A mounting method and device based on combined goods

Abstract

The present application discloses a mounting method and device based on combination products, the method comprising: rewriting the main product name in the combination product according to the four-level category requirements based on the acquired basic data of the combination product and the product category basic data, to obtain the four-level standardized product name corresponding to the combination product and the four-level category; recalling the four-level category through the first category recall method and the second category recall method according to the four-level standardized product name and the product category basic data, respectively, to recall the first candidate fourth-level category data and the second candidate fourth-level category data corresponding to the four-level standardized product name; mounting the combination product on the fourth-level category data that matches the four-level standardized product name determined based on the first candidate fourth-level category data and the second candidate fourth-level category data, as well as the basic data of the combination product, to achieve end-to-end mapping of the combination product to the four-level category, and to improve the mounting coverage and efficiency of the combination product to the four-level category.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and more particularly to a method and device for mounting a combination of products. The present application also relates to a computer storage medium and an electronic device. Background Art

[0002] With the widespread adoption of mobile internet and the continuous development of computer technology, the scope of services offered by lifestyle service applications has become increasingly broad, encompassing nearly every aspect of daily life, from food delivery and fresh produce delivery to beauty salons and home maintenance. This has led to an explosive growth in the number of products available on platforms, with millions, even tens of millions, of products available. For example, in the restaurant market, merchants need to upload extensive data on their food items. Different merchants have different service areas, and different restaurant areas offer different types of food, resulting in a massive amount of product data. With such a large volume of products, organizing and categorizing products is crucial. For example, for users, categorizing and managing products improves selection efficiency and user experience. For application service platforms, categorizing products improves data management efficiency, providing a clearer data structure and ensuring optimal user service.

[0003] With the increasing diversification and personalization of user needs, the complexity of product management has been further exacerbated. Summary of the Invention

[0004] The present application provides a method for mounting combined commodities to solve the problem in the prior art that accurate mounting of commodities cannot be achieved for dynamic commodity combinations.

[0005] In a first aspect, the present application provides a method for mounting a combination of products, comprising:

[0006] Based on the acquired basic data of the combination product and the basic data of the product category, the main product name in the combination product is rewritten according to the requirements of the fourth-level category to obtain the fourth-level standardized product name corresponding to the combination product and the fourth-level category;

[0007] According to the fourth-level standardized product name and the product category basic data, a fourth-level category recall is performed using a first category recall method and a second category recall method, respectively, to recall first candidate fourth-level category data and second candidate fourth-level category data corresponding to the fourth-level standardized product name;

[0008] The combination product is mounted on the fourth-level category data that matches the fourth-level standardized product name and is determined based on the first candidate fourth-level category data, the second candidate fourth-level category data, and the basic data of the combination product.

[0009] In some embodiments, the method of rewriting the main product name in the combination product according to the requirements of the fourth-level category based on the acquired basic data of the combination product and the basic data of the product category to obtain the fourth-level standardized product name corresponding to the combination product and the fourth-level category includes:

[0010] Splitting the combination product according to the basic data of the combination product to obtain sub-products included in the combination product;

[0011] Determine the sub-item with the highest priority in the combination of sub-items as the main item, and determine the name of the sub-item as the name of the main item;

[0012] Based on the main product name, and in combination with the product information, product image, and product category basic data in the basic data of the combination product, the main product name is rewritten according to the four-level category requirements, and the rewritten main product name is determined as the four-level standardized product name of the combination product corresponding to the four-level category.

[0013] In some embodiments, the step of prioritizing the sub-product name with the highest priority in the combination product and determining it as the main product name includes:

[0014] Prioritize the sub-products according to a priority sorting rule;

[0015] The sub-product name with the highest priority is determined as the main product name.

[0016] In some embodiments, the step of attaching the combination product to the fourth-level category data that matches the fourth-level standardized product name and is determined based on the first candidate fourth-level category data, the second candidate fourth-level category data, and the basic data of the combination product, includes:

[0017] Using the product source data and / or the corresponding product attributes corresponding to the main product name in the combination product as matching dimensions, determine whether there is target fourth-level category data that matches the fourth-level category standardized product name in the first candidate fourth-level category data and the second candidate fourth-level category data;

[0018] If so, the combination product is mounted on the target fourth-level category data.

[0019] In some embodiments, it further includes:

[0020] If not, the combination product is mounted on the fourth-level alternative category.

[0021] In some embodiments, it further includes:

[0022] If not, rewrite the main product name in the combination product according to the requirements of the third-level category based on the basic data of the combination product and the basic data of the product category, and obtain the third-level standardized product name corresponding to the combination product and the third-level category;

[0023] Recalling, based on the third-level standardized product name and the product category basic data, first candidate third-level category data and second candidate third-level category data corresponding to the third-level standardized product name using the first category recall method and / or the second category recall method;

[0024] The combination product is mounted on the third-level category data that matches the third-level standardized product name and is determined based on the first candidate third-level category data, the second candidate third-level category data, and the basic data of the combination product.

[0025] In some embodiments, the step of attaching the combination product to the third-level category data that matches the third-level standardized product name and is determined based on the first candidate third-level category data, the second candidate third-level category data, and the basic data of the combination product includes:

[0026] Using the product source data and / or the corresponding product attributes corresponding to the main product name in the combination product as matching dimensions, determine whether there is target third-level category data matching the third-level standardized product name in the first candidate third-level category data and the second candidate third-level category data;

[0027] If so, the combination product is mounted on the target third-level category data.

[0028] In some embodiments, it further includes:

[0029] If not, the combination product is mounted on the third-level alternative category.

[0030] In some embodiments, it further includes:

[0031] If not, then rewrite the main product name in the combination product according to the first-level category requirements and / or second-level category requirements based on the obtained basic data of the combination product and the basic data of the product category, and obtain the first-level standardized product name corresponding to the first-level category and / or the second-level standardized product name corresponding to the second-level category of the combination product;

[0032] Using the product source data and / or corresponding product attributes corresponding to the main product name in the combination product as matching dimensions, determine whether there is target first-level category data matching the first-level standardized product name and / or target second-level category data matching the second-level standardized product name in the first-level category data and / or the second-level category data;

[0033] If so, the combination product is mounted on the target first-level category data and / or the target second-level category data.

