Object classification method, apparatus, terminal device, and storage medium

Abstract

Embodiments of the present application provide an object classification method and device, terminal equipment and a storage medium, and relate to the technical field of data processing, which can quickly obtain accurate object categories. The method comprises the following steps: obtaining category information set by each business system of a plurality of business systems for a target entity to which a first object belongs, and matching a function attribute of the target entity; in a plurality of preset categories set in a local business system storing the first object, at least one first target preset category corresponding to each of the plurality of category information is queried; and the first object is classified according to the number of category information corresponding to each first target preset category.

Description

[Technical Field]

[0001] The embodiments of the present invention relate to the field of data processing technology, and in particular to an object classification method, a device terminal equipment, and a storage medium. [Background Technology]

[0002] With the development of the internet economy, internet platforms, based on the application scenarios of their own business systems, use data and functions to express the characteristics and functions of offline entities (such as shops, landmark buildings, public parks, etc.) in order to create objects that represent the corresponding entities online.

[0003] In application scenarios of relevant business systems (such as user store search and map base rendering), it is necessary to execute relevant business functions based on object categories. For example, e-commerce platforms recommend stores whose categories match the user's search keywords; map systems render different icons on the canvas and display them on the interface based on landmark building categories. It is evident that the accuracy of object category classification directly impacts whether the system can provide a good user experience.

[0004] However, under current technology, business systems are limited in determining object categories by the data related to the entity to which the object belongs, which is stored locally, and cannot obtain accurate object category information. [Summary of the Invention]

[0005] This invention provides an object classification method, apparatus, terminal device, and storage medium that can quickly obtain accurate object categories.

[0006] In a first aspect, embodiments of the present invention provide an object classification method applied to an electronic device. The method includes: obtaining category information set by each of a plurality of business systems for the target entity that matches the functional attributes of the target entity to which the first object belongs; querying at least one first target preset category corresponding to each of the plurality of category information in a local business system storing the first object; and classifying the first object according to the number of category information corresponding to each first target preset category.

[0007] In the above method, the first target preset category represents the local system's expression of the type of the target entity to which the first object belongs, as determined by other Internet systems. The larger the number of category information corresponding to the first target preset category, the more systems on the Internet determine that the first object belongs to the first target preset category. Therefore, the method of classifying the first object according to the number of category information corresponding to each first target preset category in this embodiment of the invention introduces the category information determined by other Internet business systems for the entity, making the prediction results more accurate and able to be traced back to the corresponding Internet business system, covering more scenarios.

[0008] In one possible implementation, the method further includes:

[0009] Obtain the name information of the first object from the local business system;

[0010] Using a preset classification model, at least one second target preset category corresponding to the first object is determined from the plurality of preset categories based on the name information;

[0011] Based on the number of category information corresponding to each preset category of the first target, the first object is classified, including:

[0012] The first object is classified according to the number of category information corresponding to each first target preset category and the at least one second target preset category.

[0013] In one possible implementation, a preset classification model is used to determine the second target preset category corresponding to the first object from among the plurality of preset categories based on the name information, including:

[0014] By concatenating the name information and the category information set by each business system for the target entity, the concatenated information is obtained;

[0015] The spliced ​​information is input into the preset classification model, and the preset classification model is used to extract category features from the spliced ​​information;

[0016] The second target preset category corresponding to the first object is determined based on the category characteristics using the preset classification model.

[0017] In one possible implementation, obtaining category information set by each business system for the target entity in multiple business systems that match the functional attributes of the target entity to which the first object belongs includes:

[0018] Query the target second object whose entity is consistent with the target entity among multiple second objects stored in a specific business system; wherein, the specific business system is any one of the multiple business systems;

[0019] Extract the category text displayed by the target second object in the specific business system as the category information set by the specific business system for the target entity.

[0020] In one possible implementation, querying at least one first target preset category corresponding to each of the multiple category information items from among multiple preset categories pre-set in the local business system storing the first object includes:

[0021] Map any category information from the plurality of category information to the node containing the associated preset category that has the highest similarity to that category information;

[0022] When the node containing the associated preset category is a leaf node in a pre-constructed category tree, the associated preset category is taken as the first target preset category; wherein, each node in the pre-constructed category tree corresponds one-to-one with the plurality of preset categories;

[0023] If the associated preset category is not a leaf node in the pre-constructed category tree, the preset category located in the child node of the node where the associated preset category is located will be used as the first target preset category.

[0024] In one possible implementation, the first object is classified according to the number of category information corresponding to each preset category of the first target, including:

[0025] When there are multiple target category information corresponding to any first target preset category, the matching probability of any first target preset category matching the first object is set in each target category information dimension for the number of first target preset categories corresponding to each target category information in the business system dimension.

[0026] By summing the matching probabilities of each preset category for the first target under different target category information dimensions, the aggregated probability value of each preset category for the first target is obtained.

[0027] The first object is classified according to the aggregation probability value of each preset category of the first target.

[0028] In one possible implementation, the method further includes:

[0029] Obtain the name information of the first object from the local business system;

[0030] Input the name information into a preset classification model, and output the confidence probability of each preset category and the first object among the multiple preset categories;

[0031] Based on the aggregate probability value of each preset category of the first target, the first object is classified, including:

[0032] The first object is classified according to the aggregate probability value of each preset category of the first target and the confidence probability of each preset category.

[0033] In one possible implementation, the first object is classified according to the aggregate probability value of each preset category of the first target and the confidence probability of each preset category, including:

[0034] When the maximum value of the aggregation probability value is greater than the first preset threshold, the maximum value of the aggregation probability value is determined as the category of the first object corresponding to the first target preset category.

[0035] When the maximum value of the aggregated probability value is less than the first preset threshold and greater than the second preset threshold, the corresponding confidence probability is superimposed on each first target preset category to obtain the comprehensive probability value of each first target preset category.

[0036] The first target preset category corresponding to the maximum value of the comprehensive probability value is determined as the category of the first object;

[0037] When the maximum value of the aggregated probability is less than the second preset threshold, the preset category corresponding to the maximum confidence probability is determined as the category of the first object.

[0038] Secondly, embodiments of the present invention provide an object classification device, disposed in an electronic device, the device comprising:

[0039] The category information acquisition module is used to acquire the category information set by each business system for the target entity in multiple business systems that match the functional attributes of the target entity to which the first object belongs.

[0040] The category query module is used to query at least one first target preset category corresponding to each of the multiple category information in the local business system storing the first object.

[0041] The object classification module is used to classify the first object according to the number of category information corresponding to each preset category of the first target.

[0042] In one possible implementation, the device further includes:

[0043] The name acquisition module is used to acquire the name information of the first object in the local business system;

[0044] The category determination module is used to determine at least one second target preset category corresponding to the first object from the plurality of preset categories based on the name information using a preset classification model;

[0045] The object classification module is specifically used to classify the first object according to the number of category information corresponding to each first target preset category and the at least one second target preset category.

[0046] In one possible implementation, the category determination module includes:

[0047] The splicing submodule is used to splice the name information and the category information set by each business system for the target entity to obtain spliced ​​information;

[0048] The input submodule is used to input the spliced ​​information into the preset classification model and use the preset classification model to extract category features from the spliced ​​information;

[0049] The category determination submodule is used to determine the second target preset category corresponding to the first object based on the category characteristics using the preset classification model.

