, the class name corresponding to the class closest to the final determination class is used as the inference point A product name confirmation step of determining a product name;

An HS item inferring step of performing an HS item item sorting procedure in such a way that the reasoning module deduces from the highest rank to the lowest rank determined in the item name checking step; And

In order to verify the HS item classification result, the learning information module searches at least one selected item name, purpose, function, and material of one or more classes selected from the classes corresponding to the HS item classification result and the upper class, An HS item verification step of completing verification through user's acknowledgment;

, Which is an artificial intelligent machine learning based HS item classification method.

The above-mentioned invention of the present invention enables a customs-related expert to carry out HS item classification without having knowledge on all fields such as plants, minerals, petrochemicals, textiles, metals, machines and transportation equipment, The manufacturer or the worker having the product can accurately classify the HS item for the manufactured article and the general person can easily perform the HS item classification only if the information on the subject article is provided.

In addition, since the present invention improves the result of classification of HS items by machine learning through artificial intelligence, more accurate and quick classification results can be obtained as the number of classification of HS items increases.

1 is a block diagram showing a configuration of a crystal system according to the present invention,
FIG. 2 is a flowchart sequentially showing determination methods for classifying HS items based on the determination system,
FIG. 3 is a view showing a first embodiment of the process of sorting HS items according to the HS speculation rule of the determination method according to the present invention,
FIG. 4 is a flowchart showing a second embodiment of classifying HS items according to the HS speculation rule of the determination method according to the present invention,
FIG. 5 is a flowchart showing a third embodiment of the HS item classification according to the HS inference rule of the determination method according to the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The above and other features and advantages of the present invention will become more apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings, There will be. The present invention is capable of various modifications and various forms, and specific embodiments are illustrated in the drawings and described in detail in the text. It is to be understood, however, that the invention is not intended to be limited to the particular forms disclosed, but on the contrary, is intended to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram showing the configuration of a crystal system according to the present invention.

The decision making system of the present embodiment includes system information for a section, a chapter, a heading, and a sub-heading that form a system of HS item classification, and inference rule information linked by a rating name and a rating And a term information DB 120 for storing the keyword and the name of the example product name for identifying the grade item, wherein the example product name includes a synonym word and an identical word, A case information DB 170 for storing case information such as cases of domestic and overseas HS item classification or cases of unconfirmed product name confirmation, An inference module 150 for searching for the DBs 110, 120, 130, and 170 and proceeding to HS item classification according to inference rules, Classification of the item or US And a learning information module 160 for searching the DBs 110, 120, 130, and 170 for the item classification of the confirmed item names.

The above-mentioned components are linked and executed so that the user can obtain accurate HS item classification results in response to input of item name and query contents. For reference, the determination system of the present embodiment may be constituted by a single program and may be installed in a general computer 200 or a mobile phone 200 ', or may be installed in a computer 200 or a mobile phone 200 ') may be connected online.

In order to more clearly illustrate the present decision system, the decision method of the present embodiment will be described below.

FIG. 2 is a flowchart sequentially illustrating decision methods for classifying HS items on the basis of the decision system. FIG. 3 is a flowchart illustrating a first procedure for classifying HS items according to the HS inference rules of the decision method according to the present invention. Fig.

S10; Step to enter name

The user executes the decision making system of this embodiment for HS item classification for a specific product name. The input / output module 140 outputs an input window for inputting the item name, and the user inputs the item name to be classified by the input window. The name of the product input by the user in this embodiment is 'Vehicle Hyundai YF Sonata' as shown in FIG.

S20; Item name confirmation step

The inference module 150 searches the classification system DB 110 and confirms the inference rule information. The reasoning module 150 of this embodiment analyzes the 'Vehicle Hyundai YF Sonata', which is the name of the vehicle, inputted by the user in accordance with the terms of the Provision 1 ', one of the inference rule information, The analysis result inference module 150 identifies the " vehicle "

The reasoning module 150 of the present embodiment that has confirmed the "vehicle" searches the classification system DB 110 to search for a class name including the "vehicle" identified in the name of the product. For reference, the reasoning module 150 of the present embodiment searches the classification system DB 110 to search for class names including " vehicles ", searches for 'passenger cars', 'vehicles', 'cars' In HS classification system, 'Heading' class which is closest to the final decision class and 'Class', 'Passenger car' which is the class name is designated as 'Heading' class.

