Data processing method and device, electronic equipment and storage medium

Abstract

Embodiments of the present application provide a data processing method and device, electronic equipment and storage medium, the method comprises: obtaining commodity information, and determining the corresponding commodity knowledge graph according to the commodity information, the commodity knowledge graph comprises nodes and the connecting edges between nodes, the nodes comprise at least one of commodity unit nodes, commodity category nodes, commodity name nodes and commodity attribute nodes; according to the commodity knowledge graph, the node features of the commodity unit nodes of the commodity are determined, and the target commodity name node corresponding to the node features of the commodity unit nodes is determined to determine the corresponding commodity name; the present method can improve the accuracy of the commodity name.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a data processing method, a data processing device, an electronic device, and a storage medium. Background Technology

[0002] Product name refers to the name of a commodity. In some scenarios, it is necessary to label commodities with their names for identification purposes. For example, in customs import and export scenarios, it is necessary to declare a product name that matches the commodity for inspection.

[0003] Currently, the name declared by the merchant is usually used as the product name for import and export. However, the product name provided by some merchants may not match the actual product, resulting in inaccurate product names. Summary of the Invention

[0004] This application provides a data processing method to improve the accuracy of product names.

[0005] Accordingly, embodiments of this application also provide a data processing device, an electronic device, and a storage medium to ensure the implementation and application of the above system.

[0006] To address the aforementioned issues, this application discloses a data processing method. The method includes: acquiring product information and determining a corresponding product knowledge graph based on the product information. The product knowledge graph includes nodes and connecting edges between nodes, and the nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. Based on the product knowledge graph, the method determines the node features of the product unit nodes and determines the target product name node corresponding to the node features of the product unit nodes, thereby determining the corresponding product name.

[0007] To address the aforementioned issues, this application discloses a data processing method. The method includes: establishing a knowledge graph based on labeled product information, the knowledge graph comprising nodes and connecting edges between nodes, the nodes including at least one of product name nodes, product unit nodes, product category nodes, product name nodes, and product attribute nodes; determining the corresponding target product name based on the knowledge graph and an analysis model, the analysis model being used to determine the node features of product unit nodes and the target product name node corresponding to the node features of the product unit nodes, thereby determining the corresponding target product name; and adjusting the analysis model based on the labeled product name and the target product name corresponding to the product information.

[0008] To address the aforementioned issues, this application discloses a data processing method, comprising: providing an interactive page to obtain product information; determining a corresponding product knowledge graph based on the product information, wherein the product knowledge graph includes nodes and connecting edges between nodes, and the nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes; determining the node features of the product unit nodes of the product based on the product knowledge graph, and determining the target product name node corresponding to the node features of the product unit nodes to determine the corresponding product name; and providing the product name.

[0009] To address the aforementioned issues, this application discloses a data processing apparatus, comprising: a knowledge graph acquisition module for acquiring product information and determining a corresponding product knowledge graph based on the product information, wherein the product knowledge graph includes nodes and connecting edges between nodes, and the nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes; and a product name acquisition module for determining the node features of the product unit nodes based on the product knowledge graph, and determining the target product name node corresponding to the node features of the product unit nodes, thereby determining the corresponding product name.

[0010] To address the aforementioned issues, this application discloses an electronic device, including: a processor; and a memory storing executable code thereon, wherein when the executable code is executed, the processor performs one or more of the methods described in the above embodiments.

[0011] To address the aforementioned issues, embodiments of this application disclose one or more machine-readable media storing executable code thereon, which, when executed, causes a processor to perform one or more of the methods described in the above embodiments.

[0012] Compared with the prior art, the embodiments of this application have the following advantages:

[0013] In this embodiment, product information of the product to be identified can be obtained, and a corresponding product knowledge graph can be constructed. The product knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. After determining the product knowledge graph, features such as product name, product category, and product attribute can be incorporated into the product unit nodes based on the product knowledge graph to determine the node features of the product unit nodes. Based on the node features of the product unit nodes, the corresponding target product name node can be determined to determine the corresponding product name. Compared with using the name declared by the merchant as the product name, this embodiment can analyze various information of the product to more accurately determine the product name corresponding to the product to be identified. In addition, this embodiment can incorporate features such as product name, product category, and product attribute into the product unit nodes, thereby increasing the amount of product features contained in the product unit nodes, and thus enabling a more accurate determination of the product name corresponding to the product. Attached Figure Description

[0014] Figure 1A This is a schematic flowchart of a data processing method according to an embodiment of this application;

[0015] Figure 1B This is a schematic flowchart of a data processing method according to another embodiment of this application;

[0016] Figure 2A This is a schematic flowchart of a data processing method according to another embodiment of this application;

[0017] Figure 2B This is a schematic flowchart of a data processing method according to another embodiment of this application;

[0018] Figure 3 This is a schematic flowchart of a data processing method according to another embodiment of this application;

[0019] Figure 4 This is a schematic flowchart of a data processing method according to another embodiment of this application;

[0020] Figure 5 This is a schematic flowchart of a data processing method according to another embodiment of this application;

[0021] Figure 6A This is a schematic flowchart of a data processing method according to another embodiment of this application;

[0022] Figure 6B This is a schematic flowchart of a data processing method according to another embodiment of this application;

[0023] Figure 7 This is a schematic diagram of the structure of a data processing apparatus according to an embodiment of this application;

[0024] Figure 8 This is a schematic diagram of the structure of a data processing apparatus according to another embodiment of this application;

[0025] Figure 9 This is a schematic diagram of the structure of a data processing apparatus according to another embodiment of this application;

[0026] Figure 10 This is a schematic diagram of the structure of an exemplary device provided in one embodiment of this application. Detailed Implementation

[0027] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0028] The embodiments of this application can be applied to the field of product name recognition, where product name refers to the name of a commodity, and product name recognition refers to recognizing the name of a commodity.