[0050] In one possible implementation, the category information acquisition module includes:

[0051] The object query submodule is used to query a target second object whose entity is consistent with the target entity among multiple second objects stored in a specific business system; wherein, the specific business system is any one of the multiple business systems;

[0052] The category information extraction submodule is used to extract the category text displayed by the target second object in the specific business system as the category information set by the specific business system for the target entity.

[0053] In one possible implementation, the category query module includes:

[0054] The mapping submodule is used to map any category information among the multiple category information to the node containing the associated preset category that has the highest similarity to that category information;

[0055] The category acquisition submodule is used to take the associated preset category as the first target preset category when the node where the associated preset category is located is a leaf node in a pre-constructed category tree; wherein, each node in the pre-constructed category tree corresponds one-to-one with the plurality of preset categories;

[0056] The category acquisition submodule is further configured to, when the associated preset category is not a leaf node in the pre-constructed category tree, take the preset category located in the child node of the node where the associated preset category is located as the first target preset category.

[0057] In one possible implementation, the object classification module includes:

[0058] The probability setting submodule is used to set the matching probability of any first target preset category matching the first object in each target category information dimension, for each first target preset category corresponding to the number of first target preset categories in the business system dimension where each target category information is located, when there are multiple target category information corresponding to the first target preset category;

[0059] The overlay submodule is used to overlay the matching probabilities of each first target preset category under different target category information dimensions to obtain the aggregated probability value of each first target preset category.

[0060] The object classification submodule is used to classify the first object according to the aggregation probability value of each first target preset category.

[0061] In one possible implementation, the device further includes:

[0062] The name acquisition module is used to acquire the name information of the first object in the local business system;

[0063] The probability acquisition module is used to input the name information into a preset classification model and output the confidence probability of each preset category and the first object among the multiple preset categories.

[0064] The object classification submodule is specifically used to classify the first object according to the aggregate probability value of each preset category of the first target and the confidence probability of each preset category.

[0065] In one possible implementation, the object classification submodule is specifically used for:

[0066] When the maximum value of the aggregation probability value is greater than the first preset threshold, the maximum value of the aggregation probability value is determined as the category of the first object corresponding to the first target preset category.

[0067] When the maximum value of the aggregated probability value is less than the first preset threshold and greater than the second preset threshold, the aggregated probability value and confidence probability corresponding to each first target preset category are superimposed to obtain the comprehensive probability value of each first target preset category.

[0068] The first target preset category corresponding to the maximum value of the comprehensive probability value is determined as the category of the first object;

[0069] When the maximum value of the aggregated probability is less than the second preset threshold, the preset category corresponding to the maximum confidence probability is determined as the category of the first object.

[0070] Thirdly, embodiments of the present invention provide a terminal device, including: at least one processor; and at least one memory communicatively connected to the processor, wherein: the memory stores program instructions executable by the processor, and the processor can execute the method provided in the first aspect by calling the program instructions.

[0071] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing computer instructions that cause the computer to perform the method provided in the first aspect.

[0072] It should be understood that the second to fourth aspects of the embodiments of the present invention are consistent with the technical solutions of the first aspect of the embodiments of the present invention, and the beneficial effects achieved by each aspect and the corresponding feasible implementation are similar, and will not be described again. [Attached Image Description]

[0073] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0074] Figure 1 This is a flowchart of the steps of the object classification method proposed in the embodiments of the present invention;

[0075] Figure 2 This is a schematic diagram of the category tree structure constructed by the local business system storing the first object in one example of the present invention;

[0076] Figure 3 This is a flowchart of another object classification method proposed in an embodiment of the present invention;

[0077] Figure 4 This is a flowchart illustrating an example of an object classification method according to the present invention;

[0078] Figure 5 This is a functional block diagram of the object classification device proposed in the embodiments of the present invention;

[0079] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention;

[0080] Figure 7 This is a schematic diagram of the structure of a terminal device provided in one embodiment of this specification.

Detailed Implementation Methods

[0081] To better understand the technical solutions in this specification, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0082] It should be understood that the described embodiments are merely some, not all, of the embodiments in this specification. All other embodiments obtained by those skilled in the art based on the embodiments in this specification without inventive effort are within the scope of protection of this specification.

[0083] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of this specification. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0084] It should be stated that the user data or related information mentioned in this technical solution is collected and processed in compliance with relevant policies and regulations and with the consent of the corresponding entities. After processing, the data is used in big data application scenarios and cannot be identified to any natural person or have a specific association with their privacy.

[0085] With the development of the Internet, online systems (such as websites and applications) abstract the functions and characteristics of offline entities (such as shops and landmarks), and use functions and data to represent the functions and characteristics of the abstracted offline entities. The objects are created and stored in the database of the online system so that when relevant instructions are received, the online system can directly call the relevant data and display it on the interactive interface.

[0086] The information used to create objects may deviate from the actual situation of the entity, resulting in low accuracy in the functions and data representing the objects. With the development of the internet, online systems are constantly iterating and updating, and the categories marked when creating objects may no longer be applicable to the new category structure of the updated online system.

[0087] Even if a machine learning model is built and the object category information is recalculated using the data stored in the system, the accuracy of the machine learning model in judging the category information is limited by the deviation between the original stored data and the actual situation of the entity, and is limited by the precision of the original stored data itself.

[0088] For example, when creating an object, restaurant A is uniformly categorized as "restaurant". As the system adapts to business development and updates its category architecture, the "restaurant" category is changed to "food" category, and subcategories such as "Chinese food", "Western food", "Southeast Asian food", and "Japanese food" are added under "food". The object categories stored in the business system do not match the updated category architecture, resulting in a situation where, when a user enters the keyword "Chinese food", restaurant A cannot be recommended.

[0089] In view of the above problems, this invention proposes an object classification method for use in electronic devices. Figure 1 This is a flowchart of the object classification method proposed in the embodiments of the present invention, as follows: Figure 1 As shown, the steps include:

[0090] Electronic devices can include terminals, servers, computer clusters, etc.

[0091] Step S101: Obtain the category information set by each business system for the target entity in multiple business systems that match the functional attributes of the target entity to which the first object belongs.

[0092] The first object is an object created by the local system for the target entity. The local system is the system where the object to be classified resides. The first object can refer to the object to be classified.

[0093] A business system can be a website, an application, etc. When the application scenario of a specific business system matches the functional attributes of the target entity, the specific business system can be considered to have matched the functional attributes of the target entity. For example, if the target entity is a Chinese restaurant, its functional attributes include catering and location. A group-buying application includes applications that recommend restaurants; therefore, a group-buying application is a business system that matches the functional attributes of a Chinese restaurant. Similarly, a map application includes applications that display the location of Chinese restaurants; therefore, a map application is also a business system that matches the functional attributes of a Chinese restaurant.

[0094] The category information set by the existing business system for the target entity can be the result of collecting relevant features of the target entity and classifying the target entity based on relevant features through machine learning; the category information set by the existing business system for the target entity can also be the category information verified by the user; or the category information set by the existing business system for the target entity can also be the category information determined by the business system after offline research.

