Automatic classification method and device, storage medium and equipment

Abstract

The embodiment of the invention discloses an automatic classification method and device, a storage medium and equipment, and the method comprises the steps: presetting reference commodity information, marking the commodity categories to which the reference commodity information belongs, and obtaining the to-be-classified commodity information when the to-be-classified commodity information is obtained; the method comprises the steps of firstly recalling reference commodity information which is relatively similar to the to-be-classified commodity information according to a first matching degree between the to-be-classified commodity information and each piece of reference commodity information, and then, according to a second matching degree between the to-be-classified commodity information and a commodity category to which each piece of recalled reference commodity information belongs, in the commodity category to which each piece of recalled reference commodity information belongs, recalling the to-be-classified commodity information. And recalling commodity categories matched with the to-be-classified commodity information, finally determining the final similarity between the to-be-classified commodity information and each recalled commodity category according to the first matching degree and the second matching degree, and determining the commodity category to which the to-be-classified commodity information belongs according to the final similarity. The to-be-classified commodity information can be automatically and accurately classified.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to an automatic classification method, device, storage medium and equipment. Background Art

[0002] At present, e-commerce platforms have been widely used around the world. Merchants can upload product information to e-commerce platforms to sell corresponding products through e-commerce platforms, and users can also purchase the products they need through e-commerce platforms.

[0003] Generally, e-commerce platforms usually maintain a huge commodity category tree, which usually contains more than 10,000 category nodes. In other words, the number of commodity categories preset in the e-commerce platform is at least more than 10,000. After merchants upload commodity information to the e-commerce platform, the e-commerce platform needs to classify the commodity information into one of the many commodity categories based on the commodity information, which is undoubtedly a huge challenge.

[0004] Therefore, the embodiments of this specification provide an automatic classification method, apparatus, storage medium and device. Summary of the invention

[0005] The embodiments of this specification provide an automatic classification method, device, storage medium and electronic device to partially solve the problems existing in the above-mentioned prior art.

[0006] The embodiments of this specification adopt the following technical solutions:

[0007] This manual provides an automatic classification method,

[0008] Get the information of the products to be classified uploaded by the merchant;

[0009] Determining a first matching degree between the to-be-classified commodity information and each reference commodity information according to the to-be-classified commodity information and each preset reference commodity information;

[0010] According to the first matching degree, determining a plurality of reference commodity information matching the commodity information to be classified as the recalled commodity information;

[0011] For each recalled product information, a product category to which the recalled product information belongs that is pre-marked is used as the recalled product category, and a second matching degree between the product information to be classified and the recalled product category is determined according to the category information of the recalled product category and the product information to be classified; and a final matching degree between the product information to be classified and the recalled product category is determined according to the second matching degree and the first matching degree between the product information to be classified and the recalled product information;

[0012] According to the final matching degree between the commodity information to be classified and each recalled commodity category, the commodity category to which the commodity information to be classified belongs is determined in each recalled commodity category.

[0013] Optionally, determining a first matching degree between the to-be-classified commodity information and each reference commodity information according to the to-be-classified commodity information and each preset reference commodity information specifically includes:

[0014] Inputting the commodity information to be classified into a pre-trained feature representation model to obtain the commodity features to be classified corresponding to the commodity information to be classified output by the feature representation model;

[0015] A first matching degree between the commodity information to be classified and each reference commodity information is determined according to the commodity characteristics to be classified and the pre-stored reference commodity characteristics corresponding to each reference commodity information.

[0016] Optionally, determining a second matching degree between the to-be-classified product information and the recalled product category according to the category information of the recalled product category and the to-be-classified product information specifically includes:

[0017] Inputting the commodity information to be classified into a pre-trained feature representation model to obtain the commodity features to be classified corresponding to the commodity information to be classified output by the feature representation model;

[0018] A second matching degree between the to-be-classified commodity information and the recalled commodity category is determined according to the to-be-classified commodity characteristics and the category characteristics corresponding to the pre-stored category information of each recalled commodity category.

[0019] Optionally, the information of commodities to be classified includes text information and image information of the commodities to be classified;

[0020] The reference product information includes text information and image information of the reference product;

[0021] The category information of the recalled product category includes text description information of the recalled product category;

[0022] The feature characterization model includes a BLIP model.

