A method, device and electronic device for recommending commodity codes for import and export commodities

Abstract

The present invention provides a method, device, and electronic device for recommending commodity codes of imported and exported goods, and relates to the technical field of commodity classification, including obtaining the commodity information to be classified of the user; inputting the commodity information to be classified into a historical declaration data model to obtain a first recommended code of several commodity information to be classified; inputting the commodity information to be classified into a professional declaration data model to obtain a second recommended code of several commodity information to be classified; determining the code recommendation scores of several target recommended codes based on a score weighted rule; when there is a code recommendation score exceeding a threshold, arranging all target recommended codes whose code recommendation scores exceed the threshold in positive order; and recommending the first N (N≥1) target recommended codes of the recommended scores arranged in positive order to the user. This application lowers the threshold for declaring commodity codes for imported and exported goods, improves the accuracy of declaring commodity codes for imported and exported goods, and thus also reduces the burden on enterprises.

Description

Technical Field

[0001] The present invention relates to the technical field of commodity classification, and in particular to a method, device and electronic equipment for recommending commodity codes for import and export commodities. Background Art

[0002] Commodity code (also known as HS code), the commodity code currently used in my country, has a total of 10 digits, of which the first 8 digits are called the main code and the last two digits are called the additional code. Commodity code is a common economic language in international trade and a common identification certificate for imported and exported goods. At present, my country's commodity code has 22 categories, 98 chapters, and a total of more than 13,000 entries. In the customs declaration business, the declared commodity code and product name must be consistent with the product name shown in the waybill, invoice, packing list and other relevant certification documents, and at the same time be consistent with the actual imported and exported goods. If the commodity code does not match, it will directly lead to different customs taxation, cause the evasion and omission of legal inspection commodities, affect the customs clearance speed of ports, reduce the customs declaration efficiency and credit of enterprises, and in serious cases, they will also bear administrative or criminal legal responsibilities. The purpose of commodity classification is to find the corresponding commodity code for imported and exported commodities, so as to accurately determine the tax rate, declaration elements, import and export control measures, applicable rules of origin and export tax rebate rate, etc.

[0003] At present, there are two ways to classify customs commodity codes. One is to perform simple fuzzy matching based on commodity keywords and commodity names declared by historical customs users, and return the data declared by historical customs users. The commodity codes may be repeated or not found, and there is no priority for the returned multiple commodity codes; the other is to return all the data declared by historical customs users containing the customs code based on the first few digits of the customs commodity code, and the commodity codes may be repeated or not found. These two methods have low accuracy due to incomplete basic data and untimely updates.

[0004] Therefore, a method, device and electronic device for recommending commodity codes for import and export commodities are proposed. Summary of the invention

[0005] This specification provides a method, device, and electronic device for recommending commodity codes for import and export commodities, which lowers the threshold for declaring commodity codes for import and export commodities, improves the accuracy of declaring commodity codes for import and export commodities, and thus reduces the burden on enterprises.

[0006] The present application provides a method for recommending commodity codes for import and export commodities using the following technical solutions, including:

[0007] Get the user's product information to be classified;

[0008] Inputting the commodity information to be classified into a historical declaration data model to obtain a plurality of first recommended codes for the commodity information to be classified;

[0009] Inputting the commodity information to be classified into a professional declaration data model to obtain a plurality of second recommended codes for the commodity information to be classified;

[0010] Determining, based on a score weighting rule, coding recommendation scores of a plurality of target recommended codings, wherein the target recommended codings include a plurality of the first recommended codings and a plurality of the second recommended codings;

[0011] Determine whether the coding recommendation score exceeds a threshold;

[0012] When there is a code recommendation score exceeding the threshold, all target recommended codes whose code recommendation scores exceed the threshold are arranged in positive order;

[0013] Recommend the target recommendation codes of the top N (N≥1) recommended scores in the positive order to the user

[0014] Optionally, inputting the commodity information to be classified into a historical declaration data model to obtain a first recommended code for the commodity information to be classified includes:

[0015] Inputting the commodity information to be classified into a pre-trained language representation sub-model to obtain a keyword vector of the commodity information to be classified;

[0016] Inputting the keyword vector into the text feature extraction sub-model to obtain the expression features of the keyword vector;

[0017] The expression feature is input into a classification prediction sub-model to obtain a first recommended code for the commodity information to be classified.

[0018] Optionally, the step of inputting the commodity information to be classified into a professional declaration data model to obtain a second recommended code for the commodity information to be classified includes:

[0019] Inputting the commodity information to be classified into a word segmentation sub-model to obtain a minimum granularity target phrase for the commodity information to be classified;

[0020] Determining whether the minimum granularity target phrase exists in the constructed import and export commodity map;

[0021] When the minimum granularity target phrase exists in the constructed import and export commodity knowledge graph, a second recommended code for the commodity information to be classified is determined based on the import and export commodity knowledge graph.

