Cross-border e-commerce commodity intelligent classification method and device based on AI algorithm

Abstract

The present invention provides a method and device for intelligent classification of cross-border e-commerce products based on an AI algorithm, wherein the method includes: obtaining search hot words and determining the mean of each preset dimension to form a data set to determine the target dimension, then updating the data classification standard based on the target dimension, and finally classifying the products based on the updated data classification standard and the AI ​​algorithm. The beneficial effects of the present invention: the real-time classification of products according to search hot words is realized, and the search hot words in different regions are different, and an intelligent classification method is set for the designated area, which can greatly meet the intelligent classification of cross-border e-commerce products in the designated area, so that users in the designated area can search for ideal products faster, improving the experience of cross-border e-commerce and users.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a method and device for intelligent classification of cross-border e-commerce products based on an AI algorithm. Background Art

[0002] With the rapid development of short video platforms such as Douyin and Kuaishou, many new products have received great exposure. However, due to cultural differences in cross-border e-commerce and the continuous updating and iteration of products, the existing classification method of e-commerce platforms is a unified classification template, which cannot intelligently classify cross-border e-commerce products and cannot meet the needs of cross-border e-commerce. Therefore, there is an urgent need for an intelligent classification method for cross-border e-commerce products based on AI algorithms. Summary of the invention

[0003] The main purpose of the present invention is to provide a method and device for intelligent classification of cross-border e-commerce products based on AI algorithm, aiming to solve the problem that the existing classification method of e-commerce platforms is a unified classification template and cannot intelligently classify cross-border e-commerce products.

[0004] The present invention provides a cross-border e-commerce commodity intelligent classification method based on AI algorithm, comprising:

[0005] Obtain multiple hot search words that appear in a specified area on a cross-border e-commerce platform, as well as product information corresponding to each of the hot search words;

[0006] Convert the search hot words and corresponding product information into hot word vectors according to Word2Vec;

[0007] Calculate the dimension mean of each preset dimension in the hot word vector corresponding to each product information;

[0008] Extract the dimension means corresponding to each of the product information according to each preset dimension to obtain a value set corresponding to each preset dimension;

[0009] Calculating the data fluctuation value of each of the numerical value sets; wherein the data fluctuation value is used to reflect the fluctuation situation in the numerical value set;

[0010] Select the preset dimension whose data fluctuation value is greater than the preset indicator value as the target dimension;

[0011] Acquire a first data classification standard before updating, and update the first data classification standard based on the target dimension to obtain a second data classification standard;

[0012] Each product is classified based on the second data classification standard and the preset AI algorithm.

[0013] Furthermore, the step of classifying each product based on the second data classification standard and the preset AI algorithm includes:

[0014] Obtain multiple product information related to the target dimension from the cross-border e-commerce platform, and use a preset linear classifier to set multiple classification functions for the same target dimension ;in, ,and , represents a constant related to the second data classification standard, Represents the bias of the t-th classification function, t is a positive integer, represents a weight vector, and the dimension of the weight vector is the same as the number of value sets, represents the t-th classification function, Indicates product information. are the preset parameters;

[0015] Calculate the Euclidean distance between each classification function and each product information, extract the maximum Euclidean distance and the minimum Euclidean distance of each classification function, and use the difference between the two as the classification distance of the corresponding classification function;

[0016] According to the formula Calculate the transformation parameters for the classification distance between two classification functions of the same target dimension ;in, represents the nth classification distance, Indicates based on The preset calculation function of

[0017] Determine the transformation parameters Is it within the preset range?

[0018] If they are within the preset range, each product is classified based on the multiple classification functions.

[0019] Furthermore, the step of calculating the Euclidean distance between each classification function and each product information includes:

[0020] Input the data fluctuation value and the plurality of the search hot words into a preset function acquisition model to obtain a corresponding conversion function; wherein the function acquisition model is trained according to different search hot words and index values ​​as inputs and according to the corresponding conversion function as output;

[0021] Performing spatial mapping on the product information according to the conversion function to obtain product data after each product information is mapped;

[0022] The Euclidean distance between each classification function and the corresponding product information is calculated based on the product data.