[0023] Optionally, determining, according to the first matching degree, a plurality of reference commodity information matching the commodity information to be classified specifically includes:

[0024] A specified number of reference product information are selected in order from high to low in the first matching degree as reference product information matching the product information to be classified.

[0025] Optionally, determining the final matching degree between the to-be-classified product information and the recalled product category according to the second matching degree and the first matching degree between the to-be-classified product information and the recalled product information specifically includes:

[0026] The second matching degree is weighted with the first matching degree between the to-be-classified product information and the recalled product information to obtain a final matching degree between the to-be-classified product information and the recalled product category.

[0027] Optionally, according to the final matching degree between the commodity information to be classified and each recalled commodity category, determining the commodity category to which the commodity information to be classified belongs in each recalled commodity category specifically includes:

[0028] Among the recalled commodity categories, the recalled commodity category with the highest final matching degree with the commodity information to be classified is determined as the commodity category to which the commodity information to be classified belongs.

[0029] This specification provides an automatic classification device, the device comprising:

[0030] The acquisition module is used to obtain the information of commodities to be classified uploaded by merchants;

[0031] A first matching module, configured to determine a first matching degree between the commodity information to be classified and each reference commodity information according to the commodity information to be classified and each preset reference commodity information;

[0032] A product recall module, configured to determine, according to the first matching degree, a plurality of reference product information matching the product information to be classified as recalled product information;

[0033] a second matching module, for each piece of recalled commodity information, using a pre-marked commodity category to which the recalled commodity information belongs as the recalled commodity category, and determining a second matching degree between the to-be-classified commodity information and the recalled commodity category according to the category information of the recalled commodity category and the to-be-classified commodity information;

[0034] a final matching module, configured to determine a final matching degree between the commodity information to be classified and the category of the recalled commodity according to the second matching degree and the first matching degree between the commodity information to be classified and the recalled commodity information;

[0035] The classification module is used to determine the commodity category to which the commodity information to be classified belongs in each recalled commodity category according to the final matching degree between the commodity information to be classified and each recalled commodity category.

[0036] This specification provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned automatic classification method is implemented.

[0037] The present specification provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned automatic classification method when executing the program.

[0038] At least one of the above technical solutions adopted in the embodiments of this specification can achieve the following beneficial effects:

[0039] The embodiment of the present specification discloses an intrusion detection method. The method pre-sets various reference product information and marks the product categories to which the reference product information belongs. When the product information to be classified is obtained, the reference product information that is relatively similar to the product information to be classified is first recalled based on the first matching degree between the product information to be classified and each reference product information. Then, based on the second matching degree between the product information to be classified and the product categories to which the recalled reference product information belongs, the product categories that are relatively matching with the product information to be classified are recalled from among the product categories to which the recalled reference product information belongs. Finally, the final similarity between the product information to be classified and each recalled product category is determined based on the above-mentioned first matching degree and second matching degree. The product category to which the product information to be classified belongs is determined based on the final similarity. The product information to be classified can be automatically and accurately classified. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The drawings described herein are used to provide a further understanding of this specification and constitute a part of this specification. The illustrative embodiments and descriptions of this specification are used to explain this specification and do not constitute an improper limitation on this specification. In the drawings:

[0041] Figure 1 A flowchart of an automatic classification method provided in an embodiment of this specification;

[0042] Figure 2 A schematic diagram of automatic classification provided in the embodiments of this specification;

[0043] Figure 3 A schematic diagram of an automatic classification device provided in an embodiment of this specification;

[0044] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this specification. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solutions and advantages of this specification more clear, the technical solutions of this specification will be clearly and completely described below in combination with the specific embodiments of this specification and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this specification.

[0046] The technical solutions provided by the embodiments of this specification are described in detail below in conjunction with the accompanying drawings.

[0047] Figure 1 A flowchart of an automatic classification method provided in this specification may include the following steps:

[0048] S100: Obtain the information of the products to be classified uploaded by the merchant.

[0049] In the embodiment of this specification, the following is performed: Figure 1 The subject of the automatic classification method shown can be any electronic device with computing capability of the e-commerce platform, such as a server, a server cluster, etc. The following description will only take the server as an example.