[0095] In one example of the present invention, the functional attribute matching of the target entity to which the first object belongs can be verified by sampling to verify whether the category information of the object labeling in the business system is accurate. When the category information set by the business system accurately represents the object category, the category information set for the target entity in the business system can be mined.

[0096] In one example of the present invention, data mining tools such as big data semantic intelligent analysis platforms can be used to obtain category information of the target entity in other business systems.

[0097] Step S102: In the local business system storing the first object, query at least one first target preset category corresponding to each of the multiple category information preset categories.

[0098] Different business systems maintain category systems suitable for their application scenarios. Since the application scenarios of different business systems are different, the category information set for the same type of entity in the category systems maintained by different business systems is different. Therefore, it is necessary to map the category information obtained from other business systems to the category system maintained by the system where the first object is located, and query which preset categories and category information in the business system where the first object is located correspond to the same type of entity.

[0099] In one example of this invention, the local business system storing the first object is an online shopping platform for electronic products. Under the category system maintained by this platform, the category for e-books is e-readers. A comprehensive e-commerce platform, as a business system matching the functional attributes of e-books, maintains a category system where e-books should be categorized as electronic products. Electronic products and e-readers refer to the same type of entity. The category information of the target entity obtained from the comprehensive e-commerce platform is "electronic products," and the first target preset category corresponding to the electronic product, "e-readers," is queried from multiple preset categories pre-set in the local business system storing the first object.

[0100] The local business system storing the first object can classify e-books and directly obtain the category information set by other business systems for the target entity to which the first object belongs. Since the category information is obtained after other business systems analyze the target entity, the first target preset category corresponding to multiple category information in the local business system storing the first object can represent the target entity type that is generally recognized by most business systems.

[0101] Step S103: Classify the first object according to the number of category information corresponding to each first target preset category.

[0102] The number of category information corresponding to the first target preset category can refer to the number of business systems that select the first target preset category to represent the type of target entity to which the first object belongs.

[0103] In one example of this invention, the business systems matching the functional attributes of the target entity to which the first object belongs include: a group-buying platform, a map system, and a food recommendation forum. The target entity name of the first object stored in the local business system is "Beitang Xiaoxiaopu Chinese Restaurant". The category information of the corresponding target entity mined from the group-buying platform, map system, and food recommendation forum are Chinese food, gourmet food, and modified Sichuan cuisine, respectively. Among the multiple preset categories pre-set in the local business system, the first target preset categories corresponding to the above category information, namely Chinese food, gourmet food, and modified Sichuan cuisine, are: staple food, staple food, and snacks, respectively. Among them, the category information corresponding to staple food includes Chinese food and gourmet food. It can be seen that the group-buying platform and map system in the three business systems matching the functional attributes of the target entity to which the first object belongs objectively determine that the type of the target entity to which the first object belongs is expressed as staple food in the local business system storing the first object. It can be considered that most business systems matching the functional attributes of the target entity objectively determine that the type of the target entity to which the first object belongs is expressed as staple food in the local business system storing the first object.

[0104] This invention classifies a first object based on the target entity to which it belongs. It mines category information set for the target entity by other business systems on the internet that store data related to the target entity. This category information is mapped from other business systems to the local business system storing the first object. The local business system queries the corresponding first target preset category. Based on the number of category information corresponding to the first target preset category, the category to which the target entity should belong under most internet business systems is selected. This invention, through the above method, quickly determines the category of the first object, unaffected by the accuracy of the data used by the local business system storing the first object.

[0105] The object classification method proposed in this invention addresses a first object to be classified in a local system by querying category information determined by multiple other systems on the Internet for the target entity to which the first object belongs. This yields the category to which the target entity should belong under the evaluation systems of other systems on the Internet. The category information determined by other Internet systems for the target entity is mapped to the local system to obtain a first target preset category. Since the first target preset category represents the local system's representation of the type determined by other Internet systems for the target entity to which the first object belongs, the larger the number of category information corresponding to the first target preset category, the more systems on the Internet determine that the first object belongs to the first target preset category. Based on the above analysis, this invention classifies the first object according to the number of category information corresponding to the first target preset category. The resulting category of the first object, classified from the perspective of the entity to which the first object belongs, is the result of factual determination of the category of the entity to which the first object belongs by most business systems, and can quickly obtain accurate classification results.

[0106] Another embodiment of the present invention proposes an implementation method for obtaining category information, wherein step S101 includes sub-steps S201 to S202.

[0107] Step S201: Query the target second object whose entity is consistent with the target entity among the multiple second objects stored in a specific business system; wherein, the specific business system is any one of the multiple business systems.

[0108] Obtain the target entity to which the first object to be classified belongs; search the Internet for business systems that have created objects for the target entity, and select a specific business system among multiple business systems that match the functional attributes of the target entity; query the second object corresponding to the target entity in the specific business system, and extract the category text of the second object as the category information of the target entity.

[0109] In one example of the present invention, querying a target second object whose entity among multiple second objects stored in a specific business system is consistent with the target entity can be performed using the following method: extracting text information from the window interface displaying the second object in the specific business system, such as the address, name, and contact information of the second object; retrieving relevant data about the first object stored in the local business system storing the first object, such as the address, name, and contact information of the first object; calculating the similarity between the text information and the relevant data; and determining whether the entity to which the second object belongs is consistent with the target entity to which the first object belongs. When the entity to which the second object belongs is consistent with the target entity, the second object is determined to be the target second object, and the category text displayed in the specific business system is obtained by filtering from the extracted text information from the window interface displaying the second object.

[0110] In one example of this invention, a specific business system contains multiple target second objects. For instance, a group-buying platform system includes a tourist attraction recommendation architecture and a food delivery architecture. The tourist attraction recommendation architecture has a browsing page for recommended restaurants around tourist attractions, and this page displays a second object A created based on restaurant A. The food delivery architecture has a window displaying a second object B created based on restaurant A. The category text "tourist snacks" for second object A and the category text "popular meat pie snacks" for second object B can be extracted as category information set by the group-buying platform system for restaurant A.

[0111] Step S202: Extract the category text displayed by the target second object in the specific business system as the category information set by the specific business system for the target entity.

[0112] This invention, based on the text information displayed by various business systems on the Internet, determines a specific business system that has a second object corresponding to a target entity, extracts the category text displayed by the specific business system for the second object, and obtains the category information set by the business system that matches the functional attributes of the target entity for the target entity. The above-mentioned method for obtaining the category information set by the business system that matches the functional attributes of the target entity for the target entity directly extracts useful information from the information displayed on the Internet, which is simple and convenient.

[0113] In another embodiment of the present invention, a method for querying at least one first target preset category corresponding to multiple category information is proposed, wherein step S102 includes sub-steps S301 to S303.

[0114] S301: Map any category information from the plurality of category information to the node containing the associated preset category that has the highest similarity to that category information.

[0115] S302: When the node containing the associated preset category is a leaf node in a pre-constructed category tree, the associated preset category is taken as the first target preset category; wherein, each node in the pre-constructed category tree corresponds one-to-one with the plurality of preset categories.