[0029] The embodiments of this application include a preparation stage and an identification stage, such as... Figure 1A As shown, in the preparation phase, a knowledge graph can be constructed based on the product information with labeled product names. This knowledge graph includes nodes and connecting edges between nodes, and each node includes at least one of the following: product name node, product unit node, product category node, product name node, and product attribute node. After constructing the knowledge graph, an analysis model is trained based on it. This model is used to determine the node features of the product unit nodes of the product to be identified and to determine the corresponding target product unit nodes, thereby determining the product name of the product to be identified. In the preparation phase, the analysis model can be adjusted based on the labeled product names and target product names corresponding to the product information. This allows the trained analysis model to determine the product name corresponding to the product information of the product to be identified during the identification phase.

[0030] In the identification phase, such as Figure 1B As shown, based on the product information of the product to be identified, the corresponding product name, product unit, product category, and product attributes can be determined, thereby constructing a product knowledge graph. The product knowledge graph includes nodes and connecting edges between nodes, and the nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. Then, using the product knowledge graph and a trained analysis model, the node features of the product unit nodes are determined, thereby identifying the product name.

[0031] Specifically, in the preparation phase, a knowledge graph can be constructed based on product information with labeled product names. This product information can include at least one of the following: product title, product attribute description, and product category description. Based on this product information, product name nodes, product unit nodes, product category nodes, product name nodes, and product attribute nodes can be constructed in the knowledge graph. The product name node includes the product name information. The product unit node includes product unit information, which corresponds to the product and can be understood as a Standard Product Unit (SPU), the smallest unit for aggregating product information. The product category node includes the category information to which the product belongs. The product name node includes the product name, and the product attribute node includes the product attribute information. After establishing nodes in the knowledge graph, connecting edges can be added between nodes based on the various product information. Specifically, a first connection edge can be established between the product name node and the product unit node based on the product title; relevant information (such as keywords, images, etc.) of two products can be displayed to consumers on the e-commerce platform, and the corresponding click-through rate can be collected to determine whether there is a relationship between the two products, and then a second connection edge can be established between different product unit nodes; a third connection edge can be established between the first product name node and the second product name node when the product title of the first product contains the product name of the second product; a fourth connection edge can be established between the product unit node and the product attribute node based on the product attribute description information; a fifth connection edge can be established between the product unit node and the product category node based on the product category description information; and a sixth connection edge can be established between the product name node and the product unit node based on the product name labeled for the product.

[0032] After establishing the knowledge graph, it can be input into the analysis model. The model can determine the node features of each node based on the connections between nodes in the knowledge graph. These node features include the features of the node itself, a first node, and a second node. The first node is the node connected to the first node, and the second node is the node connected to the first node. Specifically, the analysis model can determine the first and second nodes of each node based on the knowledge graph, and then incorporate the feature information of the first and second nodes into the node to obtain the node features. Afterward, the analysis model can determine the target product unit node corresponding to the product unit node to be identified based on the similarity between the node features of the product unit nodes, thus determining the corresponding product name. In the preparation stage, the products to be identified have been labeled, and connections have been established between product nodes and product unit nodes in the knowledge graph based on these labels. Therefore, the labeled product name corresponding to the product to be identified can be determined based on the pre-constructed knowledge graph, and the analysis model can be adjusted based on the difference between the labeled product name and the target product name to determine the trained analysis model.

[0033] In the identification phase, the product information of the product to be identified includes at least one of the following: product title, product attribute description, and product category description. Based on the product information, the corresponding product name, product unit, product category, and product attributes can be determined, and corresponding nodes can be established in the product knowledge graph. Then, based on the various product information, connecting edges are added between the nodes. Specifically, a first, second, third, fourth, and fifth connecting edge are added between the nodes. The process of adding connecting edges between nodes is similar to the process of adding connecting edges between nodes in the preparation phase described above; please refer to the above process for details, which will not be repeated here.

[0034] After determining the product knowledge graph, it can be input into a trained analysis model. The trained model can identify the first and second nodes related to each product unit node, and incorporate the feature information of these nodes into the product unit node to obtain its node features. Based on these features, a target product unit node is determined. The similarity between the target product unit node's node features and the product unit node's node features meets a preset condition (e.g., the highest similarity). Then, the corresponding product name for the target product unit node is determined for appropriate processing. For example, in a customs import / export scenario, the product name can be declared for customs clearance. It should be noted that the above embodiment describes the analysis of the product knowledge graph (and knowledge graph) using an analysis model. This application embodiment can also use other methods to analyze and process the product knowledge graph, which can be set according to requirements. For example, this application embodiment can also pre-set corresponding code segments to perform corresponding feature extraction and node feature similarity analysis actions to determine the product name. For example, the analysis model can build a knowledge graph based on product information and analyze the knowledge graph to determine the corresponding product name.

[0035] In this embodiment, product information of the product to be identified can be obtained, and a corresponding product knowledge graph can be constructed. The product knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. After determining the product knowledge graph, features such as product name, product category, and product attribute can be incorporated into the product unit nodes based on the product knowledge graph to determine the node features of the product unit nodes. Based on the node features of the product unit nodes, the corresponding target product name node can be determined to determine the corresponding product name. Compared with using the name declared by the merchant as the product name, this embodiment can analyze various information of the product to more accurately determine the product name corresponding to the product to be identified. In addition, this embodiment can incorporate features such as product name, product category, and product attribute into the product unit nodes, thereby increasing the amount of product features contained in the product unit nodes, and thus enabling a more accurate determination of the product name corresponding to the product.