[0034] In some embodiments, it further includes:

[0035] If not, the combination product is mounted on the first-level alternative category and / or the second-level alternative category.

[0036] In a second aspect, the present application further provides a mounting device based on a combination of commodities, comprising:

[0037] a rewriting unit, configured to rewrite the main product name in the combination product according to the requirements of the fourth-level category based on the acquired basic data of the combination product and the basic data of the product category, so as to obtain a fourth-level standardized product name corresponding to the combination product and the fourth-level category;

[0038] a recall unit, configured to perform a fourth-level category recall using a first category recall method and a second category recall method according to the fourth-level standardized product name and the product category basic data, and to recall first candidate fourth-level category data and second candidate fourth-level category data corresponding to the standardized product name, respectively;

[0039] The mounting unit is used to mount the combination product on the fourth-level category data that matches the fourth-level standardized product name, which is determined based on the first candidate fourth-level category data, the second candidate fourth-level category data, and the basic data of the combination product.

[0040] In a third aspect, the present application further provides a computer storage medium for storing data generated by a network platform, and a program for processing the data generated by the network platform;

[0041] When the program is read and executed by the processor, the above-mentioned mounting method based on the combined commodity is executed.

[0042] In a fourth aspect, the present application further provides an electronic device, comprising:

[0043] processor;

[0044] The memory is used to store a program for processing data generated by the network platform. When the program is read and executed by the processor, the mounting method based on the combination of products as described above is executed.

[0045] Compared with the prior art, this application has the following advantages:

[0046] The present application provides a dynamic mounting method based on combination products, which rewrites the main product name of the combination product according to the four-level category requirements based on the basic data of the combination product and the basic data of the product category, and obtains the four-level standardized product name corresponding to the four-level category. The rewritten name of the combination product can be made more accurate by combining the context information, and the accuracy of subsequent category matching can be improved. The four-level standardized name is used as the name of the combination product, and combined with the basic data of the product category, the first candidate four-level category data and the second candidate four-level category data corresponding to the four-level standardized product name are recalled through two category recall methods respectively, so that the recalled candidate four-level category data is more comprehensive and rich. According to the first candidate four-level category data and the second candidate four-level category data, as well as the basic data of the combination product, the four-level category data that matches the four-level standardized product name is determined, so that the combination product is mounted on the four-level category data, and the end-to-end mapping of the combination product to the four-level category is realized, and the coverage rate and mounting efficiency of the combination product to the four-level category are improved.

[0047] Therefore, the dynamic mounting method based on combined products provided by this application can provide a basis for mounting combined products to corresponding categories through standardized name rewriting and comprehensive category data recall, regardless of how the products are combined. Even if there is no matching fourth-level category, it can be further mounted in the third-level category through the recalled third-level category, thereby achieving elastic scaling of the mounted categories. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a flowchart of a mounting method based on combined products provided by this application.

[0049] Figure 2 This is a schematic diagram of the execution process of an embodiment of a mounting method based on combined goods provided in this application.

[0050] Figure 3 This is a structural diagram of a dynamic mounting category system according to an embodiment of a mounting method based on combined commodities provided in this application.

[0051] Figure 4 This is a structural diagram of a mounting device based on a combination of commodities provided in this application.

[0052] Figure 5 This is a structural diagram of an electronic device provided by this application. DETAILED DESCRIPTION

[0053] The following description sets forth many specific details to facilitate a thorough understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the scope of the present application. Therefore, the present application is not limited to the specific implementations disclosed below.

[0054] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. Descriptive terms such as "a," "a," "a first," and "a second," etc., used in this application and the appended claims, are not intended to limit quantity or sequence, but are used to distinguish information of the same type from one another.

[0055] Based on the background technology, we can know that as user needs become increasingly diversified and personalized, life service application platforms also need to continuously adjust the platform data architecture in combination with the adjustment of merchant product data to adapt to the diverse and personalized needs of users, thereby increasing exposure and conversion rates.

[0056] However, existing lifestyle service application platforms (including shopping platforms) are not well-suited to the category architecture after product data adjustments due to their rigid product category systems. This creates a conflict between the agility requirements of product data services and the rigid structure of the product category system. For example, they are unable to dynamically adjust categories based on dynamic changes in product data, and they are unable to integrate multiple types of data to achieve dynamic category adjustments.

[0057] In view of the above problems, this application provides a dynamic mounting method based on combined products, such as Figure 1 As shown, Figure 1 This is a flowchart of a dynamic mounting method based on combined products provided by this application, including:

[0058] Step S101: Based on the acquired basic data of the combination product and the basic data of the product category, the main product name in the combination product is rewritten according to the requirements of the fourth-level category to obtain the fourth-level standardized product name corresponding to the combination product and the fourth-level category;

[0059] Step S102: Based on the fourth-level standardized product name and the product category basic data, a fourth-level category recall is performed using a first category recall method and a second category recall method, respectively, to recall first candidate fourth-level category data and second candidate fourth-level category data corresponding to the standardized product name;

[0060] Step S103: Mounting the combination product to the fourth-level category data that matches the fourth-level standardized product name, which is determined based on the first candidate fourth-level category data, the second candidate fourth-level category data, and the basic data of the combination product.

[0061] Before describing the above steps S101 to S103 in detail, the relevant technical terms or technical nouns involved in the technical solution of this application are first explained.

[0062] CPV system: A structured model of product information in the e-commerce field, including a standardized knowledge system consisting of categories, properties, and attribute values, used to achieve digital management and efficient operation of products.

[0063] Category: Classify products into layers by combining attributes of similar dimensions.