[0050] After the merchant uploads the product information of the product he wants to sell to the e-commerce platform, the server of the e-commerce platform can use the received product information as the product information to be classified.

[0051] S102: Determine a first matching degree between the to-be-classified commodity information and each reference commodity information according to the to-be-classified commodity information and each preset reference commodity information.

[0052] In the embodiment of the present specification, the e-commerce platform needs to preset some product information as reference product information. The product categories to which these reference product information belong are known, and the product categories to which these reference product information belong can be manually marked in advance.

[0053] After obtaining the product information to be classified in step S100, the server of the e-commerce platform can determine the matching degree between the product information to be classified and each reference product information according to the product information to be classified and each reference product information as the first matching degree.

[0054] Specifically, the information of the commodity to be classified can be converted into a corresponding feature vector as the feature of the commodity to be classified, and the information of the reference commodity can also be converted into a corresponding feature vector as the feature of the reference commodity. Finally, the similarity between the feature of the commodity to be classified and the feature of the reference commodity is calculated as the first matching degree. When calculating the similarity, the Euclidean distance between the feature of the commodity to be classified and the feature of the reference commodity can be calculated to determine the similarity based on the Euclidean distance, or the cosine similarity between the feature of the commodity to be classified and the feature of the reference commodity can be calculated to determine the similarity based on the cosine similarity. In short, as long as the similarity between the feature of the commodity to be classified and the feature of the reference commodity can be calculated, it will be fine.

[0055] S104: According to the first matching degree, determine a plurality of reference commodity information matching the commodity information to be classified as the recalled commodity information.

[0056] After determining the first matching degree between the product information to be classified and each reference product information, a specified number of reference product information may be selected in descending order of the first matching degree as reference product information matching the product information to be classified, i.e., recalled product information. In other words, among the reference product information, a specified number of reference product information with the highest first matching degree is selected as recalled product information.

[0057] S106: For each piece of recalled product information, the product category to which the recalled product information belongs is pre-marked as the recalled product category, and a second matching degree between the product information to be classified and the recalled product category is determined based on the category information of the recalled product category and the product information to be classified; and a final matching degree between the product information to be classified and the recalled product category is determined based on the second matching degree and the first matching degree between the product information to be classified and the recalled product information.

[0058] As mentioned above, for each reference product information, the product category to which the reference product information belongs has been pre-marked. Therefore, in step S106, the product category to which each recalled product information belongs selected in step S104 can be used as the recalled product category. Then, based on the category information of the recalled product category and the product information to be classified, the second matching degree between the product information to be classified and each recalled product category is determined.

[0059] Similar to determining the first matching degree between the commodity information to be classified and the reference commodity information, when determining the second matching degree, the commodity information to be classified may be converted into commodity features to be classified, and the category information of the recalled commodity category may be converted into a feature vector as a category feature, and then the second matching degree between the commodity information to be classified and each recalled commodity category may be determined based on the commodity features to be classified and the category features. Specifically, the similarity between the commodity information to be classified and the recalled commodity category may be calculated by using the Euclidean distance or cosine similarity method, and then the second matching degree may be determined based on the similarity, which will not be described in detail here.

[0060] Next, for a recalled product category, the second matching degree between the product information to be classified and the recalled product category has been obtained, and the first matching degree between the product information to be classified and the recalled product information under the recalled product category has also been obtained. In order to accurately obtain the final matching degree between the product information to be classified and the recalled product category, the final matching degree between the product information to be classified and the recalled product category can be determined based on the second matching degree between the product information to be classified and the recalled product category and the first matching degree between the product information to be classified and the recalled product information under the recalled product category.

[0061] Specifically, the second matching degree between the commodity information to be classified and the category of the recalled commodity, and the first matching degree between the commodity information to be classified and the recalled commodity information under the category of the recalled commodity, may be weighted to obtain the final matching degree between the commodity information to be classified and the category of the recalled commodity. In other words, the second matching degree between the commodity information to be classified and the category of the recalled commodity obtained in step S106 is not accurate, and therefore, the second matching degree is corrected by using the first matching degree between the commodity information to be classified and the recalled commodity information under the category of the recalled commodity, so as to obtain an accurate final matching degree.