[0022] Optionally, the constructed import and export commodity map includes:

[0023] Obtain classified product information and its corresponding classified product code;

[0024] Inputting the classified commodity information into the word segmentation sub-model to obtain the minimum granularity sample phrase of the classified commodity information;

[0025] Based on the relationship between the minimum granularity sample phrases of the classified commodity information and the classified commodity codes, a commodity code mapping relationship network is formed;

[0026] Integrate several of the commodity mapping relationship networks to obtain an import and export commodity knowledge graph.

[0027] Optionally, the method further includes:

[0028] When there is no code recommendation score exceeding the threshold, a self-check window is pushed to the user end, and the self-check window includes confirmation of whether the user's to-be-classified commodity information is correct;

[0029] Receiving a self-check result of the user terminal;

[0030] When the self-check result is that the user terminal confirms the self-check window, it is determined that the information of the commodity to be classified by the user is correct, and the operation ends;

[0031] When the self-check result is that the user terminal denies the self-check window, it is determined that the user's to-be-classified commodity information is wrong, and a modification window is pushed to the user terminal;

[0032] Receiving the commodity information to be classified modified by the user terminal;

[0033] Return and input the modified information of the commodity to be classified into the constructed historical declaration data model to obtain the first recommended code of the modified information of the commodity to be classified.

[0034] Optionally, the method further includes:

[0035] Storing the user's commodity information to be classified and commodity classification information;

[0036] The user's commodity information to be classified and commodity classification information are updated to the historical declaration data model according to a preset time.

[0037] The present application provides a device for recommending commodity codes for import and export commodities, which adopts the following technical solution, including:

[0038] The acquisition module is used to obtain the user's product information to be classified;

[0039] A first input module, used for inputting the commodity information to be classified into a historical declaration data model to obtain a plurality of first recommended codes of the commodity information to be classified;

[0040] A second input module is used to input the commodity information to be classified into a professional declaration data model to obtain a plurality of second recommended codes for the commodity information to be classified;

[0041] A score determination module, used to determine the code recommendation scores of a plurality of target recommended codes based on a score weighting rule, wherein the target recommended codes include a plurality of the first recommended codes and a plurality of the second recommended codes;

[0042] A judgment module, used to judge whether the coding recommendation score exceeds a threshold;

[0043] An arrangement module, configured to arrange all target recommended codes whose recommendation scores exceed the threshold in positive order when there is a code recommendation score exceeding the threshold;

[0044] The recommendation module is used to recommend the top N (N≥1) target recommendation codes of the recommendation scores arranged in positive order to the user.

[0045] Optionally, the first input module includes:

[0046] A first input unit, used for inputting the commodity information to be classified into a pre-trained language representation sub-model to obtain a keyword vector of the commodity information to be classified;

[0047] A second input unit is used to input the keyword vector into the text feature extraction sub-model to obtain the expression feature of the keyword vector;

[0048] The prediction unit is used to input the expression feature into the classification prediction sub-model to obtain the first recommended code of the commodity information to be classified.

[0049] Optionally, the second input module includes:

[0050] A third input unit is used to input the commodity information to be classified into the word segmentation sub-model to obtain the minimum granularity target phrase of the commodity information to be classified;

[0051] A judgment unit, used for judging whether the minimum granularity target phrase exists in the constructed import and export commodity map;

[0052] A coding unit is determined, which is used to determine a second recommended coding of the commodity information to be classified based on the import and export commodity knowledge graph when the minimum granularity target phrase exists in the constructed import and export commodity knowledge graph.

[0053] Optionally, the constructed import and export commodity map includes:

[0054] Obtain classified product information and its corresponding classified product code;

[0055] Inputting the classified commodity information into the word segmentation sub-model to obtain the minimum granularity sample phrase of the classified commodity information;

[0056] Based on the relationship between the minimum granularity sample phrases of the classified commodity information and the classified commodity codes, a commodity code mapping relationship network is formed;

[0057] Integrate several of the commodity mapping relationship networks to obtain an import and export commodity knowledge graph.

[0058] Optionally, the device further comprises:

[0059] A push window module, for pushing a self-check window to a user terminal when there is no code recommendation score exceeding the threshold, wherein the self-check window includes confirmation of whether the user's to-be-classified commodity information is correct;

[0060] A receiving module, used for receiving the self-check result of the user terminal;

[0061] A confirmation module, configured to determine that the information of the commodity to be classified by the user is correct and terminate the operation when the self-check result is that the user terminal confirms the self-check window;

[0062] A denial module, configured to determine that the user's commodity information to be classified is wrong and push a modification window to the user terminal when the self-check result indicates that the user terminal denies the self-check window;

[0063] A revision module, used for receiving the information of the commodities to be classified modified by the user terminal;

[0064] The return module is used to return the historical declaration data model constructed by inputting the modified information of the commodity to be classified into the historical declaration data model to obtain the first recommended code of the modified information of the commodity to be classified.

[0065] Optionally, the device further comprises:

[0066] A storage module, used for storing the user's commodity information to be classified and commodity classification information;

[0067] The updating module is used to update the user's to-be-classified commodity information and commodity classification information to the historical declaration data model according to a preset time.

[0068] This specification also provides an electronic device, wherein the electronic device includes:

[0069] processor; and,

[0070] A memory storing computer executable instructions, which when executed cause the processor to perform any of the above methods.