[0064] Attributes and Values: Attributes define general characteristics of a product (such as color and size), while attribute values ​​are specific optional values ​​(such as red and XL).

[0065] MLLM: Multimodal Large Language Model, multimodal large language model.

[0066] Dynamic categories: Flexible categories that dynamically adjust classification levels in real time based on user service needs, merchant service needs, or application platform needs.

[0067] The dynamic mounting method based on combined goods provided in this application can be implemented in at least two specific application scenarios, such as a catering application service platform, a shopping application service platform, etc.

[0068] Regarding step S101: based on the acquired basic data of the combination product and the basic data of the product category, the main product name in the combination product is rewritten according to the requirements of the four-level category to obtain the four-level standardized product name corresponding to the combination product and the four-level category.

[0069] The combination product in step S101 can be a combination of existing products on the application service platform, or a combination of new products. The combination product can be a new product that the merchant has combined on their own, or a combination product provided by the merchant to participate in platform activities. There is no restriction on the basis for generating the combination product, and there is no restriction on the combination form of the combination product. For example, in a catering application scenario, the combination product can be a combination product consisting of meal a + meal b + meal c provided by merchant A for one person, such as: pickled mustard greens and shredded pork rice + beverage, stir-fried meat + vegetables + rice, etc.; in a shopping application scenario, the combination product can be a combination product consisting of meal a + meal b + meal c provided by merchant B, such as: shirt + skirt, sports short-sleeved shirt + sports shorts + sports shoes, etc. It can be understood from the embodiments given in combination with the two application scenarios that the combination product described in this embodiment can be a dynamically changing product, not a fixed combination, and can change dynamically with the needs of merchants, users, platforms, etc., such as the reorganization of combination products, the deletion of combination products, etc.

[0070] The commodity category basic data in step S101 may be existing categories in the platform and the relationships between categories, i.e., the CPV system. A means of classifying commodities by commodity category may be a tree structure, i.e., a category tree.

[0071] The purpose of step S101 is to rewrite the name of the combination product to obtain a standardized product name corresponding to the category requirements, thereby facilitating the establishment of an association between the combination product and the product category, that is, facilitating the attachment of the combination product to the corresponding category.

[0072] like Figure 3 As shown, in this embodiment, the second-level category is the next level category under the first-level category, and both are fixed categories. The third-level category is the next level category under the second-level category and is adjustable, while the fourth-level category is the next level category under the third-level category and can also be an adjustable category. For the application platform, the first-level and second-level categories are already fixed categories, while the fourth-level category is generated based on service needs.

[0073] In this embodiment, there is no unified product name for the combination product, and the combination product is a dynamically changing product and is not suitable for a designated name. Therefore, in order to match the combination product with the category data, the combination product name can be rewritten. The specific implementation process of step S101 may include:

[0074] Step S101-11: Split the combination product according to the basic data of the combination product to obtain the sub-products included in the combination product; for example, if the combination product is stir-fried meat + vegetables + rice, it can be split into stir-fried meat, vegetables, and rice; if the combination product is shirt + tie, it can be split into shirt and tie.

[0075] Step S101-12: The sub-product with the highest priority in the combination of the sub-products is determined as the main product, and the name of the sub-product is determined as the name of the main product; the specific implementation process may be: according to the priority sorting rule, the sub-products are prioritized, and the name of the sub-product with the highest priority is determined as the name of the main product. The priority sorting rule may include multiple implementation methods, for example: according to the price attribute of the sub-product as the priority sorting rule, the higher the price, the higher the priority; according to the use or edible attribute of the sub-product as the priority sorting rule, wherein the use attribute can be the functional attribute of the product, the stronger the function, the higher the priority; the edible attribute for the product as a meal can include staple food, main dish, side dish, soup, meat and vegetable, beverage, etc., and the priority can be determined by the importance of the edible attribute, or by combining the time of ordering, for example: during the main meal time period, the main dish has the highest priority, such as during the afternoon tea time period, the snack has the highest priority, etc., and examples are not given here; according to the combination order of the sub-products as the priority sorting rule, the priority is determined by one or more methods, and the combination order is the arrangement order, and the higher the position, the higher the priority. It is understandable that the priority sorting rules can be set according to specific application scenarios and requirements, and are not limited to the above content.

[0076] Step S101-13: Based on the main product name, and in combination with the product information, product image, and product category basic data included in the combination product basic data, the main product name is rewritten according to the four-level category requirements, and the rewritten main product name is determined as the four-level standardized product name of the combination product corresponding to the four-level category.

[0077] The specific implementation of steps S101-11 to S101-13 can be implemented using a generative model. For example, a pre-trained generative model can be used to rewrite the product name of the combination product. The prompt input to the generative model could be: "Product Name: Single Meal of Pickled Vegetables and Shredded Pork Rice Bowl + Beverage 1" and "Category Name: Fourth-Level Category." The generative model then performs the following tasks: splitting the product into: sub-products: Single Meal of Pickled Vegetables and Shredded Pork Rice Bowl, sub-product: Beverage 1; rewriting the name of the main product, Single Meal of Pickled Vegetables and Shredded Pork Rice Bowl, according to the fourth-level category. The output is: "Meal of Pickled Vegetables and Shredded Pork Rice Bowl," which is the rewritten fourth-level standardized product name of the combination product. This rewriting process removes modifiers, quantity terms, non-main product terms, and other unrelated terms to achieve a rewritten standardized name. The rewritten name serves as the standardized name (SPU name, SPU: Standard Product Unit) of the combination product.