[0062] S108: Determine the commodity category to which the commodity information to be classified belongs in each of the recalled commodity categories according to the final matching degree between the commodity information to be classified and each of the recalled commodity categories.

[0063] After obtaining the final matching degree between the product to be classified and each recalled product category, the product category to which the product to be classified belongs can be determined in each recalled product category according to the final matching degree. Specifically, in each recalled product category, the recalled product category with the highest final matching degree with the product to be classified can be determined as the product category to which the product to be classified belongs.

[0064] Through the above method, the first matching degree between the commodity information to be classified and each reference commodity information is first determined. According to the first matching degree, on the one hand, the commodity categories to which the commodity to be classified may belong can be roughly screened first. The recalled commodity categories to which each recalled commodity information with a higher first matching degree belongs are the commodity categories to which the commodity to be classified may belong. On the other hand, it also provides a basis for the subsequent correction of the second matching degree between the commodity information to be classified and each recalled commodity category. Then determine the second matching degree between the commodity information to be classified and each recalled commodity category. Then use the first matching degree to correct the second matching degree to obtain the final matching degree. Finally, according to the final matching degree between the commodity information to be classified and each recalled commodity category, determine the commodity category to which the commodity information to be classified belongs. The above method can effectively improve the accuracy of automatic classification of commodity categories to which commodity information belongs.

[0065] Further, in Figure 1 In the step S102 shown, when converting the commodity information to be classified into corresponding commodity features to be classified, a pre-trained machine learning model (hereinafter referred to as the feature representation model) can be used for conversion, that is, the commodity information to be classified can be input into the pre-trained feature representation model to obtain the commodity features to be classified corresponding to the commodity information to be classified output by the feature representation model, and then the first matching degree between the commodity information to be classified and each reference commodity information is determined based on the commodity features to be classified and the reference commodity features corresponding to each pre-saved reference commodity information.

[0066] It should be noted that the reference product features corresponding to each reference product information stored in advance are also obtained according to the reference product information through the above-mentioned feature representation model.

[0067] Similarly, in step S106, the features of the commodity to be classified corresponding to the commodity to be classified information can also be obtained through the above-mentioned feature characterization model, and then the second matching degree between the commodity to be classified information and the recalled commodity category is determined based on the features of the commodity to be classified and the category features corresponding to the category information of each pre-stored recalled commodity category. The category features corresponding to the pre-stored category information of each recalled commodity category are also obtained by using the above-mentioned feature characterization model based on the category information of the commodity category to which each reference commodity information belongs.

[0068] In addition, in actual application scenarios, the information of the commodity to be classified in the embodiments of this specification may include both text information of the commodity to be classified and image information of the commodity to be classified. Similarly, the reference commodity information may also include at least one of text information and image information of the reference commodity. The category information of the commodity category to which the reference commodity information belongs may be text description information of the commodity category.

[0069] Because the various types of product information mentioned above may include multimodal information of images and texts, the feature representation model for converting product information into product features may be a Contrastive Language-Image Pre-Training (CLIP) model.

[0070] Considering that the Bootstrapped Language-Image Pre-training (BLIP) model has better feature expression ability and better model interpretability than CLIP, the BLIP model is used as the above-mentioned feature expression model in the embodiments of this specification. Figure 2 shown.

[0071] Figure 2 A schematic diagram of automatic classification provided by an embodiment of this specification. Since the reference product information is pre-set and the product category to which the reference product information belongs is also pre-marked, the reference product information can be pre-input into the BLIP model in an offline state to obtain the reference product features corresponding to the reference product information (including text information and image information), and the reference product features are stored. Similarly, the category information of the product category to which the reference product information belongs can also be pre-input into the BLIP model in an offline state to obtain the category features and store the category features.

[0072] In the online state, when the information of goods to be classified uploaded by the merchant is obtained, the information of goods to be classified can be input into the BLIP model in real time to obtain the characteristics of the goods to be classified, and the first matching degree is determined based on the characteristics of the goods to be classified and the stored characteristics of each reference product, and the recalled product information is determined based on the first matching degree, and then the second matching degree is determined based on the characteristics of the goods to be classified and the category characteristics of the stored recalled product category, and the final matching degree is determined based on the first matching degree and the second matching degree, and the product category to which the information of goods to be classified belongs is determined accordingly.