[0036] This application's embodiments can be applied to scenarios involving the identification of product information for various types of goods. For example, it can be applied to identifying product information for customs goods, e-commerce goods, logistics goods, and secondhand goods. For instance, in the scenario of identifying product information for customs goods, this application's embodiments can obtain the product information of the customs product to be identified and construct a corresponding product knowledge graph. Then, based on the product knowledge graph, features such as product name, product category, and product attributes are integrated into the product unit nodes to obtain the node features of the product unit nodes. Based on the node features of the product unit nodes, the corresponding product name is determined and declared as the customs product name.

[0037] For another example, this application embodiment can also be applied to scenarios involving the identification of product information for e-commerce goods (or second-hand goods). It can acquire the product information of the e-commerce goods to be identified and construct a corresponding product knowledge graph. Then, based on the product knowledge graph, features such as product name, product category, and product attributes are integrated into the product unit nodes to obtain the node features of the product unit nodes. Based on the node features of the product unit nodes, the corresponding product name is determined. After determining the product name of the e-commerce goods (second-hand goods), the product name can be displayed to the merchant (or seller) so that the merchant can edit the product's listing information.

[0038] For another example, this application embodiment can also be applied to scenarios involving the identification of product information for logistics goods. It can acquire the product information of the logistics goods to be identified and construct a corresponding product knowledge graph. Then, based on the product knowledge graph, features such as product name, product category, and product attributes are integrated into the product unit nodes to obtain the node features of the product unit nodes. Based on the node features of the product unit nodes, the corresponding product name is determined. After determining the product name of the logistics goods, the logistics goods can be classified accordingly. For example, some food items require refrigerated transport. This application embodiment can determine the product name of the logistics goods and thus determine whether the logistics goods require refrigerated transport.

[0039] Based on the above embodiments, this application provides a data processing method that can be applied to a server. The method in this embodiment corresponds to the preparation stage, which can construct a knowledge graph based on labeled product information and train an analysis model based on the knowledge graph to determine the product name corresponding to the product. Specifically, as shown... Figure 2A As shown, the method includes:

[0040] Step 202: Based on the labeled product information, establish a knowledge graph. The knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of the following: product name node, product unit node, product category node, product name node, and product attribute node. The product information may include at least one of the following: product title, product attribute description information, and product category description information. This application embodiment can obtain this information in various ways, depending on the specific requirements. For example, this application embodiment can provide an interactive page for merchants to upload product information. This application embodiment can also set up an image acquisition component to capture images of products, product packaging, etc., and obtain product information through text recognition, entity recognition, etc. This application embodiment can establish corresponding nodes in the knowledge graph based on various types of product information and add connecting edges between the nodes. Specifically, as an optional embodiment, establishing a knowledge graph based on the labeled product information includes: establishing nodes based on the product information and the labeled product name, and establishing connecting edges between the nodes to form a knowledge graph. This application's embodiments can preprocess product information and then establish corresponding nodes based on the processed information. For example, for product titles, word segmentation can be performed, and stop words, special characters, and meaningless numbers can be removed. Synonyms can also be unified. For English words, part-of-speech tagging can be performed to determine the product name node. For product attribute description information, entity recognition can be performed to extract entities, and product attribute nodes can be determined based on the extracted entities. For product category information, product category nodes can be established based on product attributes and pre-set classification rules. Product name nodes can be established based on the labeled product names. The smallest unit of a product can be determined based on the product, and product unit nodes can be established for that product.

[0041] After establishing nodes, connecting edges can be added between nodes based on various product information. Specifically, as an optional embodiment, establishing connecting edges between nodes includes at least one of the following steps: establishing a first connecting edge between the product name node and the product unit node based on the product title; displaying relevant information of the first and second product units, and establishing a second connecting edge between the first and second product units based on the corresponding click-through rate; establishing a third connecting edge between the first and second product name nodes when the product title of the first product contains the product name of the second product; establishing a fourth connecting edge between the product unit node and the product attribute node based on product attribute description information; establishing a fifth connecting edge between the product unit node and the product category node based on product category description information; and establishing a sixth connecting edge between the product name node and the product unit node based on the product name labeled for the product.

[0042] For the same product's product name node, product attribute node, product name node, and product category node, connection edges can be added between these nodes and the product unit node based on the relevant information. For different product names (corresponding to different products), connection edges can be added between related product names based on whether the product title contains the product names of other products. Specifically, it can be determined whether the product title of the first product contains the product name of the second product, thereby determining whether there is a relationship between the product names of the first and second products. If there is a relationship between the product names of the first and second products, a third connection edge is established between the first and second product name nodes. For different product units (corresponding to different products), relevant information (such as keywords, image data, etc.) of different product unit nodes can be displayed as related terms to obtain the corresponding click-through rate. Based on the click-through rate, the degree of association between the displayed product unit nodes is determined, and a second connection edge is added between different product unit nodes when the degree of association between different product unit nodes meets the requirements. Click-through rate (CTR) is the percentage of clicks that reach online information (such as images, text ads, keywords, rankings, videos, etc.). It is the actual number of clicks on the information divided by the number of times the information is displayed.

[0043] After determining the knowledge graph, in step 204, based on the product knowledge graph, the node features of the product unit nodes are determined, and the target product name node corresponding to the node features of the product unit nodes is determined, and the corresponding product name is determined. This application embodiment can pre-train an analysis model to analyze the product knowledge graph, determine the connection relationships between nodes, and determine nodes related to the nodes according to the connection relationships. The feature information of the related nodes is then integrated into the feature information of the nodes to form node features. Specifically, as an optional embodiment, determining the node features of the product unit nodes includes: determining the first node and the second node of each node, where the first node is a node connected to the current node, and the second node is a node connected to the first node; determining the node features of each node based on the feature information of the nodes, the feature information of the first node, and the feature information of the second node; and extracting the node features of the product unit nodes.