[0116] S303: When the associated preset category is not a leaf node in the pre-constructed category tree, the preset category located in the child node of the node where the associated preset category is located is taken as the first target preset category.

[0117] A category tree can be constructed based on the coarse-grained nature of the business represented by different preset categories in the local business system.

[0118] Figure 2 This is a schematic diagram of the category tree structure constructed by the local business system storing the first object in one example of the present invention, such as... Figure 2 As shown, the first-level preset categories, "Food and Shopping," serve as the root nodes of the category tree, with no parent nodes. The second-level preset categories, "Chinese Food and Hot Pot," are child nodes of the "Food and Shopping" preset category. The third-level preset categories, "Cantonese Cuisine and Sichuan Cuisine," are child nodes of the "Chinese Food" preset category. The nodes containing "Cosmetics," "Clothing," and "Electronic Products" in the second-level preset categories are child nodes of the "Shopping" preset category. Similarly, the nodes containing "Mobile Phones," "Tablets," "Computers," and "Electronic Readers" in the third-level preset categories are child nodes of the "Electronic Products" preset category. The third-level preset categories, "Cantonese Cuisine," "Sichuan Cuisine," "Skincare," and "Makeup," are leaf nodes of the category tree. These preset categories, located as leaf nodes of the category tree, represent the most accurate way to express the type of the first object in the local business system application scenario where the first object is stored.

[0119] Taking the preset category of electrical appliances as an example, the sub-nodes of the node containing electrical appliances include: nodes containing mobile phones, tablets, e-readers, and computers. Among them, mobile phones, tablets, e-readers, and computers can all be subcategories of electronic products. It can be understood that electrical appliances include mobile phones, tablets, e-readers, and computers.

[0120] based on Figure 2 The category tree shown illustrates sub-steps S301 to S303.

[0121] In one example of the present invention, a business system that matches the functional attributes of a target entity includes an electronics mall and a comprehensive e-commerce platform. The target entity to which the first object belongs is an e-book store. The electronics mall creates a second object A corresponding to the e-book store. The category information obtained by extracting the category text of the second object A is e-reader. The comprehensive e-commerce platform creates a second object B corresponding to the e-book store. The category information obtained by extracting the category text of the second object B is home appliance.

[0122] exist Figure 2 The category tree shown shows that the preset category reader has the highest similarity to the e-reader. Therefore, the preset category reader is determined to be the associated preset category of the e-reader, and the e-reader is mapped to the node containing the reader. Figure 2 The category tree shows that the preset category of electrical appliances has the highest similarity to household appliances. Therefore, the preset category of electrical appliances is determined to be the preset category associated with the category information household appliances, and the household appliances are mapped to the node where the electrical appliances are located.

[0123] The e-reader is mapped to the node where the reader is located. Figure 2 The leaf nodes of the category tree shown indicate that the node containing the reader is the first target preset category. The category information set by an electronics marketplace for the second object A is... Figure 2 The category tree shown corresponds to a preset category. It can be understood that, from the perspective of a certain e-commerce business system, the target entity of the first object, the e-book store, is classified as a reader in the local business system where the first object is stored.

[0124] Home appliances are mapped to the node where the appliance product is located. Figure 2 The non-leaf nodes of the category tree shown are designated as the first target preset category for the child nodes of the node containing the electrical appliance product: the node containing the preset category mobile phone, the node containing the preset category tablet, the node containing the preset category e-reader, and the node containing the preset category computer. The category information set by a certain comprehensive e-commerce platform for the second object B is... Figure 2 The four preset categories in the category tree shown correspond to what can be understood as follows: from the perspective of a certain comprehensive e-commerce platform business system, the category of the target entity e-book store to which the first object belongs in the local business system storing the first object may be: mobile phone, tablet, e-reader, or computer.

[0125] The number of category information corresponding to each preset category of the first target is counted, resulting in the following categories: mobile phones: electrical appliances; tablets: electrical appliances; e-readers: electrical appliances and e-readers. Since e-readers correspond to both electrical appliances and e-readers, the category of "reader" is determined from the perspective of both business systems that have created e-book stores. This indicates that users generally agree that "reader" can be a category belonging to the target entity's e-book store.

[0126] This invention proposes another object classification method, which classifies the first object from the dimension of the target entity to which the first object belongs, and from the data dimension adopted by the local business system storing the first object.

[0127] To implement the object classification method proposed in this invention more intelligently, the applicant constructs a neural network model to be trained based on a text classification network (textCNN). Entity samples are collected, such as shops and landmarks. The names of the shops and landmarks are used as training samples, and the types of the collected entity samples are labeled as control samples. For example, shop categories can be labeled according to their functional attributes: clothing shops, bag shops, jewelry shops, etc.

[0128] An example of the present invention pre-constructs as follows Figure 2 The category tree shown can be used to select the preset category that best matches the store's functional attributes as the entity sample type based on the similarity between the store's functional attributes and the preset categories recorded in the nodes of the category tree.

[0129] The constructed neural network model is trained multiple times using training samples and control samples until the model converges, resulting in a preset classification model that can perform feature analysis on the names of shops or landmark buildings and output the type of shops or landmark buildings.

[0130] Figure 3 This is a flowchart of another object classification method proposed in an embodiment of the present invention, as follows: Figure 3 As shown, the steps include:

[0131] Step S401: Obtain the name information of the first object from the local business system.

[0132] The name information of the first object can be obtained by extracting the text information displayed in the window interface of the local business system storing the first object.

[0133] Step S402: Using a preset classification model, determine at least one second target preset category corresponding to the first object from among the multiple preset categories based on the name information.

[0134] The pre-defined classification model, textCNN, uses multiple convolutional kernels of different sizes to extract features from name information, capture the local relevance of names, and complete the text classification task. The second target pre-defined category is the one with the highest relevance to the name information among multiple pre-defined categories.

[0135] Step S403: Obtain the category information set by each business system for the target entity in multiple business systems that match the functional attributes of the target entity to which the first object belongs.

[0136] Step S404: In the local business system storing the first object, query at least one first target preset category corresponding to each of the multiple category information preset categories.

[0137] The electronic device executes steps S401 to S402 in any order with the electronic device executes steps S403 to S404. It can be understood that the electronic device can execute steps S401 to S402 first to obtain the second target preset category; or it can execute steps S403 to S404 first to obtain the first target preset category; or it can execute the relevant steps simultaneously to obtain the first target preset category and the second target preset category.

[0138] The electronic device executes steps S401 to S402 to obtain the type information of the first object from the data dimension adopted by the local business system storing the first object; the electronic device executes steps S403 to S404 to obtain the type information of the first object from the dimension of the target entity to which the first object belongs.

[0139] Step S405: Classify the first object according to the number of category information corresponding to each first target preset category and the at least one second target preset category.

[0140] The electronic device executes step S405, which integrates the data dimensions adopted by the local business system storing the first object to obtain information related to the type of the first object, and obtains information related to the type of the first object from the dimension of the target entity to which the first object belongs.