[0023] Furthermore, the determining of the transformation parameters After the step of whether it is within the preset range, it also includes:

[0024] If it is not within the preset range, the weight vector in the classification function is adjusted until the transformation parameter is within the preset range.

[0025] Furthermore, after the step of classifying each product based on the second data classification standard and the preset AI algorithm, the method further includes:

[0026] Get the number of traded orders for each product within a specified time period;

[0027] Setting the priority order of each product according to the number of transactions;

[0028] Receive the user's search content, and generate a push list according to the search content and priority order.

[0029] Furthermore, before the step of obtaining multiple hot search words appearing in a specified area on the cross-border e-commerce platform and the product information corresponding to each of the hot search words, the step further includes:

[0030] Monitor whether the number of times the search term appears within a set time period is greater than a preset value;

[0031] If the number of times a search term appears within a set time period is greater than a preset value, the search term is recorded as a search hot word.

[0032] Furthermore, after the step of classifying each product based on the second data classification standard and the preset AI algorithm, the method further includes:

[0033] Get the category information of each product after classification;

[0034] The category information is labeled, and corresponding labels are assigned to corresponding products.

[0035] The present invention also provides a cross-border e-commerce commodity intelligent classification device based on AI algorithm, comprising:

[0036] The first acquisition module is used to acquire multiple hot search words that appear in a specified area on the cross-border e-commerce platform, and product information corresponding to each of the hot search words;

[0037] A conversion module, used to convert the search hot words and corresponding product information into hot word vectors according to Word2Vec;

[0038] A first calculation module is used to calculate the dimension mean of each preset dimension in the hot word vector corresponding to each product information;

[0039] An extraction module, used to extract the dimension means corresponding to each of the product information according to each preset dimension, to obtain a value set corresponding to each preset dimension;

[0040] A second calculation module is used to calculate the data fluctuation value of each of the numerical value sets; wherein the data fluctuation value is used to reflect the fluctuation situation in the numerical value set;

[0041] A selection module is used to select a preset dimension whose data fluctuation value is greater than a preset index value as a target dimension;

[0042] A second acquisition module is used to acquire the first data classification standard before updating, and update the first data classification standard based on the target dimension to obtain a second data classification standard;

[0043] A classification module is used to classify each product based on the second data classification standard and a preset AI algorithm.

[0044] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0045] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

[0046] The beneficial effects of the present invention are as follows: by obtaining the search hot words and determining the mean of each preset dimension, a data set is formed to determine the target dimension, and then the data classification standard is updated based on the target dimension. Finally, the products are classified based on the updated data classification standard and AI algorithm. Thus, the real-time classification of products according to the search hot words is realized, and the search hot words in different regions are different. An intelligent classification method is set for the designated area, which can greatly meet the intelligent classification of cross-border e-commerce products in the designated area, so that users in the designated area can search for ideal products faster, improving the experience of cross-border e-commerce and users. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a flowchart of a cross-border e-commerce commodity intelligent classification method based on AI algorithm according to an embodiment of the present invention;

[0048] Figure 2 This is a schematic block diagram of the structure of a cross-border e-commerce commodity intelligent classification device based on an AI algorithm according to an embodiment of the present invention;

[0049] Figure 3 A schematic block diagram of the structure of a computer device according to an embodiment of the present application.

[0050] 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

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly. The connection can be a direct connection or an indirect connection.

[0053] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and B can represent three situations: A exists alone, A and B exist at the same time, and B exists alone.

[0054] In addition, in the present invention, descriptions such as "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0055] Reference Figure 1 The present invention proposes a cross-border e-commerce commodity intelligent classification method based on AI algorithm, comprising:

[0056] S1: Obtain multiple hot search words that appear in a specified area on a cross-border e-commerce platform, as well as product information corresponding to each of the hot search words;

[0057] S2: Convert the search hot words and corresponding product information into hot word vectors according to Word2Vec;

[0058] S3: Calculate the dimension mean of each preset dimension in the hot word vector corresponding to each product information;

[0059] S4: extracting the dimension means corresponding to each of the product information according to each preset dimension to obtain a value set corresponding to each preset dimension;

[0060] S5: Calculate the data fluctuation value of each of the numerical value sets; wherein the data fluctuation value is used to reflect the fluctuation situation in the numerical value set;

[0061] S6: Select a preset dimension whose data fluctuation value is greater than a preset index value as the target dimension;

[0062] S7: Acquire the first data classification standard before updating, and update the first data classification standard based on the target dimension to obtain a second data classification standard;

[0063] S8: Classify each product based on the second data classification standard and the preset AI algorithm.