[0044] This embodiment can filter out first nodes and second nodes connected to a node based on a knowledge graph. A first preset number of first nodes and a second preset number of second nodes can be pre-set, and the first preset number of first nodes and the second preset number of second nodes are obtained according to the preset numbers. After determining the first and second nodes for each node, the feature information of the first and second nodes can be obtained and integrated into the node's feature information to form the node's node features. This ensures that the node features contain information such as product attributes, product categories, product names, product units, and product names, increasing the amount of information contained in the node features and thus improving the accuracy of product name recognition.

[0045] To further improve the accuracy of product name recognition, this embodiment can also determine the weight of the second node and incorporate the features of the second node into the node according to the corresponding weight. Specifically, as an optional embodiment, determining the node features of each node based on the feature information of the node, the feature information of the first node, and the feature information of the second node includes: incorporating the corresponding feature information of the first node into each target node to obtain the feature vector of each target node; determining the feature incorporation weight based on the feature similarity between the feature vector of the target node and the feature vector of the second node of the target node; and incorporating the feature vector of the second node into the feature vector of the target node according to the feature incorporation weight to form the node features of the target node.

[0046] This application embodiment can acquire and fuse the feature information of each target node and the feature information of the first node of the target node to determine the feature vector of each target node. Then, it determines the feature similarity between the feature vectors of the target node and the feature vector of the second node, and determines the feature integration weight according to the similarity between the feature vectors. The higher the feature similarity between the feature vectors of the target node and the feature vectors of the second node, the closer the connection between the target node and the second node, and the higher the corresponding feature integration weight; conversely, the lower the feature integration weight, the lower the feature integration weight. After determining the feature integration weight, the integration amount of the feature vector of the second node can be determined according to the magnitude of the feature integration weight, and then the feature vector of the second node is integrated into the target node based on this integration amount, forming the node feature of the target node. This application embodiment can determine the degree of association between the target node and the second node based on the similarity between the target node and the second node, and then integrate a corresponding amount of the second node's feature into the node feature of the target node, making the node feature of the target node more accurate, thereby improving the accuracy of product name recognition. After determining the node features of each node in the knowledge graph, the node features of the product unit node can be extracted, and based on the node features of the product unit node, the corresponding target product name node can be determined to identify the corresponding target product name.

[0047] This application embodiment can match the node features of the product unit node of the product to be identified with the node features of existing product unit nodes to determine the similarity of the features. Based on the similarity, a target product unit node matching the product to be identified is determined, thus identifying the corresponding product name. The similarity can be understood as the degree of interaction between the node features of the product unit nodes. The more overlap between the nodes related to the product unit node to be identified (first node and second node) and the nodes related to existing product unit nodes (first node and second node), the higher the degree of interaction, and the higher the similarity; conversely, the lower the degree of interaction, the lower the similarity. In one optional example, one or more product unit nodes with the highest similarity can be selected as target product unit nodes. In another optional example, a similarity threshold can be preset. If the similarity between the node features of the target product unit node and the node features of other product unit nodes meets the similarity threshold, the target product unit node is determined to match the product to be identified.

[0048] After determining the target product name, in step 206, the analysis model can be adjusted based on the labeled product name and the target product name corresponding to the product information. During the preparation phase, the product to be identified has been labeled, and connections have been established between product nodes and product units in the knowledge graph based on these labels. Therefore, the labeled product name corresponding to the product to be identified can be determined based on the pre-constructed knowledge graph, and the analysis model can be adjusted based on the difference between the labeled product name and the target product name. In an optional embodiment, a loss function can be determined based on the difference between the labeled product name and the target product name, and the analysis model can be adjusted based on the loss function. The loss function characterizes the predictive quality of the analysis model; the greater the difference between the labeled product name and the target product name, the larger the loss function, and vice versa. When predicting the target product name, the analysis model generates a corresponding target confidence score, which characterizes the credibility of the target product name. If the labeled product name matches the target product name, the corresponding label confidence level is 1; if the labeled product name does not match the target product name, the corresponding label confidence level is 0. In an optional example, cross-entropy can be used as the loss function. Cross-entropy is primarily used to measure the difference between two probability distributions.

[0049] In this embodiment, labeled product information can be obtained and a knowledge graph can be constructed. The knowledge graph includes nodes and connecting edges between nodes. Each node includes at least one of the following: product name node, product unit node, product category node, product name node, and product attribute node. After constructing the knowledge graph, an analysis model is trained based on it. This analysis model is used to determine the node features of the product unit nodes of the product to be identified and to determine the corresponding target product unit node, thereby determining the target product name. The analysis model is then adjusted based on the difference between the target product name and the labeled product name. The trained analysis model can determine the corresponding product name based on the product information to be analyzed.

[0050] The following describes the data processing method of this application embodiment using a specific example. Specifically, as shown below... Figure 2B As shown in the embodiments of this application, product information can be obtained and sampled for labeling. Product names are then labeled to obtain labeled product information. To train, validate, and test the analysis model, the labeled product information can be divided according to a preset ratio (e.g., 8:1:1) to obtain a training set, a validation set, and a test set. The training set is used to train the analysis model, the validation set is used to verify whether the trained analysis model meets the validation conditions, and the test set is used to test whether the accuracy of the analysis model meets preset requirements.

[0051] In the process of training and analyzing the model based on the training set, a knowledge graph can be constructed based on the product information and labeled product names in the training set. The nodes of the knowledge graph include product unit nodes (SPU), product name nodes mapped to SPU, product category nodes, product attribute nodes, and product name nodes (which can be denoted as Token).

[0052] In the process of constructing connection edges between nodes, SPUs can obtain corresponding click-through rates by displaying search terms corresponding to two product unit nodes, and determine whether there is a relationship between product unit nodes based on the click-through rates. For example, if the click-through rates of two product unit nodes are significantly different, it is determined that there is no relationship between the two product unit nodes; if the click-through rates of two product unit nodes are relatively small, it is determined that there is a relationship between the two product unit nodes, and a connection edge is added, which can be denoted as TO. The relationship between SPUs and Tokens can be determined based on the title, which can be denoted as TITLE. The relationship between Tokens can be constructed based on the adjacency relationship in the title, which can be denoted as PMI. The relationship between the Token of the first product and the Token of the second product can be determined based on whether there is relevant information about the second product unit in the title of the first product. The PMI between Tokens can be determined according to the following formula 1.