[0071] The present specification also provides a computer-readable storage medium, wherein the computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, any of the above methods is implemented.

[0072] In order to solve the problems of high efficiency and error rate in the classification of import and export commodity codes in the prior art, the present application adopts the method of obtaining the commodity information to be classified from the user; inputting the commodity information to be classified into the historical declaration data model to obtain the first recommended codes of several commodity information to be classified; inputting the commodity information to be classified into the professional declaration data model to obtain the second recommended codes of several commodity information to be classified; determining the code recommendation scores of several target recommended codes based on the score weighting rule, wherein the target recommended codes include several first recommended codes and several second recommended codes; judging whether there are code recommendation scores exceeding the threshold; when there are code recommendation scores exceeding the threshold, arranging all target recommended codes whose code recommendation scores exceed the threshold in positive order; and recommending the first N (N≥1) target recommended codes of the recommended scores arranged in positive order to the user. Thus, the threshold for declaring the commodity codes of import and export commodities is lowered, the accuracy of declaring the commodity codes of import and export commodities is improved, and the burden on enterprises is also reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 A schematic diagram of the structure of a system for recommending commodity codes for import and export commodities involved in an embodiment of the present invention;

[0074] Figure 2 A flowchart of a first embodiment of a method for recommending commodity codes for import and export commodities provided by an embodiment of the present invention;

[0075] Figure 3 is a schematic diagram of a historical declaration data model provided by an embodiment of the present invention;

[0076] Figure 4 A detailed flow chart of step S120 of the first embodiment of a method for recommending commodity codes for import and export commodities provided by an embodiment of the present invention;

[0077] Figure 5 A detailed flow chart of step S130 of the first embodiment of a method for recommending commodity codes for import and export commodities provided by an embodiment of the present invention;

[0078] Figure 6 It is a schematic diagram of the import and export commodity map constructed according to an embodiment of the present invention;

[0079] Figure 7 A schematic diagram of the structure of a device for recommending commodity codes for import and export commodities provided by an embodiment of the present invention;

[0080] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0081] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0082] In order to solve the problems of high efficiency and error rate in the classification of import and export commodity codes in the prior art, the present application adopts the method of obtaining the commodity information to be classified from the user; inputting the commodity information to be classified into the historical declaration data model to obtain the first recommended codes of several commodity information to be classified; inputting the commodity information to be classified into the professional declaration data model to obtain the second recommended codes of several commodity information to be classified; determining the code recommendation scores of several target recommended codes based on the score weighting rule, wherein the target recommended codes include several first recommended codes and several second recommended codes; judging whether there are code recommendation scores exceeding the threshold; when there are code recommendation scores exceeding the threshold, arranging all target recommended codes whose code recommendation scores exceed the threshold in positive order; and recommending the first N (N≥1) target recommended codes of the recommended scores arranged in positive order to the user. Thus, the threshold for declaring the commodity codes of import and export commodities is lowered, the accuracy of declaring the commodity codes of import and export commodities is improved, and the burden on enterprises is also reduced.

[0083] In order to better understand the above technical solution, exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.

[0084] Those skilled in the art will understand that Figure 1 The system structure of a recommended import and export commodity code shown does not constitute a limitation on a system of recommended import and export commodity codes. A system of recommended import and export commodity codes may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0085] As a way to implement this, Figure 1 As shown, Figure 1 The present invention is a schematic diagram of a system structure for recommending commodity codes for import and export commodities according to an embodiment of the present invention.

[0086] The processor 1100 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor 1100 or an instruction in the form of software. The above processor 1100 may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The disclosed methods, steps and logic block diagrams in the embodiments of the present invention can be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The software module may be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 1200, and the processor 1100 reads the information in the memory 1200 and completes the steps of the above method in combination with its hardware.

[0087] It can be understood that the memory 1200 in the embodiment of the present invention can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM) and direct memory bus random access memory (DRRAM). The memory 1200 of the system and method described in the embodiments of the present invention is intended to include, but is not limited to, these and any other suitable types of memory.

[0088] In the embodiment of the present application, the processor 1100 may be used to call the recommended import and export commodity coding program stored in the memory 1200, and perform the following operations:

[0089] Get the user's product information to be classified;

[0090] Inputting the commodity information to be classified into a historical declaration data model to obtain a plurality of first recommended codes for the commodity information to be classified;

[0091] Inputting the commodity information to be classified into a professional declaration data model to obtain a plurality of second recommended codes for the commodity information to be classified;

[0092] Determining, based on a score weighting rule, coding recommendation scores of a plurality of target recommended codings, wherein the target recommended codings include a plurality of the first recommended codings and a plurality of the second recommended codings;

[0093] Determine whether the coding recommendation score exceeds a threshold;

[0094] When there is a code recommendation score exceeding the threshold, all target recommended codes whose code recommendation scores exceed the threshold are arranged in positive order;

[0095] The top N (N≥1) target recommended codes of the recommendation scores in the positive order are recommended to the user.