[0064] As described in step S1 above, multiple search hot words that appear in a specified area on the cross-border e-commerce platform, as well as product information corresponding to each of the search hot words, are obtained. Search hot words and their corresponding product information will continue to change with time, market trends and changes in consumer demand. Therefore, search hot words can be defined as search hot words by the number of times the search words appear in a period of time, and the products purchased by users based on the search hot words will have their product information as the product information corresponding to the search hot words. It should be noted that one search hot word can correspond to multiple product information, and one product information can also correspond to multiple search hot words. The designated area is a pre-set area, such as Africa, or a specific country, region, etc., which can be specified in advance.

[0065] As described in step S2 above, the search hot words and the corresponding product information are converted into hot word vectors according to Word2Vec. Word2Vec is a natural language processing technology that can convert words (or phrases) into vector form and capture the semantic relationship between words. Specifically, product information can be processed by natural language processing technology or deep learning models to form word vectors of products, and then the word vectors and search hot words are input into Word2Vec to obtain corresponding hot word vectors.

[0066] As described in the above step S3, since there is a corresponding relationship between the hot word vector and the product information, one product information generally corresponds to multiple hot word vectors. Therefore, the values ​​of each preset dimension in the hot word vector corresponding to a product information can be summed and averaged to obtain the dimensional mean of each preset dimension in the corresponding product information.

[0067] As described in step S4 above, the dimensional means corresponding to each of the product information are extracted according to each preset dimension to obtain a numerical set corresponding to each preset dimension. For each preset dimension, the dimensional mean of all the product information on the dimension is extracted to form a numerical set.

[0068] As described in the above step S5, the data fluctuation value of each of the numerical value sets is calculated; wherein, the data fluctuation value is used to reflect the fluctuation situation in the numerical value set, and specifically, the data fluctuation value can be calculated using statistics such as standard deviation and variance.

[0069] As described in step S6 above, a preset dimension whose data fluctuation value is greater than a preset index value is selected as the target dimension. If the data fluctuation on a certain dimension is large, it means that they are of great significance for distinguishing different categories of products, so they are used as target dimensions.

[0070] As described in step S7 above, the first data classification standard before the update is obtained, and the first data classification standard is updated based on the target dimension to obtain the second data classification standard. The first data classification standard is a classification standard pre-stored in the platform, and the first data classification standard is updated based on the target dimension to obtain the second data classification standard. The update method can be to adjust the classification boundary, add new classification features, etc., or to provide a basis for subsequent classification.

[0071] As described in step S8 above, each product is classified based on the second data classification standard and the preset AI algorithm. Each product is classified using the updated second data classification standard and the preset AI algorithm (such as support vector machine, neural network, etc.). The AI ​​algorithm will divide the products into different categories according to the vector representation and classification standard of the product information. Thereby, real-time classification of products according to search hot words is realized, and the search hot words in different regions are different. An intelligent classification method is set for the designated area, which can greatly meet the intelligent classification of cross-border e-commerce products in the designated area, so that users in the designated area can search for ideal products faster, improving the experience of cross-border e-commerce and users.