[0053]

[0054] Where #W represents the length of the SPU title character, #W(i) represents the number of times Token_i appears within a fixed-length sliding window (the length can be preset to form a fixed-length sliding window), and #w(i,j) represents the number of times Token_i and Token_j co-occur within the sliding window. A PMI relationship exists between tokens only when the PMI value is > 0. The relationship between SPU and product category nodes is named CATE; the relationship between SPU and product attribute nodes can be named ATTR; and the relationship between SPU and product name can be named PN.

[0055] After determining the knowledge graph, vectors 1 and 2 can be constructed using graph network learning and knowledge graph learning methods. Vectors 1 and 2 are then merged to obtain the node features of each node. Specifically, the first and second nodes of each node can be determined based on the knowledge graph, and the features of the first node are incorporated into the features of each node to obtain the feature vectors of each node. These feature vectors can be denoted as vector 1 (emb1). After determining the feature vectors of each node, the feature correlation between the second node and the target node can be determined, the corresponding feature integration weights can be determined, and the features to be integrated into the second node can be determined so that these features can be incorporated into the feature vector of the target node. The features to be integrated into the second node can be denoted as vector 2 (emb2), where vector 2 can be determined using the following formula 2.

[0056]

[0057] Here, the (h, r, t) triple represents the feature vector of the node that incorporates the first node t. After determining vector 1 and vector 2, vector 1 (emb1) and vector 2 (emb2) are concatenated to obtain the node features of each node. After determining the node features of each node, the node features of multiple product unit nodes can be extracted (e.g., Figure 2B The similarity between the node features of SPU1 and SPU2 is calculated, and the predicted value of the similarity between the node features of the product unit nodes is determined. Then, the corresponding loss is determined based on the predicted value and the true value to adjust the model, and the trained model is saved. The true value of the similarity between the node features of the product unit nodes is determined based on whether there are connecting edges between the product unit nodes in the knowledge graph. Specifically, the node features of two product unit nodes can be denoted as u and v, respectively, and can be determined using the following formula 3.

[0058]

[0059] The loss function can be cross-entropy, and the loss function can be determined by the following formula 4.

[0060]

[0061] After determining the loss function, the analysis model can be adjusted based on the loss function to obtain the trained analysis model. This trained model is then validated using a validation set and tested using a test set, resulting in a fully trained analysis model, which is then saved. After the analysis model is trained, it can analyze the products to be analyzed during the identification phase to determine the corresponding product names. Specifically, in this embodiment, the node features of each Product Unit (SPU) during the model training phase can be saved as a label set SPU. During the analysis of the products to be identified, the SPU to be analyzed can be input into the analysis model to determine its node features. Then, the similarity between the node features of the SPU to be analyzed and the node features of the label set SPU is determined, and the product name of the label set SPU with the highest similarity is selected as the product name of the SPU to be analyzed.

[0062] Based on the above embodiments, this application also provides a data processing method that can be applied to a server, such as... Figure 3 As shown, the method includes:

[0063] Step 302: Based on the product information and the product name labeled for the product, establish nodes and create connecting edges between the nodes to form a knowledge graph. The knowledge graph includes nodes and connecting edges between nodes, and the nodes include at least one of product name nodes, product unit nodes, product category nodes, product name nodes, and product attribute nodes. As an optional embodiment, creating connecting edges between nodes includes at least one of the following steps: creating a first connecting edge between the product name node and the product unit node based on the product title; displaying relevant information of the first product unit and the second product unit, and creating a second connecting edge between the first product unit node and the second product unit based on the corresponding click-through rate; creating a third connecting edge between the first product name node and the second product name node when the product title of the first product contains the product name of the second product; creating a fourth connecting edge between the product unit node and the product attribute node based on the product attribute description information; creating a fifth connecting edge between the product unit node and the product category node based on the product category description information; and creating a sixth connecting edge between the product name node and the product unit node based on the product name labeled for the product.

[0064] Step 304: Based on the knowledge graph, determine the first node and the second node of each node, where the first node is the node connected to the node and the second node is the node connected to the first node.

[0065] Step 306: Integrate the feature information of the corresponding first node into each target node to obtain the feature vector of each target node.

[0066] Step 308: Determine the feature integration weight based on the feature similarity between the feature vector of the target node and the feature vector of the second node of the target node.

[0067] Step 310: Based on the weighted features, incorporate the feature vector of the second node into the feature vector of the target node to form the node features of the target node.

[0068] Step 312: Extract the node features of the product unit node.

[0069] Step 314: Determine the similarity between the node features of the product unit nodes.

[0070] Step 316: Based on the similarity, determine the target product unit node corresponding to the product unit node.

[0071] Step 318: Determine the target product name node connected to the target product unit node to determine the corresponding product name.

[0072] Step 320: Adjust the analysis model based on the labeled product name and target product name corresponding to the product information.

[0073] In this embodiment, nodes can be established based on the product information and the product name labeled for the product, and connecting edges can be established between the nodes to form a knowledge graph. The knowledge graph includes nodes and connecting edges between nodes, and the nodes include at least one of product name nodes, product unit nodes, product category nodes, product name nodes, and product attribute nodes. After constructing the knowledge graph, the first node and second node of each node can be determined based on the knowledge graph, and the feature information of the first node can be integrated into each node to obtain the feature vector of each node. Then, the feature integration weight corresponding to the second node of each target node is determined, and the feature vector of the second node of the corresponding amount is integrated into the target node according to the corresponding weight to obtain the node feature of the target node. Then, the node feature of the product unit node is extracted, and the corresponding target product unit node is determined, thereby obtaining the target product name node to determine the target product name. Then, the analysis model can be adjusted based on the labeled product name and target product name corresponding to the product information. The trained analysis model can determine the corresponding product name based on the product information to be analyzed.