[0096] In one embodiment, the processor 1100 may be used to call the recommended import and export commodity code program stored in the memory 1200, and perform the following operations:

[0097] Inputting the commodity information to be classified into a pre-trained language representation sub-model to obtain a keyword vector of the commodity information to be classified;

[0098] Inputting the keyword vector into the text feature extraction sub-model to obtain the expression features of the keyword vector;

[0099] The expression feature is input into a classification prediction sub-model to obtain a first recommended code for the commodity information to be classified.

[0100] In one embodiment, the processor 1100 may be used to call the recommended import and export commodity code program stored in the memory 1200, and perform the following operations:

[0101] Inputting the commodity information to be classified into a word segmentation sub-model to obtain a minimum granularity target phrase for the commodity information to be classified;

[0102] Determining whether the minimum granularity target phrase exists in the constructed import and export commodity map;

[0103] When the minimum granularity target phrase exists in the constructed import and export commodity knowledge graph, a second recommended code for the commodity information to be classified is determined based on the import and export commodity knowledge graph.

[0104] In one embodiment, the processor 1100 may be used to call the recommended import and export commodity code program stored in the memory 1200, and perform the following operations:

[0105] Obtain classified product information and its corresponding classified product code;

[0106] Inputting the classified commodity information into the word segmentation sub-model to obtain the minimum granularity sample phrase of the classified commodity information;

[0107] Based on the relationship between the minimum granularity sample phrases of the classified commodity information and the classified commodity codes, a commodity code mapping relationship network is formed;

[0108] Integrate several of the commodity mapping relationship networks to obtain an import and export commodity knowledge graph.

[0109] In one embodiment, the processor 1100 may be used to call the recommended import and export commodity code program stored in the memory 1200, and perform the following operations:

[0110] When there is no code recommendation score exceeding the threshold, a self-check window is pushed to the user end, and the self-check window includes confirmation of whether the user's to-be-classified commodity information is correct;

[0111] Receiving a self-check result of the user terminal;

[0112] When the self-check result is that the user terminal confirms the self-check window, it is determined that the information of the commodity to be classified by the user is correct, and the operation ends;

[0113] When the self-check result is that the user terminal denies the self-check window, it is determined that the user's to-be-classified commodity information is wrong, and a modification window is pushed to the user terminal;

[0114] Receiving the commodity information to be classified modified by the user terminal;

[0115] Return and input the modified information of the commodity to be classified into the constructed historical declaration data model to obtain the first recommended code of the modified information of the commodity to be classified.

[0116] In one embodiment, the processor 1100 may be used to call the recommended import and export commodity code program stored in the memory 1200, and perform the following operations:

[0117] Storing the user's commodity information to be classified and commodity classification information;

[0118] The user's commodity information to be classified and commodity classification information are updated to the historical declaration data model according to a preset time.

[0119] Since the recommended import and export commodity coding device provided in the embodiment of the present application is the recommended import and export commodity coding device adopted by the method of the embodiment of the present application, based on the method introduced in the embodiment of the present application, the person skilled in the art can understand the specific structure and deformation of the intelligent classification device, so it will not be repeated here. All the recommended import and export commodity coding devices adopted by the method of the embodiment of the present application belong to the scope of protection of this application. The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments.

[0120] For software implementation, the technology described in the embodiments of the present invention can be implemented by a module (such as a process, function, etc.) that performs the functions described in the embodiments of the present invention. The software code can be stored in a memory and executed by a processor. The memory can be implemented in the processor or outside the processor.

[0121] Based on the above structure, an embodiment of the present invention is proposed.

[0122] Reference Figure 2 , Figure 2 A flowchart of a first embodiment of a method for recommending commodity codes for import and export commodities provided by an embodiment of the present invention includes the following steps:

[0123] Step S110, obtaining the user's product information to be classified.

[0124] In this embodiment, the information of the commodity to be classified includes relevant information of the commodity to be classified, and the relevant information includes the product name, details, and the name of the manufacturer and trading company, etc. For example, taking a smart phone as an example, the data information may be: "smart phone, 6.1 inches, IOS system, 128GB, with charger, 320.00g / unit". The acquisition method may be input using an external keyboard, input using a touch screen, or provided by voice or other methods.

[0125] Step S120: input the commodity information to be classified into a historical declaration data model to obtain a plurality of first recommended codes for the commodity information to be classified.

[0126] In this embodiment, based on the commodity information to be classified provided by the user, the historical declaration data of enterprises that matches the commodity information to be classified is searched in the historical declaration data model, so as to determine the first recommended code for the commodity information to be classified. In this embodiment, before outputting the first recommended code, the scores of several first recommended codes can be determined based on the score rule to determine whether there is a first recommended code with a score exceeding the first threshold. When there is a first recommended code with a score exceeding the first threshold, the first recommended code with a score exceeding the first threshold is output. In this embodiment, the first threshold can be a value set based on experience or a value obtained by statistical analysis. In this embodiment, referring to Figure 3 , Figure 3 Schematic diagram of the historical declaration data model provided by the embodiment of the present invention. The historical declaration data model is trained by using the historical declaration data of the enterprise to obtain the commodity code corresponding to the historical declaration data of the enterprise. The historical declaration data of the enterprise is obtained regularly from the customs declaration information network or other related websites for participating in the training.