[0072] In one embodiment, step S8 of classifying each product based on the second data classification standard and a preset AI algorithm includes:

[0073] S801: Acquire multiple product information related to the target dimension from the cross-border e-commerce platform, and set multiple classification functions using a preset linear classifier for the same target dimension ;in, ,and , represents a constant related to the second data classification standard, Represents the bias of the t-th classification function, t is a positive integer, represents a weight vector, and the dimension of the weight vector is the same as the number of value sets, represents the t-th classification function, Indicates product information. are the preset parameters;

[0074] S802: Calculate the Euclidean distance between each classification function and each product information, extract the maximum Euclidean distance and the minimum Euclidean distance of each classification function, and use the difference between the two as the classification distance of the corresponding classification function;

[0075] S803: According to the formula Calculate the transformation parameters for the classification distance between two classification functions of the same target dimension ;in, represents the nth classification distance, Indicates based on The preset calculation function of

[0076] S804: Determine the transformation parameters Is it within the preset range?

[0077] S805: If they are within the preset range, classify each product based on the multiple classification functions.

[0078] As described in the above steps S801-S805, the classification of each product is realized, multiple product information related to the target dimension is obtained from the cross-border e-commerce platform, and multiple classification functions are set for the same target dimension using a preset linear classifier, wherein the linear classifier can be Bayesian classification, LR, linear regression, single-layer perceptron, SVM (linear kernel), etc. The principle is to separate them by setting multiple hyperplanes, each classification function can be regarded as a hyperplane, and the space between the hyperplanes is the same category. The Euclidean distance between each classification function and each product information is calculated, and the maximum Euclidean distance and the minimum Euclidean distance of each classification function are extracted, and the difference between the two is used as the classification distance of the corresponding classification function, wherein the classification distance needs to satisfy that the corresponding transformation parameter belongs to the preset range, that is, if it is no longer within the preset range, it means that the classification effect is not good, and the better the classification effect, the higher the classification accuracy, the better the classification of products, and the display of similar products, so as to avoid too many similar products, which is inconvenient to display to customers.

[0079] In one embodiment, the step S802 of calculating the Euclidean distance between each classification function and each product information includes:

[0080] S8021: Input the data fluctuation value and the plurality of search hot words into a preset function acquisition model to obtain a corresponding conversion function; wherein the function acquisition model is trained according to different search hot words and index values ​​as inputs and the corresponding conversion function as output;

[0081] S8022: Perform spatial mapping on the product information according to the conversion function to obtain product data after each product information is mapped;

[0082] S8023: Calculate the Euclidean distance between each classification function and the corresponding product information based on the product data.

[0083] As described in the above steps S8021-S8023, the calculation of the Euclidean distance, that is, the preset function corresponding to the data fluctuation value and the search hot word is realized. It is not a linear relationship. Therefore, the corresponding relationship can be constructed through a neural network model, and the conversion function is obtained according to the data fluctuation value and the search hot word. Common conversion functions include polynomial conversion function, linear conversion function, Gauss radial basis conversion function, etc.

[0084] In one embodiment, the determining of the transformation parameters After step S804 of determining whether the data is within a preset range, the following steps are further included:

[0085] S8051: If it is not within the preset range, adjust the weight vector in the classification function until the transformation parameter is within the preset range.

[0086] As described in step S8051 above, the classification function is adjusted. If it is not within the preset range, it means that the weight vector setting is unreasonable and needs to be adjusted. The adjustment range can be adjusted according to the transformation parameter value, that is, the adjustment amplitude and direction are determined according to the value of the transformation parameter, so that the weight parameter can be adjusted more quickly.

[0087] In one embodiment, after the step S8 of classifying each product based on the second data classification standard and the preset AI algorithm, the method further includes:

[0088] S901: Obtain the number of traded orders for each product within a specified time period;

[0089] S902: Setting the priority order of each product according to the number of completed orders;

[0090] S903: Receive the user's search content, and generate a push list according to the search content and priority order.

[0091] As described in the above steps S901-S903, the generation of the push list is realized. In order to provide a better user experience, the number of transactions for each product within a specified time period can be obtained, and the priority order of the products can be set according to the number of transactions. Products with more transactions are given a higher priority, and then related products are determined based on the user's search content, and then a push list is generated based on the priority order.

[0092] In one embodiment, before the step S1 of obtaining multiple hot search words appearing in a specified area on the cross-border e-commerce platform and product information corresponding to each of the hot search words, the step further includes:

[0093] S001: monitor whether the number of times the search term appears within a set time period is greater than a preset value;

[0094] S002: If the number of times the search word appears within a set time period is greater than a preset value, the search word is recorded as a search hot word.