[0141] by Figure 2 Taking the category tree shown as an example, the first target preset category, obtained from the dimension of the target entity to which the first object belongs, includes skincare and makeup. The first target preset category, makeup, corresponds to two types of information. It can be understood that multiple business systems on the Internet that match the functional attributes of the target entity to which the first object belongs generally determine that makeup is more representative of the category of the target entity to which the first object belongs. The second target preset category output by the preset classification model includes makeup. By integrating the above information, the first object is classified, and the category of the first object is determined to be makeup.

[0142] An example of the present invention proposes an implementation method for obtaining a second target preset category based on name information, wherein step S402 includes sub-steps S501 to S503.

[0143] S501: Combine the name information and the category information set by each business system for the target entity to obtain the combined information.

[0144] For example, the name information of the first object is extracted as "Beitang Xiaoxiaopu". Multiple business systems that match the functional attributes of the target entity "Beitang Xiaoxiaopu" include: a group-buying platform and a food recommendation website. The group-buying platform sets the category information for "Beitang Xiaoxiaopu" as "Chinese restaurant," while the food recommendation website sets the category information as "seafood porridge hotspot." The information "Beitang Xiaoxiaopu," "Chinese restaurant," and "seafood porridge hotspot" are then concatenated to obtain the concatenated information.

[0145] The word2vec model, VSM vector space model, neural network language model (NNLM) and other models can be used to convert Beitang Xiaoxiaopu, Chinese restaurants and seafood porridge check-in spots into vector representations, or to convert spliced ​​information into vector representations.

[0146] Step 502: Input the spliced ​​information into the preset classification model, and use the preset classification model to extract category features from the spliced ​​information.

[0147] Step S503: Using the preset classification model, determine the second target preset category corresponding to the first object based on the category characteristics.

[0148] The input vector of the aforementioned pre-defined classification model includes features from the data used by the local business system to create the first object, and features from the category information determined by other internet business systems for the first object. The pre-defined classification model extracts features from the concatenated information and fuses the features from both. Even if there are errors in the data used by the local business system to create the first object, the pre-defined classification model can correct these errors based on the category information extracted from other internet business systems, thus ensuring the accuracy of the output of the pre-defined classification model.

[0149] Another embodiment of the present invention proposes that after querying at least one first target preset category corresponding to each of the multiple category information in the local system storing the first object, the matching probability of mapping the category information to each first target preset category can be set according to the number of first target preset categories corresponding to the category information.

[0150] When the number of target category information corresponding to any first target preset category is one, the probability of mapping from category information to the first target preset category is set to 100%.

[0151] When there are multiple target category information corresponding to any first target preset category, the matching probability of any first target preset category matching the first object is set for the number of first target preset categories corresponding to the business system dimension of each target category information.

[0152] For example, such as Figure 2The category information corresponding to the first target preset category "makeup" shown includes: beauty. Within its respective business system dimension, the first target preset category "beauty" also includes skincare. The category information "beauty" within its respective business system is stored in an external business system other than the first object. It can be understood that, within the business system dimension to which the category information "beauty" belongs, the probability of determining the type of the first object as "makeup" is 50%.

[0153] The matching probabilities of each first target preset category under different target category information dimensions are superimposed to obtain the aggregate probability value of each first target preset category.

[0154] When there are enough business systems on the Internet that match the functional attributes of the target entity to which the first object belongs, the aggregated probability value obtained by superimposing the matching probabilities of each first target preset category under different target category information dimensions can be regarded as the support rate of websites or applications on the Internet related to the target entity to which the first object belongs for the first target preset category being the first object category.

[0155] The first object is classified according to the aggregation probability value of each preset category of the first target.

[0156] This invention proposes a preset classification model that can output the confidence probability that different preset categories can be identified as a second target preset category. The steps include:

[0157] Step S601: Obtain the name information of the first object from the local business system.

[0158] Step S602: Input the name information into a preset classification model and output the confidence probability of each preset category and the first object.

[0159] The pre-defined classification model textCNN uses multiple convolutional kernels of different sizes to extract semantic features from name information, calculates the correlation between semantic features and each pre-defined category, and outputs the confidence probability of each pre-defined category and the first object based on the correlation between semantic features and each pre-defined category.

[0160] The confidence probability of a preset category represents the probability that the preset category, calculated by the preset classification model based on the relevant data of the first object (such as name information), can accurately represent the type of the first object.

[0161] Step S603: Classify the first object according to the aggregate probability value of each preset category of the first target and the confidence probability of each preset category.

[0162] Step S603 includes sub-steps S701 to S703:

[0163] S701: When the maximum value of the aggregation probability value is greater than the first preset threshold, the maximum value of the aggregation probability value is determined as the category of the first object corresponding to the first target preset category.

[0164] S702: When the maximum value of the aggregated probability value is less than the first preset threshold and greater than the second preset threshold, the confidence probability corresponding to each first target preset category is superimposed to obtain the comprehensive probability value of each first target preset category; the first target preset category corresponding to the maximum value of the comprehensive probability value is determined as the category of the first object.

[0165] S703: When the maximum value of the aggregated probability value is less than the second preset threshold, the preset category corresponding to the maximum confidence probability is determined as the category of the first object.

[0166] The local business system categorizes the first object based on relevant data, obtaining a confidence probability that the first object belongs to a preset category from the local business system's perspective. It then incorporates category information determined by other internet business systems to obtain a confidence probability that the first object belongs to a preset category from the external system's perspective. Finally, it combines the identification results from both the local and external systems to obtain a comprehensive probability value for the first object belonging to the preset category. This method, by combining results from both local and external systems, is more reliable. The textCNN model algorithm uses general features and incorporates category information from internet intelligence, covering more scenarios. Furthermore, it introduces more intuitive internet category information, making it easier to trace the prediction results. Due to the convenience of determining the category information of the first object from the external system's perspective, optimization of subsequent steps can be completed quickly once the model is stable.

[0167] Figure 4 This is a flowchart of an example object classification method of the present invention, such as... Figure 4 As shown, the steps include:

[0168] K11: Building as Figure 2 The category tree shown includes multiple preset categories: catering, shopping, Chinese food, hot pot, beauty, clothing, electrical appliances, Cantonese cuisine, Sichuan cuisine, etc.

[0169] K12: Obtain the target entity to which the first object to be classified belongs: e-book store. Internet mining yielded multiple business systems matching the functional attributes of the target entity, including an online shopping platform for electronic products, a comprehensive e-commerce platform, and a map system.

[0170] K13: Extract category information for e-book stores from an online electronics shopping platform, a comprehensive e-commerce platform, and a map system. The online electronics shopping platform categorizes e-book stores as e-readers, the comprehensive e-commerce platform categorizes them as home appliances, and the map system categorizes them as digital product stores.

[0171] K14: Map category information 1 to category information k (e-readers, home appliances, digital product stores) to the business system where the first object to be classified resides. Figure 2 The category tree shown shows that the first target preset category corresponding to e-readers includes: e-readers. The matching probability of the category information from a certain online shopping platform for electronic products pointing to the first target preset category e-readers is set to 100%.

[0172] K15: The first target preset category for querying home appliances includes: mobile phones, tablets, e-readers, and computers. The matching probability of the first target preset category "e-reader" is set to 25% when the category information of a certain comprehensive e-commerce platform points to the first target preset category "mobile phone", the matching probability of the first target preset category "tablet", and so on.