[0078] In this embodiment, the basic data of commodity categories may include primary category data, secondary category data, tertiary category data, and quaternary category data, etc., which belong to the category system pre-built by the application platform. It can be understood that the primary category data and the secondary category data are fixed categories, and the tertiary category data and the quaternary category data are adjustable categories. In this embodiment, taking food as an example, the primary / secondary category data may include the following examples:

[0079] {"Japanese and Korean cuisine":"Japanese and Korean cuisine"};

[0080] {"Hamburger Pizza": "Pizza, hamburger, western food, light meal"};

[0081] {"Fried Chicken Skewers": "Fried food, grilled food, special snacks"};

[0082] {"rice noodles": "rice noodles, river noodles, cold noodles, vermicelli noodles, snail noodles, noodles"};

[0083] {“Meal Set”: “Cold Dish Set, Rice / Pan Rice / Wooden Barrel Rice, Dry Pot / Clay Pot Rice, Bibimbap, Fried Rice, Grilled / Baked Dish Set, Roasted / Braised Dish Set, Pan-Fried / Stir-Fried Dish Set, Steamed / Boiled / Stewed Dish Set, Spicy Hot Pot / Fried Pot / Maocai”};

[0084] {“Drinks and Desserts”: “Frozen Desserts, Coffee, Milk Tea, Fruit Tea, Fruit and Vegetable Juice, Milk and Soy Milk Products, Cakes, Sweet Soups, Pure Tea, Flower Tea”};

[0085] {"dumplings, porridge, and cakes": "steamed buns, soup dumplings, soup, glutinous rice balls, special dim sum, porridge, rice noodle rolls, dumplings, pot stickers, cakes, buns, steamed buns, wontons, wontons, and wontons"};

[0086] Fourth-level category: Standardized names can be exemplified as follows:

[0087] {"Specialty Dim Sum": "Pork huoshao, noodle soup, mixed rice wrappers, fried glutinous rice balls, glutinous rice cakes, glutinous rice, fresh meat dumplings, Longjing fragrant green rice balls, huoshao, donkey meat huoshao..."};

[0088] {"rice roll":"Meat and egg rice roll, ham rice roll, lean meat rice roll, minced meat rice roll, rice roll, egg rice roll, egg and pork rice roll..."};

[0089] {"Light Meal": "Seared longhair with grain rice, chicken and beef salad, egg salad, steak and chicken breast salad, pork floss and cheese sandwich..."};

[0090] {"Bibimbap": "Chopped Pepper Bibimbap, Classic Stone Pot Bibimbap, Beef Stone Pot Bibimbap, Korean Stone Pot Bibimbap, Fish Flavored Shredded Pork Bibimbap, Chicken and Multigrain Bibimbap, Salad and Fried Chicken Bibimbap, Beef Bibimbap, Tuna Bibimbap, Chicken Bibimbap..."};

[0091] It is understandable that the second-level categories are further divided based on the first-level categories, the third-level categories are further divided based on the second-level categories, and the fourth-level categories are further divided based on the third-level categories. The third-level and fourth-level categories can be adjusted and changed based on the application requirements of specific service data.

[0092] In the process of training the generative model, it can also be trained through multi-task training. For example: by training on the first and second level category data, the model can be generated in a controllable manner, directly hitting the actual first and second level categories; by training on the fourth level category data, the model can rewrite the name of the product, thereby filtering out non-core information, especially for set meals such as "Stir-fried Pork + Vegetables + Rice". Through end-to-end training, it can be directly rewritten as the standard dish name "Stir-fried Pork", so that the fourth level category focuses on the core information of the product. Figure 2 As shown in , the product multimedia information (such as pictures, videos, etc.) and product text information included in the combination product can be input into the generative model for rewriting.

[0093] Regarding step S102: according to the four-level standardized product name and the product category basic data, four-level category recall is performed through the first category recall method and the second category recall method respectively, and the first candidate four-level category data and the second candidate four-level category data corresponding to the standardized product name are recalled respectively.

[0094] like Figure 2 As shown, Figure 2This is a schematic diagram of the execution process of an embodiment of a dynamic mounting method based on combined products provided by this application. In step S102, at least two recalls are performed on the rewritten four-level standard product name and product category basic data, respectively, to improve the comprehensiveness and possibility of recall based on the existing product category basic data. The specific implementation process can be:

[0095] Using the four-level standardized product names and the product category basic data as input data for a first intent understanding macromodel in a first category recall method, and obtaining first candidate four-level category data output by the first intent understanding macromodel;

[0096] The four-level standardized product names and the product category basic data are used as input data of the second intention understanding model in the second type of recall method to obtain the second candidate four-level category data output by the second intention understanding model.

[0097] The first intention understanding big model and the second intention understanding big model can be different multimodal big models, or different unimodal big models, or a mixed multimodal and unimodal big model. For example, the first intention understanding big model is a multimodal big model, and the second intention understanding big model is a unimodal big model. The first intention understanding big model and the second intention understanding big model can be selected in combination with the input data of specific intention understanding. In this embodiment, the first intention understanding big model and the second intention understanding big model can be a twin-tower model.

[0098] like Figure 2 As shown, the first intention understanding model includes: a multimodal model A and a text model B. The input data of the multimodal model A is the four-level standardized product name, and the input data of the text model B is the basic data of the product category. That is, the query request of the input data is converted into a text vector according to the multimodal model A, and the text vector is used as the input data of the text model B for vector recall to obtain the first candidate four-level category data.

[0099] like Figure 2 As shown, the second intention understanding model includes: a multimodal model C and a text model D. The input data of the multimodal model C is the four-level standardized product name, and the input data of the text model D is the basic data of the product category. That is, the query request of the input data is converted into a text vector according to the multimodal model C, and the text vector is used as the input data of the text model D for vector recall to obtain the second candidate four-level category data.

[0100] It is understandable that the above description only uses two recall methods and recall candidate categories as examples. In fact, the recall methods and the number and methods of candidate categories can be adjusted according to the type of specific service data, scenarios, etc. Each candidate category data, whether it is level four or level three, can include multiple, for example: the first candidate level four category data includes multiple candidate level four category data, the second candidate level four category data includes multiple candidate level four category data, etc. In other words, the first candidate level four category data and the second candidate level four category data are data sets, and each candidate set includes one or more candidate category data. Similarly, the other candidate category data involved later also apply.