[0095] As described in the above steps S001-S002, the marking of hot search words is realized, that is, monitoring whether the number of times the search word appears within the set time period is greater than the preset value. If it is greater than the preset value, it is recorded as a hot search word. It should be noted that the set time period is the most recent time period, that is, real-time monitoring is performed the day after tomorrow. When the number of times the search word appears is higher than the preset value within a period of time, it is identified as a hot search word.

[0096] In one embodiment, after the step S8 of classifying each product based on the second data classification standard and the preset AI algorithm, the method further includes:

[0097] S911: Obtain category information of each product after classification;

[0098] S912: Label the category information and assign corresponding labels to corresponding products.

[0099] As described in the above steps S911-S912, labeling of each product is achieved, that is, the category information of each product after classification is obtained, the category information is labeled, and the corresponding label is assigned to the corresponding product.

[0100] Reference Figure 2 The present invention also provides a cross-border e-commerce commodity intelligent classification device based on AI algorithm, comprising:

[0101] The first acquisition module 10 is used to acquire multiple hot search words that appear in a specified area on the cross-border e-commerce platform, and product information corresponding to each of the hot search words;

[0102] A conversion module 20, for converting the search hot words and corresponding product information into hot word vectors according to Word2Vec;

[0103] A first calculation module 30 is used to calculate the dimension mean of each preset dimension in the hot word vector corresponding to each product information;

[0104] An extraction module 40 is used to extract the dimension means corresponding to each of the product information according to each preset dimension to obtain a value set corresponding to each preset dimension;

[0105] A second calculation module 50 is used to calculate the data fluctuation value of each of the numerical value sets; wherein the data fluctuation value is used to reflect the fluctuation situation in the numerical value set;

[0106] A selection module 60, configured to select a preset dimension whose data fluctuation value is greater than a preset index value as a target dimension;

[0107] A second acquisition module 70 is used to acquire the first data classification standard before updating, and update the first data classification standard based on the target dimension to obtain a second data classification standard;

[0108] The classification module 80 is used to classify each product based on the second data classification standard and a preset AI algorithm.

[0109] In one embodiment, the classification module 80 includes:

[0110] The product information acquisition submodule is used to obtain multiple product information related to the target dimension from the cross-border e-commerce platform, and use a preset linear classifier to set multiple classification functions for the same target dimension. ;in, ,and , represents a constant related to the second data classification standard, Represents the bias of the t-th classification function, t is a positive integer, represents a weight vector, and the dimension of the weight vector is the same as the number of value sets, represents the t-th classification function, Indicates product information. are the preset parameters;

[0111] The Euclidean distance calculation submodule is used to calculate the Euclidean distance between each classification function and each product information, and extract the maximum Euclidean distance and the minimum Euclidean distance of each classification function, and use the difference between the two as the classification distance of the corresponding classification function;

[0112] The transformation parameter calculation submodule is used to calculate the Calculate the transformation parameters for the classification distance between two classification functions of the same target dimension ;in, represents the nth classification distance, Indicates based on The preset calculation function of

[0113] The transformation parameter judgment submodule is used to judge the transformation parameters Is it within the preset range?

[0114] The classification submodule is used to classify each product based on the multiple classification functions if it is within a preset range.

[0115] In one embodiment, the Euclidean distance calculation submodule includes:

[0116] An input unit, used to input the data fluctuation value and the plurality of the search hot words into a preset function acquisition model to obtain a corresponding conversion function; wherein the function acquisition model is trained according to different search hot words and index values ​​as inputs and according to the corresponding conversion function as output;

[0117] A mapping unit, used to perform spatial mapping on the product information according to the conversion function to obtain product data after each product information is mapped;

[0118] A calculation unit is used to calculate the Euclidean distance between each classification function and the corresponding product information based on the product data.

[0119] In one embodiment, the classification module 80 further includes:

[0120] The adjustment submodule is used to adjust the weight vector in the classification function if it is not within the preset range, until the transformation parameter is within the preset range.