[0127] Reference Figure 4 , Figure 4 A detailed flow chart of step S120 of the first embodiment of a method for recommending commodity codes for import and export commodities provided by an embodiment of the present invention includes the following steps:

[0128] Step S121: input the commodity information to be classified into a pre-trained language representation sub-model to obtain a keyword vector of the commodity information to be classified.

[0129] In this embodiment, the pre-trained language representation sub-model includes BERT (Bidirectional Encoder Representation from Transformers), which is a pre-trained language representation model. It emphasizes that it is no longer pre-trained by using the traditional unidirectional language model or the shallow splicing of two unidirectional language models as in the past, but a new masked language (MLM, Masked Language Model) is used to generate deep bidirectional language representation. The information of the products to be classified is converted into keyword vectors through BERT's Embedding. Embedding is a way to convert discrete variables into continuous vector representations.

[0130] Step S122: input the keyword vector into the text feature extraction sub-model to obtain the expression features of the keyword vector.

[0131] In this embodiment, the sub-model for extracting text features includes TextCNN, which has a more outstanding performance in text classification problems. Intuitively, TextCNN obtains the feature representation of n-grams in sentences through one-dimensional convolution. TextCNN has a strong ability to extract shallow text features. It works well in short text fields such as search and dialogue fields when focusing on intent classification. It is widely used and fast, and is generally the first choice. For long text fields, TextCNN mainly relies on filter windows to extract features, has limited capabilities in long-distance modeling, and is insensitive to word order. TextCNN converts keyword vectors into expression features of keyword vectors.

[0132] Step S123: input the expression feature into a classification prediction sub-model to obtain a first recommended code for the commodity information to be classified.

[0133] In this embodiment, the classification prediction sub-model includes Softmax, which is used in the multi-classification process. It maps the outputs of multiple neurons to the interval (0, 1), which can be understood as probability, so as to perform multi-classification. The classification corresponding to the expression feature is determined by Softmax, and the corresponding first recommended code is determined based on the corresponding classification.

[0134] Step S130, inputting the commodity information to be classified into a professional declaration data model to obtain a plurality of second recommended codes for the commodity information to be classified.

[0135] In this embodiment, the customs commodity codes and their corresponding commodity names, commodity specifications, etc. formulated by the customs bureau that match the commodity information to be classified provided by the user are searched in the professional declaration data model, so as to determine the second recommended code of the commodity information to be classified. In this embodiment, before outputting the second recommended code, the scores of several second recommended codes can be determined based on the score rule, and it is judged whether there is a second recommended code with a score exceeding the second threshold. When there is a second recommended code with a score exceeding the second threshold, the second recommended code with a score exceeding the second threshold is output. In this embodiment, the second threshold can be a value set based on experience or a value obtained by statistical analysis. In this embodiment, the historical declaration data model is trained using the customs commodity codes and their corresponding commodity names, commodity specifications, etc. formulated by the customs bureau to obtain the commodity codes corresponding to the commodity names, commodity specifications, etc. The customs commodity codes and their corresponding commodity names, commodity specifications, etc. formulated by the customs bureau are regularly obtained from the customs declaration information network or other relevant websites for participation in training.

[0136] Reference Figure 5 , Figure 5A detailed flow chart of step S130 of the first embodiment of a method for recommending commodity codes for import and export commodities provided by an embodiment of the present invention includes the following steps:

[0137] Step S131, inputting the commodity information to be classified into a word segmentation sub-model to obtain a minimum granularity target phrase of the commodity information to be classified.

[0138] In the present embodiment, the word segmentation submodel includes an HMM model (Hidden Markov Model), a CRF model (Conditional Random Field Algorithm), and an LSTM model (Long Short-Term Memory). The HMM model is a dynamic Bayesian network with the simplest structure. It is a particularly famous directed graph structure, mainly used for modeling time series data, and is widely used in fields such as speech recognition and natural language processing. In the word segmentation algorithm, hidden Markov is often used as an algorithm that can discover new words. Through massive data learning, it can identify names, place names, new words on the Internet, etc. one by one, and has a wide range of application scenarios. CRF combines the characteristics of the maximum entropy model and the hidden Markov model, and is an undirected graph model. In recent years, it has achieved good results in sequence labeling tasks such as word segmentation, part-of-speech tagging, and named entity recognition. LSTM is a special RNN, which is mainly used to solve the gradient disappearance and gradient explosion problems in the long sequence training process. In this embodiment, the information of the goods to be classified is input into the LSTM model to obtain the minimum granularity target phrase of the goods to be classified. For example, "21-inch LCD screen without touch function weighing 1.27 kg / piece" is input into the LSTM model to obtain "21 inches", "1.27 kg / piece", "without touch", and "LCD screen".

[0139] Step S132, determining whether the minimum granularity target phrase exists in the constructed import and export commodity map.

[0140] Step S133: when the minimum granularity target phrase exists in the constructed import and export commodity knowledge graph, a second recommended code for the commodity information to be classified is determined based on the import and export commodity knowledge graph.