[0101] Regarding step S103: mounting the combination product to the fourth-level category data that matches the fourth-level standardized product name, determined based on the first candidate fourth-level category data, the second candidate fourth-level category data, and the basic data of the combination product.

[0102] The purpose of step S103 is to attach the combined product to the matching fourth-level category data. This process can be implemented through a multimodal macro model. For example, the first candidate fourth-level category and the second candidate fourth-level category, as well as the basic data of the combined product, are used as input data of the multimodal macro model. The task is to find the fourth-level category data that matches the fourth-level standardized product name. The specific implementation process can include multiple implementation methods, as follows:

[0103] Step S103-11: Using the product source data and / or the corresponding product attributes corresponding to the main product name in the combination product as matching dimensions, determine whether there is target fourth-level category data that matches the fourth-level category standardized product name in the first candidate fourth-level category data and the second candidate fourth-level category data;

[0104] Step S103-12: If yes, then the combination product is mounted on the target fourth-level category data.

[0105] The product source data may refer to food ingredients, processing methods, etc. for food products, and the product attributes may include taste attributes, seasonal attributes, regional attributes, etc.

[0106] The specific implementation process of step S103-11 may include:

[0107] Step S103-111: extracting source data features and / or attribute features of the main product from the main product name in the combination product;

[0108] Step S103-112: determining the similarity between the main product name and the first candidate fourth-level category data and the second candidate fourth-level category data respectively based on the source data characteristics and / or attribute characteristics;

[0109] Steps S103-113: Based on the similarity, determine whether target fourth-level category data matching the standardized product name of the fourth-level category exists in the first candidate fourth-level category data and the second candidate fourth-level category data. Specifically, the similarity value is compared with a similarity threshold. If the similarity value is greater than or equal to the similarity threshold, the target fourth-level category data is considered to exist; otherwise, the target fourth-level category data is considered to not exist.

[0110] It should be noted that in determining whether there is target fourth-level category data that matches the standardized product name of the fourth-level category in the first candidate fourth-level category data and the second candidate fourth-level category data, the product source data and / or corresponding product attributes corresponding to the main product name in the combination product can be used as the matching dimension for judgment, and the merchant information can also be used as auxiliary information for judgment. For example: Product Name: Sweet and Spicy Three-Piece Set (Duck Neck + Duck Collarbone + Duck Wings), Merchant Name: Luxianshuo·Quzhou Duck Head (Wulian Xiyuan Store), Candidate Fourth-Level Categories: Braised Duck, Roast Duck, Roasted Pig's Trotter, Oreo Sundae, Fried Tofu with Meat, based on the similarity, the target fourth-level category data can be determined to be: Braised Duck.

[0111] like Figure 3 As shown, it can be understood that when the combination product can be mounted on the target fourth-level category data that matches the standardized product name of the fourth-level category, it can be mapped to the first, second, and third-level categories according to the category system, thereby improving the execution efficiency of the combination product classification task. If the target fourth-level category data does not exist, the execution method of step S103-11 can include multiple methods, as follows:

[0112] Method 1 includes:

[0113] Step S103-a: When there is no target fourth-level category data that matches the standardized product name of the fourth-level category in the first candidate fourth-level category data and the second candidate fourth-level category data, the combination product can be mounted on the fourth-level alternative category, for example: the fourth-level category "other categories", and the other categories are categories named after other categories.

[0114] Method 2 includes:

[0115] Step S103-b1: When there is no target fourth-level category data matching the standardized product name of the fourth-level category in the first candidate fourth-level category data and the second candidate fourth-level category data, the main product name in the combination product is rewritten according to the third-level category requirements based on the obtained basic data of the combination product and the product category basic data, to obtain the third-level standardized product name corresponding to the third-level category of the combination product;

[0116] Step S103-b2: Based on the third-level standardized product name and the product category basic data, first candidate third-level category data and second candidate third-level category data corresponding to the third-level standardized product name are respectively recalled using the first category recall method and / or the second category recall method;

[0117] Step S103-b3: Mounting the combination product to the third-level category data that matches the third-level standardized product name, determined based on the first candidate third-level category data, the second candidate third-level category data, and the basic data of the combination product.

[0118] The specific implementation process of step S103-b3 is similar to the specific implementation process of step S103-11, that is, it includes:

[0119] Step S103-b31: Using the product source data and / or corresponding product attributes corresponding to the main product name in the combination product as matching dimensions, determine whether there is target third-level category data that matches the third-level standardized product name in the first candidate third-level category data and the second candidate third-level category data;

[0120] Step S103-b32: If yes, then the combination product is mounted on the target third-level category data.

[0121] When the judgment result of step S103 - b31 is no, multiple implementation methods may be included.

[0122] Method 1 includes:

[0123] Step S103-c: When there is no target third-level category data that matches the third-level standardized product name in the first candidate third-level category data and the second candidate third-level category data, the combination product can be mounted on the third-level alternative category, for example: the third-level category "other categories", and the other categories are categories named after other categories.

[0124] Method 2 includes:

[0125] Step S103-d1: When target third-level category data matching the third-level standardized product name does not exist in the first candidate third-level category data and the second candidate third-level category data, rewrite the main product name in the combination product according to the first-level category requirements and / or second-level category requirements based on the acquired combination product basic data and product category basic data, to obtain the first-level standardized product name corresponding to the first-level category and / or the second-level standardized product name corresponding to the second-level category of the combination product;

[0126] Step S103-d2: Using the product source data and / or the corresponding product attributes corresponding to the main product name in the combination product as matching dimensions, determine whether there is target first-level category data matching the first-level standardized product name and / or target second-level category data matching the second-level standardized product name in the first-level category data and / or the second-level category data;

[0127] Step S103 - d3 : If yes, then the combination product is mounted on the target first-level category data and / or the target second-level category data.