The reasoning module 150 begins to infer the 'heading' class 'passenger car' as the starting point.

S30; HS item inference step

The reasoning module 150 may be used to refer to the 'Vehicle / Aircraft / Ship and Transportation Equipment Associated Products (Part 17)' to which a 'Passenger Car' having a 'Heading' , And confirms the "Chapter" and "Heading" grades leading to the "Passenger Car" with the "Heading" rating, followed by the "Sub-Heading" And finally determine HS classification.

More specifically, if the inference module 150 determines that a 'vehicle / aircraft / ship and a transportation device related article (Part 17)' is classified as a 'section' The inference rule information stored in the classification system DB 110 is searched to confirm whether or not the vehicle and its accessories and accessories of the ' For reference, the reasoning rule information stored in the classification system DB 110 includes an item criterion that is excluded from the corresponding grade for each grade. The reasoning module 150 confirms that the item is excluded from the 'chapter' And executes a query about the reference through the input / output module 140. The reasoning rule information is rewritten in the description related to the contention.

As a result of the inquiry, if it is confirmed that the item name inputted by the user is not excluded from the corresponding grade, the inference module 150 classifies 'Chapter' grade between 'Section' grade and 'Heading' grade 'Vehicle and its accessories and accessories (Chapter 87)' is determined, and in this implementation, the transportation car is checked incidentally.

Subsequently, the reasoning module 150 searches system information in the classification system DB 110 to determine a 'Sub-Heading' level, which is a lower class of 'Heading' A query is made to select one of the "spark ignition type piston internal combustion engine" and the "compression ignition type internal combustion engine", which is a grade name corresponding to the "Sub-Heading" grade. However, since the rating name is not easily understood by the general person, the inference module 150 searches the term information DB 120 for the note of the grade name, the query word corresponding to the grade name, and the query keyword corresponding thereto, You may query with a supplementary explanation while keeping the grade name. For reference, the "spark ignition type piston internal combustion engine" is a "gasoline engine" and the "compression ignition type piston internal combustion engine" is a "diesel engine".

When the user selects the gasoline engine in the above query, the reasoning module 150 determines the 'sub-heading' grade to be 'spark ignition type piston internal combustion engine (8703.23)'. Subsequently, the reasoning module 150 searches the lower grades of the 'Sub-Heading' class according to the system information, and selects 'cylinder capacity 1,500 cc to 3,000 cc (No. 8703.23)' And 'new car (No. 8703.23.1010)' in order.

The 'new car' (No. 8703.23.1010) confirmed above is the final decision of HSK (Harmonized System of Korea) which is the last 4 digits code added in Korea to the world standard HS Code, The HSK code of the YF Sonata shall be determined in accordance with No. 8703 (Results) No. 8703.23.1010.

S40; The first step in determining whether you meet

The inference module 150 queries the user for satisfaction of the determined HS item classification. If the HS item category is unsatisfied, the inference module 150 queries the incomplete reason through the input / output module 140 in the form of a single-item or a double-item.

However, if the user satisfies the HS item classification decision, proceed to the next step.

FIG. 4 is a diagram sequentially showing a second embodiment of classifying HS items according to the HS speculation rule of the determination method according to the present invention. FIG. 5 is a flowchart illustrating a method of classifying HS items according to the HS speculation rule of the determination method according to the present invention. FIG. 3 is a view showing a third embodiment of the present invention.

S90; Step to present incomplete information

The reasoning module 150 inputs the unsatisfactory information to the inquiry of the inference module 150, and the inference module 150 compares the incomplete information with the item name input step S10, the item name confirmation step S20, Step S30 is repeated.

For example, it can be informed that the HS item classification can not be determined in the HS item inference step S30 and can not be determined by the contention. Of course, the user expresses dissatisfaction in the first satisfaction check step (S40), and also inputs the unsatisfactory information.