[0141] Reference Figure 6 , Figure 6 : is a schematic diagram of an import and export commodity map constructed according to an embodiment of the present invention. In this embodiment, the constructed import and export commodity map includes:

[0142] Obtain classified product information and its corresponding classified product code;

[0143] Inputting the classified commodity information into the word segmentation sub-model to obtain the minimum granularity sample phrase of the classified commodity information;

[0144] Based on the relationship between the minimum granularity sample phrases of the classified commodity information and the classified commodity codes, a commodity code mapping relationship network is formed;

[0145] Integrate several of the commodity mapping relationship networks to obtain an import and export commodity knowledge graph.

[0146] In this embodiment, the classified commodity information and its corresponding classified commodity codes, i.e., the customs commodity codes formulated by the customs bureau and their corresponding commodity names, commodity specifications, etc., are input into the LSTM model to obtain the minimum granularity target phrases corresponding to the commodity names, commodity specifications, etc. Based on the relationship between the minimum granularity sample phrases of the classified commodity information and the classified commodity codes, a commodity code mapping relationship network is formed; several commodity mapping relationship networks are integrated to obtain the import and export commodity knowledge graph.

[0147] Step S140, determining the coding recommendation scores of a plurality of target recommended codings based on a score weighting rule, wherein the target recommended codings include a plurality of the first recommended codings and a plurality of the second recommended codings.

[0148] In this embodiment, the score weighting rule may be to first determine the correlation between a number of first recommendation codes, a number of second recommendation codes and the information of the commodity to be classified, and the higher the correlation, the higher the recommendation score. Then, different weight values ​​are assigned to the recommendation scores of the first recommendation code and the second recommendation code, for example, the recommendation score of the first recommendation code is assigned a weight value of 0.3, and the recommendation score of the second recommendation code is assigned a weight value of 0.7. That is, when the first recommendation code and the second recommendation code are repeated, the recommendation score of the code is 0.3*the recommendation score of the first recommendation code + 0.7*the recommendation score of the second recommendation code; when the first recommendation code and the second recommendation code are not repeated, the recommendation score of the first recommendation code is 0.3*the recommendation score of the first recommendation code, and the recommendation score of the second recommendation code is 0.7*the recommendation score of the second recommendation code.

[0149] Step S150, determining whether there is a coding recommendation score exceeding a threshold.

[0150] Step S160 , when there is a code recommendation score exceeding the threshold, all target recommended codes whose code recommendation scores exceed the threshold are arranged in positive order.

[0151] Step S170: recommending the top N (N≥1) target recommended codes in the positive order of recommendation scores to the user.

[0152] In this embodiment, the method further includes:

[0153] When there is no code recommendation score exceeding the threshold, a self-check window is pushed to the user end, and the self-check window includes confirmation of whether the user's to-be-classified commodity information is correct;

[0154] Receiving a self-check result of the user terminal;

[0155] When the self-check result is that the user terminal confirms the self-check window, it is determined that the information of the commodity to be classified by the user is correct, and the operation ends;

[0156] When the self-check result is that the user terminal denies the self-check window, it is determined that the user's to-be-classified commodity information is wrong, and a modification window is pushed to the user terminal;

[0157] Receiving the commodity information to be classified modified by the user terminal;

[0158] Return and input the modified information of the commodity to be classified into the constructed historical declaration data model to obtain the first recommended code of the modified information of the commodity to be classified.

[0159] In this embodiment, the threshold value can be a value set based on experience, or a value obtained by statistical analysis. It is determined whether there is a code recommendation score exceeding the threshold value. If so, the target recommendation codes with code recommendation scores exceeding the threshold value are arranged in positive order, and the target recommendation codes with the top N recommended scores in positive order are recommended to the user. The top N can be selected based on the user's preferences, or can be set based on experience values. Specifically, the top 5 target recommendation codes with the recommended scores in positive order can be recommended to the user. If it does not exist, a self-check window is pushed to the user so that the user can self-check the input commodity information to be classified to see if there are any typos, etc., and receive the user's self-check results. When the user self-checks and confirms that there is no error, the commodity information to be classified may be a new commodity and there is no historical data, that is, no recommendation is made; when the user self-checks and confirms that there is an error, that is, the commodity information to be classified input by the user has typos, etc., the user makes modifications based on the modification window, and then returns to step S120.

[0160] In this embodiment, the method further includes:

[0161] Storing the user's commodity information to be classified and commodity classification information;

[0162] The user's commodity information to be classified and commodity classification information are updated to the historical declaration data model according to a preset time.

[0163] In this embodiment, the user's information on commodities to be classified and the final commodity classification information are stored. After a preset storage time, the user's information on commodities to be classified and the final commodity classification information are updated to the historical declaration data model to establish self-update and self-iteration of system data.

[0164] Based on the same inventive concept, the present invention also provides a device for recommending commodity codes for import and export commodities, such as Figure 7 As shown, Figure 7 A schematic diagram of the structure of a device for recommending commodity codes for import and export commodities provided in an embodiment of the present invention, the system includes: an acquisition module 10, a first input module 20, a second input module 30, a score determination module 40, a judgment module 50, an arrangement module 60 and a recommendation module 70.