[0128] If there is no target first-level category data that matches the first-level standardized product name and / or target second-level category data that matches the second-level standardized product name in the first-level category data and / or second-level category data, the combination product can be mounted on the first-level alternative category and / or the second-level alternative category, for example: the "other categories" of the first-level category and / or the second-level category, and the other categories are categories named after other categories.

[0129] Based on the above, in the process of classifying the association relationship between the combination products and the category system, dynamic mounting can be achieved. It is not limited to the existence of four-level categories corresponding to the combination products in the category system in advance. Moreover, the combination products are a dynamically changing product and do not belong to the inherent products included in the category. Therefore, the above method can realize the dynamic mounting of the combination products in the category system, with the first and second level categories as the stability layer and the third and fourth level categories as the agility layer, to realize the dynamic mounting of the combination products with a mixed structure.

[0130] It should be noted that, in addition to recalling the candidate fourth-level categories in step S102 in the above embodiment, the candidate third-level categories can also be recalled at the same time, that is: when rewriting the main product name in the combination product, the fourth-level category can be rewritten and the third-level category can be rewritten. When recalling the candidate categories, the first candidate fourth-level category data and the second candidate fourth-level category data corresponding to the fourth-level standardized product name can be recalled, and the first candidate third-level category data and the second candidate third-level category data corresponding to the third-level standardized product name can also be recalled. Therefore, when executing step S103-b1, that is: when the first candidate fourth-level category data and the second candidate fourth-level category data do not have the target fourth-level category data that matches the fourth-level category standardized product name, step S103-b31 can be directly executed, that is, to determine whether the first candidate third-level category data and the second candidate third-level category data have the target third-level category data that matches the third-level standardized product name.

[0131] Similarly, when rewriting the name of the main product in the combination product in step S101, it can also include: rewriting the name of the main product according to the requirements of the first-level category and / or the requirements of the second-level category, and obtaining the first-level standardized product name corresponding to the first-level category and / or the second-level standardized product name corresponding to the second-level category of the combination product. When there is no target fourth-level category data that matches the fourth-level category standardized product name in the first candidate fourth-level category data and the second candidate fourth-level category data, and there is no target third-level category data that matches the third-level standardized product name in the first candidate third-level category data and the second candidate third-level category data, it can be directly determined whether there is target first-level category data that matches the first-level standardized product name and / or target second-level category data that matches the second-level standardized product name in the first category data and / or the second category data.

[0132] Therefore, the rewriting can be performed sequentially according to whether the target category exists in the candidate categories, or the rewriting of the combined product name can be performed simultaneously for the fourth-level category, the third-level category, the second-level category, and the first-level category. In other words, one implementation method can be serial implementation, and another implementation method can be parallel implementation.

[0133] After determining the mounting category of the combination product, you can also add or delete the fourth-level category and / or the third-level category. By counting the third-level category and / or the fourth-level category, determine whether to add or delete the edges and leaf nodes of the third-level and / or the fourth-level category. For example: if the number of third-level and / or fourth-level categories determined for the combination product is greater than the set threshold, you can add edges and leaf nodes to the third-level and / or fourth-level categories, and establish the relationship between the leaf nodes and the third-level and / or fourth-level categories. The category data can be updated by changing the standardized name of the product and setting its relationship with the third-level and / or fourth-level category. Therefore, in the face of diversified combination products, the diversification and richness of the category system can be achieved through dynamic mounting, and the third-level and / or fourth-level categories of the category system can also be dynamically changed as the combination products change. This improves the efficiency of linking products to categories when creating new and diverse service combinations based on the application service platform. It also meets the iteration requirements of daily, event, and holiday categories for these combined products, allowing for on-demand elastic scaling of categories for different service products or objects. This delivers a superior user experience for both the business end (merchant users) and the consumer end (purchasers). For the platform, this also enables efficient data category construction and management for merchants' ever-changing service products.

[0134] Based on the above, it can be understood that in the dynamic mounting method based on combined products provided by this application, if the third-level and fourth-level category nodes and their associations are modified or changed, there is no need to change or modify the standardized name of the product. Instead, the standardized name of the product can be used to recall new categories through the intent understanding model, thereby updating the category data. If the basic data of the product changes, the standardized name of the product is re-predicted, and the new category is recalled through the generative model, intent understanding model, etc. to update the category data.

[0135] The above is a description of an embodiment of a dynamic mounting method based on a combination of products provided by the present application. This method embodiment rewrites the main product name of the combination product according to the four-level category requirements based on the acquired basic data of the combination product and the basic data of the product category, and obtains a four-level standardized product name corresponding to the four-level category. It can make the rewritten name of the combination product more accurate by combining context information, and can improve the accuracy of subsequent category matching. The four-level standardized name is used as the name of the combination product, and combined with the basic data of the product category, the first candidate four-level category data and the second candidate four-level category data corresponding to the four-level standardized product name are recalled by two category recall methods respectively, so that the recalled candidate four-level category data is more comprehensive and rich. According to the first candidate four-level category data and the second candidate four-level category data, as well as the basic data of the combination product, the four-level category data that matches the four-level standardized product name is determined, so that the combination product is mounted on the four-level category data, and end-to-end mapping of the combination product to the four-level category is realized, and the coverage rate and mounting efficiency of the combination product to the four-level category are improved.

[0136] Therefore, the dynamic mounting method based on combined products provided by this application can provide a basis for mounting combined products to corresponding categories through standardized name rewriting and comprehensive category data recall, regardless of how the products are combined. Even if there is no matching fourth-level category, it can be further mounted in the third-level category through the recalled third-level category, thereby achieving elastic scaling of the mounted categories.

[0137] The above is a detailed description of an embodiment of a dynamic mounting method based on a combination of commodities provided by this application. Corresponding to the embodiment of a dynamic mounting method based on a combination of commodities provided above, this application also discloses an embodiment of a dynamic mounting device based on a combination of commodities, please refer to Figure 4 Since the device embodiment is basically similar to the method embodiment, the description is relatively simple. For relevant details, please refer to the partial description of the method embodiment. The device embodiment described below is only illustrative.