[0173] K16: The first target preset category for querying digital product stores includes mobile phones, tablets, e-readers, and computers. Set the matching probability of the first target preset category "reader" from the category information of a certain map system to the first target preset category "mobile phone" ...

[0174] K17: The local business system storing the first object extracts the name information of the first object, concatenates the name information, e-reader, home appliance, and digital product store to obtain the concatenated information, inputs the concatenated information into the preset classification model, and outputs as follows: Figure 2 The confidence probability distribution of each preset category and the first object in the category tree shown is as follows: [Food and Beverage - 0%, Shopping - 0%, Chinese Food - 2%, Hot Pot - 2%, Beauty Products - 3%, Clothing - 1%, Electrical Appliances - 0%, Cantonese Cuisine - 1%, Sichuan Cuisine - 1%, Skincare - 2%, Makeup - 1%, Mobile Phones - 10%, Tablets - 10%, Computers - 5%, Readers - 62%].

[0175] K18: By superimposing the matching probabilities calculated from K14 to K16, we obtain the matching probability distribution of each first target preset category based on other business systems besides the local business system storing the first object: [Mobile phone - 50%, Tablet - 50%, Computer - 50%, Reader - 150%].

[0176] K19: When the aggregated probability value (reader -150%) is greater than the first preset threshold, the reader is identified as the category of the first object. When the reader -150% is less than the first preset threshold but greater than the second preset threshold, the matching probability distributions are superimposed: [Mobile phone -50%, Tablet -50%, Computer -50%, Reader -150%] and the confidence probability distributions: [Dining -0%, Shopping -0%, Chinese food -2%, Hot pot -2%, Beauty -3%, Clothing -1%, Electrical appliances -0%, Cantonese cuisine -1%, Sichuan cuisine -1%, Skincare -2%, Makeup -1%, Mobile phone -10%, Tablet -10%, Computer -5%, Reader -62%], to obtain the comprehensive probability value for each first target preset category; the reader corresponding to the maximum comprehensive probability value is identified as the category of the first object. When the reader -150% is less than the second preset threshold, the preset category corresponding to the maximum confidence probability is identified as the category of the first object.

[0177] Figure 5 This is a functional block diagram of the object classification device proposed in an embodiment of the present invention. The object classification device is installed in an electronic device, such as... Figure 5 As shown, the device includes:

[0178] The category information acquisition module 51 is used to acquire the category information set by each business system for the target entity in multiple business systems that match the functional attributes of the target entity to which the first object belongs.

[0179] Category query module 52 is used to query at least one first target preset category corresponding to each of the multiple category information in a local business system that stores the first object.

[0180] The object classification module 53 is used to classify the first object according to the number of category information corresponding to each first target preset category.

[0181] Figure 5 The object classification apparatus provided in the illustrated embodiment can be used to execute this specification. Figures 1 to 4 The implementation principle and technical effects of the method embodiment shown can be further referred to the relevant description in the method embodiment.

[0182] Optionally, the device further includes:

[0183] The name acquisition module is used to acquire the name information of the first object in the local business system;

[0184] The category determination module is used to determine at least one second target preset category corresponding to the first object from the plurality of preset categories based on the name information using a preset classification model;

[0185] The object classification module is specifically used to classify the first object according to the number of category information corresponding to each first target preset category and the at least one second target preset category.

[0186] Optionally, the category determination module includes:

[0187] The splicing submodule is used to splice the name information and the category information set by each business system for the target entity to obtain spliced ​​information;

[0188] The input submodule is used to input the spliced ​​information into the preset classification model and use the preset classification model to extract category features from the spliced ​​information;

[0189] The category determination submodule is used to determine the second target preset category corresponding to the first object based on the category characteristics using the preset classification model.

[0190] Optionally, the category information acquisition module includes:

[0191] The object query submodule is used to query a target second object whose entity is consistent with the target entity among multiple second objects stored in a specific business system; wherein, the specific business system is any one of the multiple business systems;

[0192] The category information extraction submodule is used to extract the category text displayed by the target second object in the specific business system as the category information set by the specific business system for the target entity.

[0193] Optionally, the category query module includes:

[0194] The mapping submodule is used to map any category information among the multiple category information to the node containing the associated preset category that has the highest similarity to that category information;

[0195] The category acquisition submodule is used to take the associated preset category as the first target preset category when the node where the associated preset category is located is a leaf node in a pre-constructed category tree; wherein, each node in the pre-constructed category tree corresponds one-to-one with the plurality of preset categories;

[0196] The category acquisition submodule is further configured to, when the associated preset category is not a leaf node in the pre-constructed category tree, take the preset category located in the child node of the node where the associated preset category is located as the first target preset category.

[0197] Optionally, the object classification module includes:

[0198] The probability setting submodule is used to set the matching probability of any first target preset category matching the first object in each target category information dimension, for each first target preset category corresponding to the number of first target preset categories in the business system dimension where each target category information is located, when there are multiple target category information corresponding to the first target preset category;

[0199] The overlay submodule is used to overlay the matching probabilities of each first target preset category under different target category information dimensions to obtain the aggregated probability value of each first target preset category.

[0200] The object classification submodule is used to classify the first object according to the aggregation probability value of each first target preset category.

[0201] Optionally, the device further includes:

[0202] The name acquisition module is used to acquire the name information of the first object in the local business system;

[0203] The probability acquisition module is used to input the name information into a preset classification model and output the confidence probability of each preset category and the first object among the multiple preset categories.

[0204] The object classification submodule is specifically used to classify the first object according to the aggregate probability value of each preset category of the first target and the confidence probability of each preset category.

[0205] Optionally, the object classification submodule is specifically used for:

[0206] When the maximum value of the aggregation probability value is greater than the first preset threshold, the maximum value of the aggregation probability value is determined as the category of the first object corresponding to the first target preset category.

[0207] When the maximum value of the aggregated probability value is less than the first preset threshold and greater than the second preset threshold, the aggregated probability value and confidence probability corresponding to each first target preset category are superimposed to obtain the comprehensive probability value of each first target preset category.

[0208] The first target preset category corresponding to the maximum value of the comprehensive probability value is determined as the category of the first object;

[0209] When the maximum value of the aggregated probability is less than the second preset threshold, the preset category corresponding to the maximum confidence probability is determined as the category of the first object.

[0210] The apparatus provided in the above embodiments is used to execute the technical solutions of the above method embodiments. Its implementation principle and technical effects can be further referred to the relevant descriptions in the method embodiments, and will not be repeated here.

[0211] The apparatus provided in the above embodiments may be, for example, a chip or a chip module. The apparatus provided in the above embodiments is used to execute the technical solutions of the above-described method embodiments. Its implementation principles and technical effects can be further referred to the relevant descriptions in the method embodiments, and will not be repeated here.