[0074] Based on the above embodiments, this application also provides a data processing method that can be applied to a server. In the identification stage, this method can construct a knowledge graph based on the product information of the product to be identified, and determine the node features corresponding to the product unit nodes based on the knowledge graph, thereby determining the corresponding product name. Specifically, as shown... Figure 4 As shown, the method includes:

[0075] Step 402: Obtain product information and determine the corresponding product knowledge graph based on the product information. The product knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. This application embodiment can create nodes in the product knowledge graph based on various product information and add connecting edges between the nodes. Specifically, as an optional embodiment, determining the corresponding product knowledge graph based on product information includes: creating nodes based on product information and establishing connecting edges between nodes to form a product knowledge graph. This application embodiment can establish connection edges between nodes based on various information of the product. Specifically, as an optional embodiment, establishing connection edges between nodes includes at least one of the following steps: establishing a first connection edge between the product name node and the product unit node based on the product title; displaying relevant information of the first product unit and the second product unit, and establishing a second connection edge between the first product unit node and the second product unit based on the corresponding click-through rate; establishing a third connection edge between the first product name node and the second product name node when the product title of the first product contains the product name of the second product; establishing a fourth connection edge between the product unit node and the product attribute node based on the product attribute description information; and establishing a fifth connection edge between the product unit node and the product category node based on the product category description information.

[0076] Step 404: Based on the product knowledge graph, determine the node features of the product unit nodes and the target product name node corresponding to the node features of the product unit nodes to determine the corresponding product name. This embodiment of the application can analyze the connection relationships between nodes in the product knowledge graph, thereby determining the first and second nodes related to the product unit nodes, and then incorporating the features of the first and second nodes into the product unit nodes to determine the node features of the product unit nodes. Specifically, as an optional embodiment, determining the node features of the product unit nodes based on the product knowledge graph includes: determining the first and second nodes related to the product unit nodes, where the first node is a node connected to the product unit node, and the second node is a node connected to the first node; and determining the node features of the product unit nodes based on the feature information of the first node and the feature information of the second node.

[0077] After determining the node features of a product unit node, target product unit nodes similar to the product unit node can be identified, thereby determining the corresponding product name node and obtaining the corresponding product name. Specifically, in one optional embodiment, the similarity between the node features of the product unit node to be analyzed and the node features of existing product unit nodes (product unit nodes saved during the training phase) can be determined, and the similarity can be sorted to select one or more of the most similar existing product unit nodes as target product unit nodes, and the corresponding product name can be determined. In another optional example, the similarity between the node features of the product unit node and the node features of the target product unit node can be determined, and this similarity can be compared with a pre-set similarity threshold to identify target product unit nodes similar to the product unit node, thereby determining the corresponding product name.

[0078] In an optional embodiment, this application embodiment may provide an interactive page to obtain product information of the product to be analyzed. The server then analyzes the product information to obtain the product name and sends the product name back through the interactive page. This allows for appropriate processing based on the analyzed product name.

[0079] This application embodiment can train an analysis model using labeled product information (product information labeled with product names), and use the trained analysis model to process the product information to determine the corresponding product name. Specifically, as an optional embodiment, the product name is determined after analyzing the product information based on the trained analysis model. The method further includes: determining a knowledge graph based on the labeled product information, the knowledge graph including nodes and connecting edges between nodes, the nodes including at least one of product name nodes, product unit nodes, product category nodes, product name nodes, and product attribute nodes; training an analysis model based on the knowledge graph, the analysis model being used to determine the node features of product unit nodes, and determining the target product unit node corresponding to the node features of the product unit nodes, thereby determining the corresponding target product name. This application embodiment can obtain labeled product information, construct a corresponding knowledge graph, and train an analysis model based on the knowledge graph. The knowledge graph can be input into the analysis model to determine the product name prediction result. Then, the difference between the product name prediction result and the product name labeling result is determined to adjust the analysis model. After training the analysis model, it can be used to process the product information of the goods to be identified. Specifically, the network architecture and model parameters of the trained analysis model can be configured on different devices for corresponding processing. For example, the trained analysis model can be configured on terminals, servers, and other devices for appropriate processing.

[0080] This application embodiment can identify product information and provide the identified product name to the merchant so that the merchant can submit the corresponding product name. However, the actual identified product name may be inaccurate. Therefore, in this application embodiment, the submitted product name and the identified product name can be collected and statistically analyzed. When the submitted product name and the actual product name of multiple products are inconsistent, a prompt can be issued for appropriate processing. Specifically, as an optional embodiment, the method further includes: obtaining the submitted product name; obtaining the product name determined based on product information; based on the product name, counting the number of discrepancies in the submitted product names, and outputting a prompt message when the number of discrepancies meets a discrepancy threshold. When multiple submitted product names are inconsistent with the product name identified by the analysis model, a prompt message can be output for appropriate processing. For example, the prompt message can be output to the server, which can then correct the analysis model based on the prompt message to make the analysis model more suitable for the merchant. For example, prompts can be sent to staff so that they can perform manual verification. If the merchant's declaration is incorrect, the staff can provide guidance to the merchant. If the merchant's declaration is correct but the analysis model is misidentified, the staff can retrain an analysis model for the merchant to obtain a more suitable analysis model for the merchant's products.

[0081] The implementation methods of this application are similar to those of the above method embodiments. For specific implementation methods, please refer to the implementation methods of the above method embodiments, which will not be repeated here.