[0138] like Figure 4 As shown, Figure 4 This is a structural diagram of a dynamic mounting device based on a combination of commodities provided by the present application, which includes:

[0139] The rewriting unit 401 is configured to rewrite the main product name in the combination product according to the requirements of the fourth-level category based on the acquired basic data of the combination product and the basic data of the product category, thereby obtaining a fourth-level standardized product name corresponding to the combination product and the fourth-level category;

[0140] A recall unit 402 is configured to perform a fourth-level category recall based on the fourth-level standardized product name and the product category basic data using a first category recall method and a second category recall method, respectively, to recall first candidate fourth-level category data and second candidate fourth-level category data corresponding to the standardized product name;

[0141] The mounting unit 403 is configured to mount the combination product onto the fourth-level category data that matches the fourth-level standardized product name and is determined based on the first candidate fourth-level category data, the second candidate fourth-level category data, and the basic data of the combination product.

[0142] The specific implementation process of the rewriting unit 401 may include: a splitting subunit, a first determining subunit, and a second determining subunit;

[0143] The splitting subunit is configured to split the combination product according to the basic data of the combination product to obtain the sub-products included in the combination product;

[0144] The first determining sub-unit is configured to determine the sub-product with the highest priority in the combination of sub-products as the main product, and determine the name of the sub-product as the name of the main product. Specific implementation may include: prioritizing the sub-products according to a priority sorting rule; and determining the name of the sub-product with the highest priority as the name of the main product.

[0145] The second determination subunit is used to rewrite the main product name according to the requirements of the four-level category based on the main product name and in combination with the product information, product image and the product category basic data in the combination product basic data, and determine the rewritten main product name as the fourth-level standardized product name of the combination product corresponding to the fourth-level category.

[0146] For the specific content of the rewriting unit 401, please refer to the description of step S101 above, which will not be described in detail here.

[0147] For the specific content of the recall unit 402, please refer to the description of step S102 above, which will not be described in detail here.

[0148] The specific implementation process of the mounting unit 403 may include: a judgment subunit and a mounting subunit;

[0149] The judgment subunit is configured to determine whether there is target fourth-level category data that matches the fourth-level category standardized product name in the first candidate fourth-level category data and the second candidate fourth-level category data, using the product source data and / or the corresponding product attributes corresponding to the main product name in the combination product as a matching dimension;

[0150] The mounting subunit is configured to mount the combination product onto the target fourth-level category data when the judgment result of the judgment subunit is yes.

[0151] The mounting subunit further includes: when the judgment result of the judgment subunit is negative, mounting the combination product on the fourth-level alternative category.

[0152] When the judgment result of the judgment subunit is negative, the method may further include: a third-level rewriting subunit, configured to rewrite the main product name in the combination product according to the third-level category requirements based on the acquired basic data of the combination product and the basic data of the product category, to obtain a third-level standardized product name corresponding to the combination product and the third-level category;

[0153] a third-level recall subunit, configured to recall, based on the third-level standardized product name and the product category basic data, first candidate third-level category data and second candidate third-level category data corresponding to the third-level standardized product name using the first category recall method and / or the second category recall method;

[0154] The third-level mounting sub-unit is used to mount the combination product to the third-level category data that matches the third-level standardized product name, determined based on the first candidate third-level category data and the second candidate third-level category data, as well as the basic data of the combination product. Specifically, it includes: using the product source data and / or the corresponding product attributes corresponding to the main product name in the combination product as the matching dimension, determining whether there is target third-level category data that matches the third-level standardized product name in the first candidate third-level category data and the second candidate third-level category data; if so, mounting the combination product on the target third-level category data. If not, mounting the combination product on the third-level alternative category.

[0155] If there is no target third-level category data matching the third-level standardized product name in the first candidate third-level category data and the second candidate third-level category data, the method further includes:

[0156] Based on the acquired basic data of the combination product and the basic data of the product category, rewrite the main product name in the combination product according to the requirements of the first-level category and / or the requirements of the second-level category to obtain the first-level standardized product name corresponding to the first-level category and / or the second-level standardized product name corresponding to the second-level category of the combination product;

[0157] Using the product source data and / or corresponding product attributes corresponding to the main product name in the combination product as matching dimensions, determine whether there is target first-level category data matching the first-level standardized product name and / or target second-level category data matching the second-level standardized product name in the first-level category data and / or the second-level category data;

[0158] If yes, the combination product is mounted on the target primary category data and / or the target secondary category data. If no, the combination product is mounted on the primary candidate category and / or the secondary candidate category.

[0159] For the specific content of the mounting unit 403, please refer to the content of the above step S103, which will not be described in detail here.

[0160] Based on the above, the present application also provides a computer storage medium for storing data generated by a network platform and a program for processing the data generated by the network platform;

[0161] When the program is read and executed by the processor, the program executes the relevant steps in the above-mentioned dynamic mounting method based on combined commodities.

[0162] Based on the above, the present application also provides an electronic device, such as Figure 5 As shown, Figure 5 This is a schematic diagram of the structure of an electronic device provided by this application, which includes:

[0163] Processor 501;

[0164] The memory 502 is used to store a program for processing data generated by the network platform. When the program is read and executed by the processor, the program executes the relevant steps in the above-mentioned dynamic mounting method based on combined products.

[0165] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0166] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0167] Memory may include non-permanent storage in a computer-readable medium, in the form of random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0168] 1. Computer-readable media, including both permanent and non-permanent, removable and non-removable media, can be implemented using 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 RAM (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 technologies, compact disc read-only memory (CD-ROM), digital versatile disc (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 non-transitory computer-readable media, such as modulated data signals and carrier waves.

[0169] 2. Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0170] Although the present application is disclosed as above with the preferred embodiments, it is not intended to limit the present application. Any person skilled in the art may make possible changes and modifications without departing from the spirit and scope of the present application. Therefore, the scope of protection of the present application shall be based on the scope defined by the claims of the present application.