[0212] Regarding the modules / units included in the various devices described in the above embodiments, they can be software modules / units, hardware modules / units, or a combination of both. For example, for devices applied to or integrated into a chip, all modules / units can be implemented using hardware methods such as circuits, or at least some modules / units can be implemented using software programs running on a processor integrated within the chip, while the remaining modules / units can be implemented using hardware methods such as circuits. For devices applied to or integrated into a chip module, all modules / units can be implemented using hardware methods such as circuits. Different modules / units can be located in the same component (e.g., chip, circuit module, etc.) or different components of the chip module, or at least some modules / units can be implemented using hardware methods such as circuits. The implementation can be done through a software program that runs on a processor integrated within the chip module, while the remaining modules / units can be implemented using hardware methods such as circuits. For each device applied to or integrated into an electronic device, each of its modules / units can be implemented using hardware methods such as circuits. Different modules / units can be located in the same component (e.g., chip, circuit module, etc.) or different components within the electronic device. Alternatively, at least some modules / units can be implemented through a software program that runs on a processor integrated within the electronic device, while the remaining (if any) modules / units can be implemented using hardware methods such as circuits.

[0213] Figure 6This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The electronic device 600 includes a processor 610, a memory 611, and a computer program stored in the memory 611 and executable on the processor 610. When the processor 610 executes the program, it implements the steps in the aforementioned method embodiment. The electronic device provided in this embodiment can be used to execute the technical solution of the method embodiment shown above. Its implementation principle and technical effect can be further referred to the relevant description in the method embodiment, which will not be repeated here.

[0214] Figure 7 This is a schematic diagram of the structure of a terminal device provided in one embodiment of this specification, as shown below. Figure 7 As shown, the terminal device described above may include at least one processor; and at least one memory communicatively connected to the processor, wherein the memory stores program instructions executable by the processor, and the processor can execute this specification by calling the program instructions. Figures 1-4 The illustrated embodiment provides an object classification method.

[0215] It is understood that the structures illustrated in the embodiments of the present invention do not constitute a specific limitation on the terminal device 100. In other embodiments of the present invention, the terminal device 100 may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0216] like Figure 7 As shown, the terminal device 100 may include a processor 110, an external memory interface 120, an internal memory 121, a mobile communication module 150, a wireless communication module 160, an indicator 192, a camera 193, a display screen 194, etc.

[0217] Processor 110 may include one or more processing units, such as application processors (APs), modem processors, graphics processing units (GPUs), image signal processors (ISPs), controllers, video codecs, digital signal processors (DSPs), baseband processors, and / or neural network processing units (NPUs). These different processing units may be independent devices or integrated into one or more processors.

[0218] The controller can generate operation control signals based on the instruction opcode and timing signals to complete the control of instruction fetching and execution.

[0219] The processor 110 may also include a memory for storing instructions and data. In some embodiments, the memory in the processor 110 is a cache memory. This memory can store instructions or data that the processor 110 has just used or that are used repeatedly. If the processor 110 needs to use the instruction or data again, it can retrieve it directly from the memory. This avoids repeated accesses, reduces the waiting time of the processor 110, and thus improves the efficiency of the system.

[0220] Processor 110 executes various functional applications and data processing by running programs stored in internal memory 121, such as implementing the present invention. Figures 1-4 The illustrated embodiment provides an object classification method.

[0221] The wireless communication function of the terminal device 100 can be implemented through antenna 1, antenna 2, mobile communication module 150, wireless communication module 160, modem processor and baseband processor, etc.

[0222] Antennas 1 and 2 are used to transmit and receive electromagnetic wave signals. Each antenna in terminal device 100 can be used to cover one or more communication frequency bands. Different antennas can also be multiplexed to improve antenna utilization. For example, antenna 1 can be multiplexed as a diversity antenna for a wireless local area network. In some other embodiments, the antennas can be used in conjunction with a tuning switch.

[0223] Terminal device 100 implements display functions through a GPU, display screen 194, and application processor. The GPU is a microprocessor for image processing, connected to the display screen 194 and the application processor. The GPU is used to perform mathematical and geometric calculations and for graphics rendering. Processor 110 may include one or more GPUs, which execute program instructions to generate or modify display information.

[0224] Display screen 194 is used to display images, videos, etc. Display screen 194 includes a display panel. The display panel may be a liquid crystal display (LCD), an organic light-emitting diode (OLED), an active-matrix organic light-emitting diode (AMOLED), a flexible light-emitting diode (FLED), a miniature LED, a microLED, a quantum dot light-emitting diode (QLED), etc. In some embodiments, terminal device 100 may include one or N displays 194, where N is a positive integer greater than 1.

[0225] Terminal device 100 can perform shooting functions through ISP, camera 193, video codec, GPU, display 194 and application processor.

[0226] The ISP (Image Signal Processor) is used to process data fed back from the camera 193. For example, when taking a picture, the shutter is opened, and light is transmitted through the lens to the camera's photosensitive element. The light signal is converted into an electrical signal, and the camera's photosensitive element transmits the electrical signal to the ISP for processing, transforming it into an image visible to the naked eye. The ISP can also perform algorithmic optimization of image noise, brightness, and skin tone. The ISP can also optimize parameters such as exposure and color temperature of the shooting scene. In some embodiments, the ISP can be set in the camera 193.

[0227] Camera 193 is used to capture still images or videos. An object is projected onto a photosensitive element by generating an optical image through the lens. The photosensitive element can be a charge-coupled device (CCD) or a complementary metal-oxide-semiconductor (CMOS) phototransistor. The photosensitive element converts the light signal into an electrical signal, which is then passed to an ISP for conversion into a digital image signal. The ISP outputs the digital image signal to a DSP for processing. The DSP converts the digital image signal into image signals in standard RGB, YUV, or other formats. In some embodiments, the terminal device 100 may include one or N cameras 193, where N is a positive integer greater than 1.

[0228] A digital signal processor (DSP) is used to process digital signals. Besides digital image signals, it can also process other digital signals. For example, when terminal device 100 selects a frequency, the DSP can perform Fourier transforms on the frequency energy.

[0229] Video codecs are used to compress or decompress digital video. Terminal device 100 may support one or more video codecs. Thus, terminal device 100 can play or record videos in various encoding formats, such as Moving Picture Experts Group (MPEG) 1, MPEG 2, MPEG 3, MPEG 4, etc.

[0230] Internal memory 121 can be used to store computer executable program code, which includes instructions. Internal memory 121 may include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback, image playback, etc.), etc. The data storage area may store data created during the use of terminal device 100 (such as audio data, phonebook, etc.). Furthermore, internal memory 121 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, universal flash storage (UFS), etc. Processor 110 executes various functional applications and data processing of terminal device 100 by running instructions stored in internal memory 121 and / or instructions stored in memory located in the processor.

[0231] This invention provides a computer-readable storage medium storing computer instructions that cause a computer to execute the present specification. Figures 1-4 The illustrated embodiment provides an object classification method.

[0232] The aforementioned computer-readable storage medium may be any combination of one or more computer-readable media. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) or flash memory, optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium that contains or stores a program that may be used by or in connection with an instruction execution system, apparatus, or device.

[0233] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including—but not limited to—electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of transmitting, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0234] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including—but not limited to—wireless, wire, optical fiber, radio frequency (RF), etc., or any suitable combination thereof.

[0235] Computer program code for performing the operations described herein can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as "C" or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0236] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0237] In the description of the embodiments of the present invention, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this specification. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0238] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this specification, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0239] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this specification includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which the embodiments of this specification pertain.