[0121] In one embodiment, the cross-border e-commerce commodity intelligent classification device based on AI algorithm further includes:

[0122] The transaction number acquisition module is used to obtain the transaction number of each product within a specified time period;

[0123] A priority order setting module, used to set the priority order of each product according to the number of transaction orders;

[0124] The push list generation module is used to receive the user's search content and generate a push list according to the search content and priority order.

[0125] In one embodiment, the cross-border e-commerce commodity intelligent classification device based on AI algorithm further includes:

[0126] A monitoring module, used to monitor whether the number of times the search term appears within a set time period is greater than a preset value;

[0127] The marking module is used to record the search term as a hot search term if the number of times the search term appears within a set time period is greater than a preset value.

[0128] In one embodiment, the cross-border e-commerce commodity intelligent classification device based on AI algorithm further includes:

[0129] Category information acquisition module, used to obtain category information of each product after classification;

[0130] The labeling module is used to label the category information and assign corresponding labels to corresponding products.

[0131] The beneficial effects of the present invention are as follows: by obtaining the search hot words and determining the mean of each preset dimension, a data set is formed to determine the target dimension, and then the data classification standard is updated based on the target dimension, and finally the product is classified based on the updated data classification standard and AI algorithm. Thus, the real-time classification of products according to the search hot words is realized, and the search hot words in different regions are different. The intelligent classification method is set for the designated area, which can greatly meet the intelligent classification of cross-border e-commerce products in the designated area, so that users in the designated area can search for ideal products faster, improving the experience of cross-border e-commerce and users.

[0132] Reference Figure 3 In an embodiment of the present application, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store various product information, etc. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the cross-border e-commerce commodity intelligent classification method based on the AI ​​algorithm described in any of the above embodiments can be implemented.

[0133] Those skilled in the art will understand that Figure 3 The structure shown in is merely a block diagram of a portion of the structure related to the present application solution and does not constitute a limitation on the computer device to which the present application solution is applied.

[0134] An embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the cross-border e-commerce product intelligent classification method based on the AI ​​algorithm described in any of the above embodiments can be implemented.

[0135] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided in this application and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0136] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method 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, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, device, article or method including the element.

[0137] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0138] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0139] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention shall be included in the scope of the claims of the present invention.

Claims

1. A cross-border e-commerce commodity intelligent classification method based on AI algorithm, characterized in that: include: Obtain multiple hot search words that appear in a specified area on a cross-border e-commerce platform, as well as product information corresponding to each of the hot search words; Convert the search hot words and corresponding product information into hot word vectors according to Word2Vec; Calculate the dimension mean of each preset dimension in the hot word vector corresponding to each product information; Extract the dimension means corresponding to each of the product information according to each preset dimension to obtain a value set corresponding to each preset dimension; Calculating the data fluctuation value of each of the numerical value sets; wherein the data fluctuation value is used to reflect the fluctuation situation in the numerical value set; Select the preset dimension whose data fluctuation value is greater than the preset indicator value as the target dimension; Acquire a first data classification standard before updating, and update the first data classification standard based on the target dimension to obtain a second data classification standard; Classify each product based on the second data classification standard and a preset AI algorithm; The step of classifying each product based on the second data classification standard and the preset AI algorithm includes: Obtain multiple product information related to the target dimension from the cross-border e-commerce platform, and use a preset linear classifier to set multiple classification functions for the same target dimension ;in, , represents a constant related to the second data classification standard, Represents the bias of the t-th classification function, t is a positive integer, represents a weight vector, and the dimension of the weight vector is the same as the number of value sets, represents the t-th classification function, Indicates product information. are the preset parameters; Calculate the Euclidean distance between each classification function and each product information, extract the maximum Euclidean distance and the minimum Euclidean distance of each classification function, and use the difference between the two as the classification distance of the corresponding classification function; According to the formula Calculate the transformation parameters for the classification distance between two classification functions of the same target dimension ;in, represents the nth classification distance, Indicates based on The preset calculation function of Determine the transformation parameters Is it within the preset range? If they are within the preset range, each product is classified based on the multiple classification functions.