Claims

1. A mounting method based on combined goods, characterized in that: include: Based on the acquired basic data of the combination product and the basic data of the product category, the main product name in the combination product is rewritten according to the requirements of the four-level category to obtain the four-level standardized product name corresponding to the combination product and the four-level category; including: splitting the combination product according to the basic data of the combination product to obtain sub-products included in the combination product; determining the sub-product with the highest priority in the combination product as the main product, and determining the name of the sub-product as the main product name; based on the main product name, in combination with the product information, product image in the basic data of the combination product, and the basic data of the product category, rewriting the main product name according to the requirements of the four-level category, and determining the rewritten main product name as the four-level standardized product name of the combination product corresponding to the four-level category; According to the fourth-level standardized product name and the product category basic data, a fourth-level category recall is performed using a first category recall method and a second category recall method, respectively, to recall first candidate fourth-level category data and second candidate fourth-level category data corresponding to the fourth-level standardized product name; The combination product is mounted on the fourth-level category data that matches the fourth-level standardized product name and is determined based on the first candidate fourth-level category data, the second candidate fourth-level category data, and the basic data of the combination product.

2. The method for mounting a combination of products according to claim 1, characterized in that: The step of prioritizing the sub-product name with the highest priority in the combination of products and determining it as the main product name includes: Prioritize the sub-products according to a priority sorting rule; The sub-product name with the highest priority is determined as the main product name.

3. The method for mounting a combination of products according to claim 1, characterized in that: The step of mounting the combination product on the fourth-level category data that matches the fourth-level standardized product name and is determined based on the first candidate fourth-level category data, the second candidate fourth-level category data, and the basic data of the combination product includes: Using the product source data and / or the corresponding product attributes corresponding to the main product name in the combination product as matching dimensions, determine whether there is target fourth-level category data matching the fourth-level standardized product name in the first candidate fourth-level category data and the second candidate fourth-level category data; If so, the combination product is mounted on the target fourth-level category data.

4. The method for mounting a combination of commodities according to claim 3, characterized in that: Also includes: If not, the combination product is mounted on the fourth-level alternative category.

5. The method for mounting a combination of commodities according to claim 3, characterized in that: Also includes: If not, rewrite the main product name in the combination product according to the requirements of the third-level category based on the basic data of the combination product and the basic data of the product category, and obtain the third-level standardized product name corresponding to the combination product and the third-level category; Recalling, based on the third-level standardized product name and the product category basic data, first candidate third-level category data and second candidate third-level category data corresponding to the third-level standardized product name using the first category recall method and / or the second category recall method; The combination product is mounted on the third-level category data that matches the third-level standardized product name and is determined based on the first candidate third-level category data, the second candidate third-level category data, and the basic data of the combination product.

6. The method for mounting a combination of commodities according to claim 5, characterized in that: The step of attaching the combination product to the third-level category data that matches the third-level standardized product name and is determined based on the first candidate third-level category data, the second candidate third-level category data, and the basic data of the combination product, includes: Using the product source data and / or the corresponding product attributes corresponding to the main product name in the combination product as matching dimensions, determine whether there is target third-level category data matching the third-level standardized product name in the first candidate third-level category data and the second candidate third-level category data; If so, the combination product is mounted on the target third-level category data.

7. The method for mounting combined commodities according to claim 6, characterized in that: Also includes: If not, the combination product is mounted on the third-level alternative category.

8. The method for mounting combined commodities according to claim 3 or 6, characterized in that: Also includes: If not, then rewrite the main product name in the combination product according to the first-level category requirements and / or second-level category requirements based on the obtained basic data of the combination product and the basic data of the product category, and obtain the first-level standardized product name corresponding to the first-level category and / or the second-level standardized product name corresponding to the second-level category of the combination product; Using the product source data and / or corresponding product attributes corresponding to the main product name in the combination product as matching dimensions, determine whether there is target first-level category data matching the first-level standardized product name and / or target second-level category data matching the second-level standardized product name in the first-level category data and / or the second-level category data; If so, the combination product is mounted on the target first-level category data and / or the target second-level category data.

9. The method for mounting a combination of commodities according to claim 8, characterized in that: Also includes: If not, the combination product is mounted on the first-level alternative category and / or the second-level alternative category.

10. A mounting device based on a combination of commodities, characterized in that: include: A rewriting unit is configured to rewrite the name of the main product in the combination product according to the requirements of the four-level category based on the acquired basic data of the combination product and the basic data of the product category, so as to obtain the four-level standardized product name corresponding to the combination product and the four-level category; the rewriting unit comprises: splitting the combination product according to the basic data of the combination product to obtain the sub-products included in the combination product; determining the sub-product with the highest priority in the combination product as the main product, and determining the name of the sub-product as the main product name; rewriting the main product name according to the requirements of the four-level category based on the main product name and in combination with the product information, product image in the basic data of the combination product and the basic data of the product category, and determining the rewritten main product name as the four-level standardized product name of the combination product corresponding to the four-level category; a recall unit, configured to perform a fourth-level category recall using a first category recall method and a second category recall method according to the fourth-level standardized product name and the product category basic data, and to recall first candidate fourth-level category data and second candidate fourth-level category data corresponding to the standardized product name, respectively; The mounting unit is used to mount the combination product on the fourth-level category data that matches the fourth-level standardized product name, which is determined based on the first candidate fourth-level category data, the second candidate fourth-level category data, and the basic data of the combination product.

11. A computer storage medium for storing data generated by a network platform and a program for processing the data generated by the network platform; When the program is read and executed by a processor, the program executes the mounting method based on combined commodities as described in any one of claims 1 to 9 above.

12. An electronic device comprising: processor; A memory for storing a program for processing data generated by a network platform, wherein when the program is read and executed by the processor, the mounting method based on the combination product as described in any one of claims 1 to 9 is executed.

Images