[0240] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0241] It should be noted that the terminals involved in the embodiments of the present invention may include, but are not limited to, personal computers (PCs), personal digital assistants (PDAs), wireless handheld devices, tablet computers, mobile phones, MP3 players, MP4 players, etc.

[0242] In the several embodiments provided in this specification, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0243] Furthermore, the functional units in the various embodiments of this specification can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units.

[0244] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods described in the various embodiments of this specification. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0245] The above description is merely a preferred embodiment of this specification and is not intended to limit this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of protection of this specification.

Claims

1. An object classification method, characterized in that, Applied to electronic devices, the method includes: For offline entities, data and functions are used to express the characteristics and functions of the entities, so as to create objects that represent the corresponding entities online; Obtain the category information set by each business system for the target entity in multiple business systems that match the functional attributes of the target entity to which the first object belongs; In the local business system storing the first object, query at least one first target preset category corresponding to each of the multiple category information; The first object is classified according to the number of category information corresponding to each preset category of the first target; In the local business system storing the first object, querying at least one first target preset category corresponding to each of the multiple category information preset categories includes: Map any category information from the plurality of category information to the node containing the associated preset category that has the highest similarity to that category information; When the node containing the associated preset category is a leaf node in a pre-constructed category tree, the associated preset category is taken as the first target preset category; wherein, each node in the pre-constructed category tree corresponds one-to-one with the plurality of preset categories; When the associated preset category is not a leaf node in the pre-constructed category tree, the preset category located in the child node of the node where the associated preset category is located is taken as the first target preset category; the first object is classified according to the number of category information corresponding to each first target preset category, including: When there are multiple target category information corresponding to any first target preset category, the matching probability of any first target preset category matching the first object is set in each target category information dimension for the number of first target preset categories corresponding to each target category information in the business system dimension. By summing the matching probabilities of each preset category for the first target under different target category information dimensions, the aggregated probability value of each preset category for the first target is obtained. The first object is classified according to the aggregation probability value of each preset category of the first target; The method further includes: The name information of the first object is obtained from the local business system. The name information is obtained by extracting the text information of the local business system that displays the first object. Input the name information into a preset classification model, and output the confidence probability of each preset category and the first object among the multiple preset categories; Based on the aggregate probability value of each preset category of the first target, the first object is classified, including: The first object is classified according to the aggregate probability value of each preset category of the first target and the confidence probability of each preset category; The first object is classified according to the aggregate probability value of each preset category of the first target and the confidence probability of each preset category, including: When the maximum value of the aggregation probability value is greater than the first preset threshold, the maximum value of the aggregation probability value is determined as the category of the first object corresponding to the first target preset category. When the maximum value of the aggregated probability value is less than the first preset threshold and greater than the second preset threshold, the corresponding confidence probability is superimposed on each first target preset category to obtain the comprehensive probability value of each first target preset category. The first target preset category corresponding to the maximum value of the comprehensive probability value is determined as the category of the first object; When the maximum value of the aggregated probability is less than the second preset threshold, the preset category corresponding to the maximum confidence probability is determined as the category of the first object.

2. The method according to claim 1, characterized in that, The method further includes: Obtain the name information of the first object from the local business system; Using a preset classification model, at least one second target preset category corresponding to the first object is determined from the plurality of preset categories based on the name information; Based on the number of category information corresponding to each preset category of the first target, the first object is classified, including: The first object is classified according to the number of category information corresponding to each first target preset category and the at least one second target preset category.

3. The method according to claim 2, characterized in that, Using a preset classification model based on the name information, determining the second target preset category corresponding to the first object from among the plurality of preset categories includes: By concatenating the name information and the category information set by each business system for the target entity, the concatenated information is obtained; The spliced ​​information is input into the preset classification model, and the preset classification model is used to extract category features from the spliced ​​information; The second target preset category corresponding to the first object is determined based on the category characteristics using the preset classification model.

4. The method according to claim 1, characterized in that, Obtain the category information set by each business system for the target entity in multiple business systems that match the functional attributes of the target entity to which the first object belongs, including: Query the target second object whose entity is consistent with the target entity among multiple second objects stored in a specific business system; wherein, the specific business system is any one of the multiple business systems; Extract the category text displayed by the target second object in the specific business system as the category information set by the specific business system for the target entity.

5. An object classification device, characterized in that, The device, located in an electronic device, includes: For offline entities, data and functions are used to express the characteristics and functions of the entities, so as to create objects that represent the corresponding entities online; The category information acquisition module is used to acquire the category information set by each business system for the target entity in multiple business systems that match the functional attributes of the target entity to which the first object belongs. The category query module is used to query at least one first target preset category corresponding to each of the multiple category information in the local business system storing the first object. The object classification module is used to classify the first object according to the number of category information corresponding to each preset category of the first target; In the local business system storing the first object, querying at least one first target preset category corresponding to each of the multiple category information preset categories includes: Map any category information from the plurality of category information to the node containing the associated preset category that has the highest similarity to that category information; When the node containing the associated preset category is a leaf node in a pre-constructed category tree, the associated preset category is taken as the first target preset category; wherein, each node in the pre-constructed category tree corresponds one-to-one with the plurality of preset categories; When the associated preset category is not a leaf node in the pre-constructed category tree, the preset category located in the child node of the node where the associated preset category is located will be used as the first target preset category. Based on the number of category information corresponding to each preset category of the first target, the first object is classified, including: When there are multiple target category information corresponding to any first target preset category, the matching probability of any first target preset category matching the first object is set in each target category information dimension for the number of first target preset categories corresponding to each target category information in the business system dimension. By summing the matching probabilities of each preset category for the first target under different target category information dimensions, the aggregated probability value of each preset category for the first target is obtained. The first object is classified according to the aggregation probability value of each preset category of the first target; Also includes: The name information of the first object is obtained from the local business system. The name information is obtained by extracting the text information of the local business system that displays the first object. Input the name information into a preset classification model, and output the confidence probability of each preset category and the first object among the multiple preset categories; Based on the aggregate probability value of each preset category of the first target, the first object is classified, including: The first object is classified according to the aggregate probability value of each preset category of the first target and the confidence probability of each preset category; The first object is classified according to the aggregate probability value of each preset category of the first target and the confidence probability of each preset category, including: When the maximum value of the aggregation probability value is greater than the first preset threshold, the maximum value of the aggregation probability value is determined as the category of the first object corresponding to the first target preset category. When the maximum value of the aggregated probability value is less than the first preset threshold and greater than the second preset threshold, the corresponding confidence probability is superimposed on each first target preset category to obtain the comprehensive probability value of each first target preset category. The first target preset category corresponding to the maximum value of the comprehensive probability value is determined as the category of the first object; When the maximum value of the aggregated probability is less than the second preset threshold, the preset category corresponding to the maximum confidence probability is determined as the category of the first object.

6. A terminal device, comprising: At least one processor; And at least one memory communicatively connected to the processor, characterized in that the memory stores program instructions executable by the processor, which, when invoked, can perform the method as described in any one of claims 1 to 4.

7. A computer-readable storage medium storing computer instructions, characterized in that, The computer instructions cause the computer to perform the method as described in any one of claims 1 to 4.

Images