[0082] In this embodiment, product information of the product to be identified can be obtained, and a corresponding product knowledge graph can be constructed. The product knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. After determining the product knowledge graph, features such as product name, product category, and product attribute can be incorporated into the product unit nodes based on the product knowledge graph to determine the node features of the product unit nodes. Based on the node features of the product unit nodes, the corresponding target product name node can be determined to determine the corresponding product name. Compared with using the name declared by the merchant as the product name, this embodiment can analyze various information of the product to more accurately determine the product name corresponding to the product to be identified. In addition, this embodiment can incorporate features such as product name, product category, and product attribute into the product unit nodes, thereby increasing the amount of product features contained in the product unit nodes, and thus enabling a more accurate determination of the product name corresponding to the product.

[0083] Based on the above embodiments, this application also provides a data processing method that can be applied to a server, such as... Figure 5 As shown, the method includes:

[0084] Step 502: Based on the product information, create nodes and establish connection edges between nodes to form a product knowledge graph. The product knowledge graph includes nodes and connection edges between nodes, and the nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. Specifically, as an optional embodiment, establishing connection edges between nodes includes at least one of the following steps: establishing a first connection edge between the product name node and the product unit node based on the product title; displaying relevant information of the first product unit and the second product unit, and establishing a second connection edge between the first product unit node and the second product unit based on the corresponding click-through rate; establishing a third connection edge between the first product name node and the second product name node when the product title of the first product contains the product name of the second product; establishing a fourth connection edge between the product unit node and the product attribute node based on the product attribute description information; and establishing a fifth connection edge between the product unit node and the product category node based on the product category description information.

[0085] Step 504: Based on the product knowledge graph, determine the first node and the second node related to the product unit node. The first node is a node connected to the node, and the second node is a node connected to the first node.

[0086] Step 506: Determine the node characteristics of the commodity unit node based on the feature information of the first node and the feature information of the second node.

[0087] Step 508: Based on the node characteristics of the product unit node, determine the target product unit node corresponding to the product unit node.

[0088] Step 510: Determine the target product name node connected to the target product unit node to determine the corresponding product name.

[0089] In this embodiment, product information of the product to be identified can be obtained, and a corresponding product knowledge graph can be constructed. The product knowledge graph includes nodes and connecting edges between nodes. After determining the product knowledge graph, the first node and each second node related to the product unit node can be determined based on the product knowledge graph. Then, based on the feature information of the first node and the feature information of the second node, features such as product name, product category, and product attributes are incorporated into the product unit node to determine the node features of the product unit node. Based on the node features of the product unit node, the corresponding target product unit node is determined. Then, based on the target product unit node, the corresponding target product name node is determined to determine the corresponding product name.

[0090] Based on the above embodiments, this application also provides a data processing method that can be applied to a server. This method provides an interactive page to obtain product information and, based on the product information of the product to be identified, constructs a knowledge graph. Then, based on the knowledge graph, it determines the node features corresponding to the product unit nodes, thereby determining the corresponding product name. Specifically, as shown... Figure 6A As shown, the method includes:

[0091] Step 602: Provide an interactive page to obtain product information.

[0092] Step 604: Based on the product information, determine the corresponding product knowledge graph. The product knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes.

[0093] Step 606: Based on the product knowledge graph, determine the node features of the product unit nodes and the target product name node corresponding to the node features of the product unit nodes, so as to determine the corresponding product name.

[0094] Step 608: Provide the product name.

[0095] The implementation methods of this application are similar to those of the above method embodiments. For specific implementation methods, please refer to the implementation methods of the above method embodiments, which will not be repeated here.

[0096] like Figure 6B As shown in this embodiment, the server can provide an interactive page to the terminal. This page may include an information upload control. The terminal user can trigger the upload control to upload various product information, which is then transmitted to the server. Upon receiving the product information, the server can construct a corresponding product knowledge graph. This knowledge graph includes nodes and connecting edges between them. Each node includes at least one of the following: product unit node, product category node, product name node, and product attribute node. After determining the product knowledge graph, the server can incorporate features such as product name, product category, and product attribute into the product unit nodes to determine their node characteristics. Based on these characteristics, the server can then determine the corresponding target product name node to identify the product name. Once the server determines the product name, it can send it back to the terminal for display on the interactive page. Alternatively, as an optional embodiment, a product name adjustment control can be set in the interactive page to obtain the corrected product name and upload it to the server. After obtaining the corrected product name, the server can adjust the analysis model based on the corrected product name to further improve the accuracy of the analysis model.

[0097] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of this application are not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of this application.

[0098] Based on the above embodiments, this embodiment also provides a data processing apparatus, referring to... Figure 7 Specifically, it can include the following modules:

[0099] The knowledge graph acquisition module 702 is used to acquire product information and determine the corresponding product knowledge graph based on the product information. The product knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes.

[0100] The product name acquisition module 704 is used to determine the node features of the product unit node of the product based on the product knowledge graph, and to determine the target product name node corresponding to the node features of the product unit node, so as to determine the corresponding product name.

[0101] In summary, this embodiment of the application can obtain product information of the product to be identified and construct a corresponding product knowledge graph. The product knowledge graph includes nodes and connecting edges between nodes, and the nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. After determining the product knowledge graph, features such as product name, product category, and product attribute can be incorporated into the product unit nodes based on the product knowledge graph to determine the node features of the product unit nodes. Based on the node features of the product unit nodes, the corresponding target product name node can be determined to determine the corresponding product name. Compared with using the name declared by the merchant as the product name, this embodiment of the application can analyze various information of the product to more accurately determine the product name corresponding to the product to be identified. In addition, this embodiment of the application can incorporate features such as product name, product category, and product attribute into the product unit nodes, thereby increasing the amount of product features contained in the product unit nodes, and thus enabling a more accurate determination of the product name corresponding to the product.