[0073] The above is an automatic classification method provided in an embodiment of this specification. Based on the same idea, this specification also provides corresponding devices, storage media and electronic devices.

[0074] Figure 3 A schematic diagram of an automatic classification device provided in an embodiment of this specification, the device comprising:

[0075] The acquisition module 301 is used to acquire the commodity information to be classified uploaded by the merchant;

[0076] A first matching module 302, configured to determine a first matching degree between the commodity information to be classified and each reference commodity information according to the commodity information to be classified and each preset reference commodity information;

[0077] A product recall module 303 is used to determine, according to the first matching degree, a plurality of reference product information matching the product information to be classified as recalled product information;

[0078] The second matching module 304 is used to, for each piece of recalled commodity information, use the commodity category to which the recalled commodity information belongs as the recalled commodity category, and determine a second matching degree between the commodity information to be classified and the recalled commodity category according to the category information of the recalled commodity category and the commodity information to be classified;

[0079] A final matching module 305, configured to determine a final matching degree between the to-be-classified product information and the recalled product category according to the second matching degree and the first matching degree between the to-be-classified product information and the recalled product information;

[0080] The classification module 306 is used to determine the commodity category to which the commodity information to be classified belongs in each recalled commodity category according to the final matching degree between the commodity information to be classified and each recalled commodity category.

[0081] Optionally, the first matching module 302 is used to input the commodity information to be classified into a pre-trained feature representation model to obtain the commodity features to be classified corresponding to the commodity information to be classified output by the feature representation model; and determine the first matching degree between the commodity information to be classified and each reference commodity information according to the commodity features to be classified and the reference commodity features corresponding to each reference commodity information that are pre-saved.

[0082] Optionally, the second matching module 304 is used to input the commodity information to be classified into a pre-trained feature representation model to obtain the commodity features to be classified corresponding to the commodity information to be classified output by the feature representation model; and determine the second matching degree between the commodity information to be classified and the recalled commodity category according to the category features corresponding to the commodity features to be classified and the category information of each recalled commodity category saved in advance.

[0083] Optionally, the information of commodities to be classified includes text information and image information of the commodities to be classified;

[0084] The reference product information includes text information and image information of the reference product;

[0085] The category information of the recalled product category includes text description information of the recalled product category;

[0086] The feature characterization model includes a BLIP model.

[0087] Optionally, the product recall module 303 is used to select a specified number of reference product information in descending order of the first matching degree as reference product information matching the product information to be classified.

[0088] Optionally, the final matching module 305 is configured to weight the second matching degree and the first matching degree between the to-be-classified product information and the recalled product information to obtain a final matching degree between the to-be-classified product information and the recalled product category.

[0089] Optionally, the classification module 306 is configured to determine, among the recalled product categories, the recalled product category with the highest final matching degree with the product information to be classified as the product category to which the product information to be classified belongs.

[0090] This specification also provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, it can be used to perform the above Figure 1 Automatic classification method provided.

[0091] based on Figure 1 Based on the Figure 1 The automatic classification method shown in the present specification also provides Figure 4 The structural diagram of the electronic device shown in FIG. Figure 4 At the hardware level, the electronic device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 The automatic classification method described.

[0092] Of course, in addition to software implementation, this specification does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc., that is to say, the executor of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0093] In the 1990s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the method flow). However, with the development of technology, many improvements to the method flow today can be regarded as direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask a chip manufacturer to design and produce a dedicated integrated circuit chip. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0094] The controller can be implemented in any appropriate manner, for example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in a purely computer-readable program code manner, the controller can be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, this controller can be considered as a hardware component, and the devices included therein for implementing various functions can also be regarded as structures within the hardware component. Or even, the devices for implementing various functions can be regarded as both software modules for implementing the method and structures within the hardware component.

[0095] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0096] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0097] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this specification 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.) that contain computer-usable program code.

[0098] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes 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.

[0099] 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.

[0100] 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 The steps for the functions specified in one or more boxes.

[0101] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0102] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0103] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0104] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0105] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, this specification 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.

[0106] This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0107] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0108] The above description is only an embodiment of the present specification and is not intended to limit the present specification. For those skilled in the art, the present specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification shall be included in the scope of the claims of the present specification.