[0165] The acquisition module 10 is used to acquire the user's commodity information to be classified;

[0166] A first input module 20, used for inputting the commodity information to be classified into the historical declaration data model to obtain a plurality of first recommended codes of the commodity information to be classified;

[0167] The second input module 30 is used to input the commodity information to be classified into the professional declaration data model to obtain a plurality of second recommended codes of the commodity information to be classified;

[0168] A score determination module 40, configured to determine code recommendation scores of a plurality of target recommended codes based on a score weighting rule, wherein the target recommended codes include a plurality of the first recommended codes and a plurality of the second recommended codes;

[0169] A judging module 50, used to judge whether the coding recommendation score exceeds a threshold;

[0170] An arrangement module 60 is used for arranging all target recommended codes whose recommendation scores exceed the threshold in positive order when there is a code recommendation score exceeding the threshold;

[0171] The recommendation module 70 is used to recommend the top N (N≥1) target recommended codes of the recommendation scores arranged in the positive order to the user.

[0172] Optionally, the first input module 20 includes:

[0173] A first input unit, used for inputting the commodity information to be classified into a pre-trained language representation sub-model to obtain a keyword vector of the commodity information to be classified;

[0174] A second input unit is used to input the keyword vector into the text feature extraction sub-model to obtain the expression feature of the keyword vector;

[0175] The prediction unit is used to input the expression feature into the classification prediction sub-model to obtain the first recommended code of the commodity information to be classified.

[0176] Optionally, the second input module 30 includes:

[0177] A third input unit is used to input the commodity information to be classified into the word segmentation sub-model to obtain the minimum granularity target phrase of the commodity information to be classified;

[0178] A judgment unit, used for judging whether the minimum granularity target phrase exists in the constructed import and export commodity map;

[0179] A coding unit is determined, which is used to determine a second recommended coding of the commodity information to be classified based on the import and export commodity knowledge graph when the minimum granularity target phrase exists in the constructed import and export commodity knowledge graph.

[0180] Optionally, the constructed import and export commodity map includes:

[0181] Obtain classified product information and its corresponding classified product code;

[0182] Inputting the classified commodity information into the word segmentation sub-model to obtain the minimum granularity sample phrase of the classified commodity information;

[0183] Based on the relationship between the minimum granularity sample phrases of the classified commodity information and the classified commodity codes, a commodity code mapping relationship network is formed;

[0184] Integrate several of the commodity mapping relationship networks to obtain an import and export commodity knowledge graph.

[0185] Optionally, the device further comprises:

[0186] A push window module, for pushing a self-check window to a user terminal when there is no code recommendation score exceeding the threshold, wherein the self-check window includes confirmation of whether the user's to-be-classified commodity information is correct;

[0187] A receiving module, used for receiving the self-check result of the user terminal;

[0188] A confirmation module, configured to determine that the information of the commodity to be classified by the user is correct and terminate the operation when the self-check result is that the user terminal confirms the self-check window;

[0189] A denial module, configured to determine that the user's commodity information to be classified is wrong and push a modification window to the user terminal when the self-check result indicates that the user terminal denies the self-check window;

[0190] A revision module, used for receiving the information of the commodities to be classified modified by the user terminal;

[0191] The return module is used to return the historical declaration data model constructed by inputting the modified information of the commodity to be classified into the historical declaration data model to obtain the first recommended code of the modified information of the commodity to be classified.

[0192] Optionally, the device further comprises:

[0193] A storage module, used for storing the user's commodity information to be classified and commodity classification information;

[0194] The updating module is used to update the user's to-be-classified commodity information and commodity classification information to the historical declaration data model according to a preset time.

[0195] The functions of the device in the embodiment of the present invention have been described in the above method embodiment, so for details not provided in the description of this embodiment, please refer to the relevant description in the above embodiment, and no further description will be given here.

[0196] Based on the same inventive concept, an embodiment of the present application further provides a computer-readable storage medium, which stores a program for recommending commodity codes for import and export commodities. When the program for recommending commodity codes for import and export commodities is executed by a processor, the various steps of the method for recommending commodity codes for import and export commodities as described above are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0197] Since the computer-readable storage medium provided in the embodiments of the present application is a computer-readable storage medium used to implement the method of the embodiments of the present application, based on the method introduced in the embodiments of the present application, those skilled in the art can understand the specific structure and deformation of the computer-readable storage medium, so it is not repeated here. All computer-readable storage media used in the methods of the embodiments of the present application belong to the scope of protection of this application.

[0198] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0199] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0200] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0201] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0202] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0203] It should be noted that in the claims, any reference signs placed between brackets shall not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claim. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The invention may be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In a unit claim enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third etc. does not indicate any order. These words may be interpreted as names.