An example of the reason why the HS item classification determination can not be performed is the HS Code contention of the HS item classification decision presented by the reasoning module 150. For example, as shown in FIGS. 4 and 5, when the user inputs 'Hyundai YF Sonata automatic transmission' as the name of the product in the item name input step S10, Confirms the product name in the item name confirmation step (S20). The reasoning module 150 confirms the term " automatic transmission " which can not be confirmed in the system information in the item name checking process. Therefore, the reasoning module 150 searches for terms similar or identical to the product names in the classification system DB 110, and inquires the user about the terms thus retrieved. Subsequently, the reasoning module 150 changes the " automatic transmission " to " gearbox " through the inquiry, and starts the inference in the HS item inference step S30.

However, the "gear box" determined by the product name is also registered in No. 8708.40 shown in FIG. 5 in addition to the item 8483.40 shown in FIG. 4, and a competition occurs in the HS item classification determination. The "gearbox" (result) specified in No. 8483.40 hereby applies the rules to the Article 16, "Machinery / electrical equipment and parts thereof", followed by "Machinery and their parts (Chapter 84)" and " / Gears / gearboxes / clutches etc. (heading 84.83) ', infer the results. In this process, the inference module 150 searches the classification system DB 110 to check the classification inference rule information, and generates inference rule information linked to the 'transmission / gear / gear box / clutch (heading 84.83) In rule 5 of B of the Note, "Excluded standard car gearboxes are classified in heading 87.08".

The reasoning module 150 outputs a query to the user through the input / output module 140 as to whether or not the gearbox is a gearbox for a vehicle, and the user inputs the name of the automatic transmission for the Hyundai YF sonatas, . Eventually, the speculation module 150 determines the HS item category for the name of the item, in accordance with No. 8708.40.

Another example of why HS item classification decisions can not be made is explained.

The reason why HS item classification decision can not be made is unclear about the products that are subject to HS item classification decision other than competition. For example, the target article is a new article that has not existed in the past and is not defined in the classification system DB 110, or the user lacks information on the target article.

In this case, in the item name confirmation step (S20), the learning information module 160 searches for the synonyms or synonyms of the product name inputted by the user in the term information DB 120, The user is asked to confirm the item name. However, if the item name is not confirmed or the user does not satisfy the inquiry in the HS item inference S30 and the first satisfaction check step S40, the learning information module 160 searches the feature information DB 130 ), And searches the case information DB 170 for domestic and foreign item classification cases or unconfirmed item name confirmation cases and compares them.

Also, in the HS item inference step S30, the learning information module 160 inquires about the use and the function, as shown in FIG. 4, The learning information module 160 finds the " hybrid integrated circuit " of No. 8542.32 through inquiry about the use and function, and then, based on the classification rule of the general rule No. 4, Can be determined.

The item name confirmation step S20 determines the item name according to the procedure described above. The HS item item reasoning step S30 determines the HS item classification according to the procedure described above. The learning information module 160 determines the item name The confirmation procedure and the HS item classification determination procedure are stored in the case information DB 170 as case information.

Through the iterative inference, the first satisfaction check step (S40) confirms the satisfaction of the user and proceeds with the subsequent procedure.

S60; Verification Step Verification

The inference module 150 inquires of the user whether or not the HS item classification determination result is verified and requests the verification to proceed with the subsequent procedure. When the verification is abandoned, the inference module 150 concludes the HS item item classification determination result.

S70; HS Item Verification Phase

The user requesting the verification in the verification step (S60) is a case of unsatisfactory or suspicious result of the determination of the HS item classification, which may be a reason of the contention or uncertainty of the HS item classification determination result.

In this case, the learning information module 160 searches the feature information DB 130 to check feature information about the use, function, and material of the grade to be dissatisfied and its surrounding classes, Query. For example, as shown in FIG. 4, when the reasoning module 150 determines the classification of 'gear / gear box other transmission (No. 8483.40)' and 'transmission gear / gear , The feature information on each item of the 'Gearbox / clutch (8483)' rating, and the reasoning rule information of the above grades, and if the 'gearbox' Or not.

If the user responds that the name of the entered item name is for automobile, the HS item classification determination result is a misjudgment, so the learning information module 160 proceeds HS item classification again according to the inference rule information.