Claims

1. An automatic classification method, the method comprising: Get the information of the products to be classified uploaded by the merchant; Determining a first matching degree between the to-be-classified commodity information and each reference commodity information according to the to-be-classified commodity information and each preset reference commodity information; According to the first matching degree, determining a plurality of reference commodity information matching the commodity information to be classified as the recalled commodity information; For each recalled product information, a product category to which the recalled product information belongs that is pre-marked is used as the recalled product category, and a second matching degree between the product information to be classified and the recalled product category is determined according to the category information of the recalled product category and the product information to be classified; and a final matching degree between the product information to be classified and the recalled product category is determined according to the second matching degree and the first matching degree between the product information to be classified and the recalled product information; According to the final matching degree between the commodity information to be classified and each recalled commodity category, the commodity category to which the commodity information to be classified belongs is determined in each recalled commodity category.

2. The method according to claim 1, determining the first matching degree between the to-be-classified commodity information and each reference commodity information according to the to-be-classified commodity information and each preset reference commodity information, specifically comprising: Inputting the commodity information to be classified into a pre-trained feature representation model to obtain the commodity features to be classified corresponding to the commodity information to be classified output by the feature representation model; A first matching degree between the commodity information to be classified and each reference commodity information is determined according to the commodity characteristics to be classified and the pre-stored reference commodity characteristics corresponding to each reference commodity information.

3. The method according to claim 1, determining a second matching degree between the to-be-classified product information and the recalled product category according to the category information of the recalled product category and the to-be-classified product information, specifically comprising: Inputting the commodity information to be classified into a pre-trained feature representation model to obtain the commodity features to be classified corresponding to the commodity information to be classified output by the feature representation model; A second matching degree between the to-be-classified commodity information and the recalled commodity category is determined according to the to-be-classified commodity characteristics and the category characteristics corresponding to the pre-stored category information of each recalled commodity category.

4. The method according to claim 2 or 3, wherein the information of the commodities to be classified includes text information and image information of the commodities to be classified; The reference product information includes text information and image information of the reference product; The category information of the recalled product category includes text description information of the recalled product category; The feature characterization model includes a BLIP model.

5. The method according to claim 1, determining a plurality of reference commodity information matching the commodity information to be classified according to the first matching degree, specifically comprising: A specified number of reference product information are selected in order from high to low in the first matching degree as reference product information matching the product information to be classified.

6. The method according to claim 1, determining the final matching degree between the product information to be classified and the recalled product category according to the second matching degree and the first matching degree between the product information to be classified and the recalled product information, specifically comprising: The second matching degree is weighted with the first matching degree between the to-be-classified product information and the recalled product information to obtain a final matching degree between the to-be-classified product information and the recalled product category.

7. The method according to claim 1, determining the commodity category to which the commodity information to be classified belongs in each recalled commodity category according to the final matching degree between the commodity information to be classified and each recalled commodity category, specifically comprising: Among the recalled commodity categories, the recalled commodity category with the highest final matching degree with the commodity information to be classified is determined as the commodity category to which the commodity information to be classified belongs.

8. An automatic classification device, comprising: The acquisition module is used to obtain the information of commodities to be classified uploaded by merchants; A first matching module, configured to determine a first matching degree between the commodity information to be classified and each reference commodity information according to the commodity information to be classified and each preset reference commodity information; A product recall module, configured to determine, according to the first matching degree, a plurality of reference product information matching the product information to be classified as recalled product information; a second matching module, for each piece of recalled commodity information, using a pre-marked commodity category to which the recalled commodity information belongs as the recalled commodity category, and determining a second matching degree between the to-be-classified commodity information and the recalled commodity category according to the category information of the recalled commodity category and the to-be-classified commodity information; a final matching module, configured to determine a final matching degree between the commodity information to be classified and the category of the recalled commodity according to the second matching degree and the first matching degree between the commodity information to be classified and the recalled commodity information; The classification module is used to determine the commodity category to which the commodity information to be classified belongs in each recalled commodity category according to the final matching degree between the commodity information to be classified and each recalled commodity category.

9. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the method according to any one of claims 1 to 7.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 7 when executing the program.

Images