[0204] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0205] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for recommending commodity codes for import and export commodities, characterized in that: include: Get the user's product information to be classified; Inputting the commodity information to be classified into a historical declaration data model to obtain a plurality of first recommended codes for the commodity information to be classified; Inputting the commodity information to be classified into a professional declaration data model to obtain a plurality of second recommended codes of the commodity information to be classified, including: inputting the commodity information to be classified into a word segmentation sub-model to obtain a minimum granularity target phrase of the commodity information to be classified; judging whether the minimum granularity target phrase exists in the constructed import and export commodity knowledge graph; when the minimum granularity target phrase exists in the constructed import and export commodity knowledge graph, determining the second recommended code of the commodity information to be classified based on the import and export commodity knowledge graph; Determining, based on a score weighting rule, coding recommendation scores of a plurality of target recommended codings, wherein the target recommended codings include a plurality of the first recommended codings and a plurality of the second recommended codings; Determine whether the coding recommendation score exceeds a threshold; When there is a code recommendation score exceeding the threshold, all target recommended codes whose code recommendation scores exceed the threshold are arranged in positive order; The top N target recommendation codes in the positive order of recommendation scores are recommended to the user, where N≥1.

2. The method for recommending commodity codes for import and export commodities according to claim 1, characterized in that: The step of inputting the commodity information to be classified into a historical declaration data model to obtain a first recommended code for the commodity information to be classified includes: Inputting the commodity information to be classified into a pre-trained language representation sub-model to obtain a keyword vector of the commodity information to be classified; Inputting the keyword vector into the text feature extraction sub-model to obtain the expression features of the keyword vector; The expression feature is input into a classification prediction sub-model to obtain a first recommended code for the commodity information to be classified.

3. The method for recommending commodity codes for import and export commodities according to claim 2, characterized in that: The constructed import and export commodity knowledge graph includes: Obtain classified product information and its corresponding classified product code; Inputting the classified commodity information into the word segmentation sub-model to obtain the minimum granularity sample phrase of the classified commodity information; Based on the relationship between the minimum granularity sample phrases of the classified commodity information and the classified commodity codes, a commodity code mapping relationship network is formed; Integrate several of the commodity mapping relationship networks to obtain an import and export commodity knowledge graph.

4. The method for recommending commodity codes for import and export commodities according to claim 3, characterized in that: Also includes: When there is no code recommendation score exceeding the threshold, a self-check window is pushed to the user end, and the self-check window includes confirmation of whether the user's to-be-classified commodity information is correct; Receiving a self-check result of the user terminal; When the self-check result is that the user terminal confirms the self-check window, it is determined that the information of the commodity to be classified by the user is correct, and the operation ends; When the self-check result is that the user terminal denies the self-check window, it is determined that the user's to-be-classified commodity information is wrong, and a modification window is pushed to the user terminal; Receiving the commodity information to be classified modified by the user terminal; Return and input the modified information of the commodity to be classified into the constructed historical declaration data model to obtain the first recommended code of the modified information of the commodity to be classified.

5. The method for recommending commodity codes for import and export commodities according to claim 4, characterized in that: Also includes: Storing the user's commodity information to be classified and commodity classification information; The user's commodity information to be classified and commodity classification information are updated to the historical declaration data model according to a preset time.

6. A device for recommending commodity codes for import and export commodities, characterized in that: include: The acquisition module is used to obtain the user's product information to be classified; A first input module, used for inputting the commodity information to be classified into a historical declaration data model to obtain a plurality of first recommended codes of the commodity information to be classified; The second input module is used to input the commodity information to be classified into a professional declaration data model to obtain a plurality of second recommended codes of the commodity information to be classified, including: inputting the commodity information to be classified into a word segmentation sub-model to obtain a minimum granularity target phrase of the commodity information to be classified; judging whether the minimum granularity target phrase exists in the constructed import and export commodity knowledge graph; when the minimum granularity target phrase exists in the constructed import and export commodity knowledge graph, determining the second recommended code of the commodity information to be classified based on the import and export commodity knowledge graph; A score determination module, used to determine the code recommendation scores of a plurality of target recommended codes based on a score weighting rule, wherein the target recommended codes include a plurality of the first recommended codes and a plurality of the second recommended codes; A judgment module, used to judge whether the coding recommendation score exceeds a threshold; An arrangement module, configured to arrange all target recommended codes whose recommendation scores exceed the threshold in positive order when there is a code recommendation score exceeding the threshold; The recommendation module is used to recommend the top N target recommendation codes of the recommendation scores arranged in positive order to the user, wherein N≥1.

7. The device for recommending commodity codes for import and export commodities according to claim 6, characterized in that: The first input module comprises: A first input unit, used for inputting the commodity information to be classified into a pre-trained language representation sub-model to obtain a keyword vector of the commodity information to be classified; A second input unit is used to input the keyword vector into the text feature extraction sub-model to obtain the expression feature of the keyword vector; The prediction unit is used to input the expression feature into the classification prediction sub-model to obtain the first recommended code of the commodity information to be classified.

8. An electronic device, wherein: The electronic device includes: processor; and, A memory storing computer executable instructions which, when executed, cause the processor to perform a method according to any one of claims 1 to 5.

9. A computer-readable storage medium, wherein: The computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, the method of any one of claims 1 to 5 is implemented.

Images