On the other hand, the learning information module 160 stores the item name, the HS item classification determination procedure and the feature information about the feature information, the response of the user corresponding thereto, and the re-inferred HS item classification determination result to the feature information DB 130 and the case information DB 170 as well as the reasoning rule information of the classification system DB 110 are added and updated. When the HS item verification procedure is performed under the same conditions, the information is searched and utilized.

S80; Steps to check for secondary fulfillment

The learning information module 160 inquires of the user about the re-determination or verification of the HS item classification, and when it is unsatisfactory, it inquires the reason of non-satisfaction through the input / output module 140 in the form of a multiple choice or a reply, The unsatisfactory information presentation step S50 can be continued.

However, if the user satisfies the HS item classification decision, proceed to the next step.

S90; HS item classification decision step

The inference module 150 outputs the HS item classification determination result satisfied by the user through the input / output module, and can output additional contents such as the applied tariff rate.

While the present invention has been described in connection with what is presently considered to be practical exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (8)

Classification system that stores system information on Section, Head, and Sub-Heading grades that make up HS item classification, class name, and inference rule information linked by class DB;
A term information DB for storing a query keyword and an example product name for the item identification of the grade, wherein the example product name is a term information DB including synonyms and synonyms;
A feature information DB for storing feature information on the use, functions, and materials of the classified items;
A case information DB for storing case information about cases of HS classification of domestic and overseas products or cases of confirmation of unidentified product names;
An input / output module for executing a query and a response procedure with a user;
The HS item classification is performed through the inquiry according to the information of the DBs, the item name inputted by the user is analyzed to search one or more class names in the classification system DB including the terms in the product name, The class name corresponding to the class closest to the final decision class is determined as the inference time point and the keyword of the product name entered by the user or the query keyword corresponding thereto and the corresponding query keyword are retrieved from the term information DB, An inference module for querying or querying for words that the user can recognize; And
A learning information module for performing a query and processing a query according to information of the DBs for at least one selected from the conflict item elimination of the HS item classification, the uncertain item name confirmation, and the verification of the HS item classification result;
Wherein the artificial intelligence machine learning based HS item classification determination system comprises: The method according to claim 1,
Wherein the reasoning module deduces HS items from the highest rank to the lowest rank in the rating system of the HS commodity classification based on the name of the commodity inputted by the user and proceeds the HS commodity classification procedure to the name of the commodity;
Wherein the learning information module verifies HS item classification results in ascending order from a lower class to a higher class in the course of verification of the HS item classification result;
Which is based on artificial intelligence machine learning. The method according to claim 1,
Wherein the learning information module stores the confirmation procedure performed in the unclear item name confirmation process as case information in the case information DB;
Which is based on artificial intelligence machine learning. delete delete The method according to claim 1,
The reasoning module searches for inference rule information of a class in the classification system DB to resolve conflicts generated in the process of classifying HS item items and, through inquiry and response confirmation based on the reasoning rule information, Determining one of the HS item classes as the final result;
Which is based on artificial intelligence machine learning. The method according to claim 1,
Wherein the learning information module searches the at least one selected item among the grade name of the grade in which the contention occurred and the item use, function, and material of the upper grade grade, and inquires the user for the contention eliminated in the verification progress of the HS item classification result, Determining one of the competing HS item categories as a final result through user acknowledgment;
Which is based on artificial intelligence machine learning. The reasoning module analyzes the item name inputted through the input / output module and searches the classification system DB for the class name including the term in the name of item. If the class names are within the designated grade range, The class name corresponding to the class name is determined as the inference point and the class name synonym or the corresponding keyword corresponding to the class name presented to the user in the query process is searched in the term information DB to search the class name, To be included in the query;
An inferencing module for inferring an HS item item classification procedure from a higher rank to a lower rank of a grade determined in the item name confirmation step; And
In order to verify the HS item classification result, the learning information module searches at least one selected item name, purpose, function, and material of one or more classes selected from the classes corresponding to the HS item classification result and the upper class, An HS item validation step of verifying the user's response through verification;
The method comprising the steps of: determining an HS item classification based on artificial intelligence machine learning.

Images