2. The cross-border e-commerce commodity intelligent classification method based on AI algorithm as claimed in claim 1, characterized in that: The step of calculating the Euclidean distance between each classification function and each product information includes: Input the data fluctuation value and the plurality of the search hot words into a preset function acquisition model to obtain a corresponding conversion function; wherein the function acquisition model is trained according to different search hot words and index values ​​as inputs and according to the corresponding conversion function as output; Performing spatial mapping on the product information according to the conversion function to obtain product data after each product information is mapped; The Euclidean distance between each classification function and the corresponding product information is calculated based on the product data.

3. The cross-border e-commerce commodity intelligent classification method based on AI algorithm as claimed in claim 1, characterized in that: The determining of the transformation parameters After the step of whether it is within the preset range, it also includes: If it is not within the preset range, the weight vector in the classification function is adjusted until the transformation parameter is within the preset range.

4. The cross-border e-commerce commodity intelligent classification method based on AI algorithm as claimed in claim 1, characterized in that: After the step of classifying each product based on the second data classification standard and the preset AI algorithm, the method further includes: Get the number of traded orders for each product within a specified time period; Setting the priority order of each product according to the number of transactions; Receive the user's search content, and generate a push list according to the search content and priority order.

5. The cross-border e-commerce commodity intelligent classification method based on AI algorithm as claimed in claim 1, characterized in that: Before the step of obtaining multiple hot search words appearing in a specified area on the cross-border e-commerce platform and product information corresponding to each of the hot search words, the step further includes: Monitor whether the number of times the search term appears within a set time period is greater than a preset value; If the number of times a search term appears within a set time period is greater than a preset value, the search term is recorded as a search hot word.

6. The cross-border e-commerce commodity intelligent classification method based on AI algorithm as claimed in claim 1, characterized in that: After the step of classifying each product based on the second data classification standard and the preset AI algorithm, the method further includes: Get the category information of each product after classification; The category information is labeled, and corresponding labels are assigned to corresponding products.

7. A cross-border e-commerce commodity intelligent classification device based on AI algorithm, characterized in that: include: The first acquisition module is used to acquire multiple hot search words that appear in a specified area on the cross-border e-commerce platform, and product information corresponding to each of the hot search words; A conversion module, used to convert the search hot words and corresponding product information into hot word vectors according to Word2Vec; A first calculation module is used to calculate the dimension mean of each preset dimension in the hot word vector corresponding to each product information; An extraction module, used to extract the dimension means corresponding to each of the product information according to each preset dimension, to obtain a value set corresponding to each preset dimension; A second calculation module is used to calculate the data fluctuation value of each of the numerical value sets; wherein the data fluctuation value is used to reflect the fluctuation situation in the numerical value set; A selection module is used to select a preset dimension whose data fluctuation value is greater than a preset index value as a target dimension; A second acquisition module is used to acquire the first data classification standard before updating, and update the first data classification standard based on the target dimension to obtain a second data classification standard; A classification module, used to classify each product based on the second data classification standard and a preset AI algorithm; The classification module comprises: The product information acquisition submodule is used to obtain multiple product information related to the target dimension from the cross-border e-commerce platform, and use a preset linear classifier to set multiple classification functions for the same target dimension. ;in, , represents a constant related to the second data classification standard, Represents the bias of the t-th classification function, t is a positive integer, represents a weight vector, and the dimension of the weight vector is the same as the number of value sets, represents the t-th classification function, Indicates product information. are the preset parameters; The Euclidean distance calculation submodule is used to calculate the Euclidean distance between each classification function and each product information, and extract the maximum Euclidean distance and the minimum Euclidean distance of each classification function, and use the difference between the two as the classification distance of the corresponding classification function; The transformation parameter calculation submodule is used to calculate the Calculate the transformation parameters for the classification distance between two classification functions of the same target dimension ;in, represents the nth classification distance, Indicates based on The preset calculation function of The transformation parameter judgment submodule is used to judge the transformation parameters Is it within the preset range? The classification submodule is used to classify each product based on the multiple classification functions if it is within a preset range.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images