[0102] Based on the above embodiments, this embodiment also provides a data processing device, which may specifically include the following modules:

[0103] The knowledge graph construction and processing module is used to create nodes based on product information and establish connection edges between nodes to form a product knowledge graph. The product knowledge graph includes nodes and connection edges between nodes, and the nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. Specifically, as an optional embodiment, the knowledge graph construction and processing module is specifically used to complete at least one of the following steps: establishing a first connection edge between the product name node and the product unit node based on the product title; displaying relevant information of the first product unit and the second product unit, and establishing a second connection edge between the first product unit node and the second product unit based on the corresponding click-through rate; establishing a third connection edge between the first product name node and the second product name node when the product title of the first product contains the product name of the second product; establishing a fourth connection edge between the product unit node and the product attribute node based on product attribute description information; and establishing a fifth connection edge between the product unit node and the product category node based on product category description information.

[0104] The related node acquisition and processing module is used to determine, based on the product knowledge graph, a first node and a second node related to the product unit node, where the first node is a node connected to the product unit node, and the second node is a node connected to the first node. The node feature acquisition and processing module is used to determine the node features of the product unit node based on the feature information of the first node and the feature information of the second node. The target unit acquisition and processing module is used to determine the target product unit node corresponding to the product unit node based on the node features of the product unit node. The product name acquisition and processing module is used to determine the target product name node connected to the target product unit node, thereby determining the corresponding product name.

[0105] In this embodiment, product information of the product to be identified can be obtained, and a corresponding product knowledge graph can be constructed. The product knowledge graph includes nodes and connecting edges between nodes. After determining the product knowledge graph, the first node and each second node related to the product unit node can be determined based on the product knowledge graph. Then, based on the feature information of the first node and the feature information of the second node, features such as product name, product category, and product attributes are incorporated into the product unit node to determine the node features of the product unit node. Based on the node features of the product unit node, the corresponding target product unit node is determined. Then, based on the target product unit node, the corresponding target product name node is determined to determine the corresponding product name.

[0106] Based on the above embodiments, this embodiment also provides a data processing apparatus, referring to... Figure 8 Specifically, it can include the following modules:

[0107] The knowledge graph determination module 802 is used to establish a knowledge graph based on the labeled product information. The knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of the following: product name node, product unit node, product category node, product name node, and product attribute node.

[0108] The product name determination module 804 is used to determine the corresponding target product name based on the knowledge graph and the analysis model. The analysis model is used to determine the node characteristics of the product unit node and the target product name node corresponding to the node characteristics of the product unit node, so as to determine the corresponding target product name.

[0109] The model loss determination module 806 is used to adjust the analysis model based on the labeled product name and the target product name corresponding to the product information.

[0110] In this embodiment, labeled product information can be obtained and a knowledge graph can be constructed. The knowledge graph includes nodes and connecting edges between nodes. Each node includes at least one of the following: product name node, product unit node, product category node, product name node, and product attribute node. After constructing the knowledge graph, an analysis model is trained based on it. This analysis model determines the node features of the product unit nodes of the product to be identified and identifies the corresponding target product unit nodes, thereby determining the target product name node and thus the target product name. The analysis model is then adjusted based on the difference between the target product name and the labeled product names. The trained analysis model can determine the corresponding product name based on the product information to be analyzed.

[0111] Based on the above embodiments, this embodiment also provides a data processing device, which may specifically include the following modules:

[0112] A knowledge graph building and processing module is used to create nodes based on product information and product names labeled for the products, and to establish connecting edges between the nodes to form a knowledge graph. The knowledge graph includes nodes and connecting edges between nodes, and the nodes include at least one of product name nodes, product unit nodes, product category nodes, product name nodes, and product attribute nodes. As an optional embodiment, the knowledge graph building and processing module is specifically used to complete at least one of the following steps: establishing a first connecting edge between the product name node and the product unit node based on the product title; displaying relevant information of the first and second product units, and establishing a second connecting edge between the first and second product unit nodes based on the corresponding click-through rate; establishing a third connecting edge between the first and second product name nodes when the product title of the first product contains the product name of the second product; establishing a fourth connecting edge between the product unit node and the product attribute node based on product attribute description information; establishing a fifth connecting edge between the product unit node and the product category node based on product category description information; and establishing a sixth connecting edge between the product name node and the product unit node based on the product name labeled for the products.

[0113] The related node determination processing module is used to determine the first node and the second node of each node based on the knowledge graph, where the first node is a node connected to the target node, and the second node is a node connected to the first node. The first feature integration processing module is used to integrate the feature information of the corresponding first node into each target node to obtain the feature vector of each target node. The second feature integration processing module is used to determine the feature integration weight based on the feature similarity between the feature vector of the target node and the feature vector of the second node of the target node. The node feature determination processing module is used to integrate the feature vector of the second node into the feature vector of the target node according to the feature integration weight, forming the node feature of the target node. The node feature extraction processing module is used to extract the node features of the product unit node.

[0114] The feature similarity acquisition processing module is used to determine the similarity between the node features of the product unit nodes. The target unit determination processing module is used to determine the target product unit node corresponding to the product unit node based on the similarity. The product name determination processing module is used to determine the target product name node connected to the target product unit node, thereby determining the corresponding product name. The analysis model adjustment processing module is used to adjust the analysis model based on the labeled product name and the target product name corresponding to the product information.

[0115] In this embodiment, nodes can be established based on the product information and the product name labeled for the product, and connecting edges can be established between the nodes to form a knowledge graph. The knowledge graph includes nodes and connecting edges between nodes, and the nodes include at least one of product name nodes, product unit nodes, product category nodes, product name nodes, and product attribute nodes. After constructing the knowledge graph, the first node and second node of each node can be determined based on the knowledge graph, and the feature information of the first node can be integrated into each node to obtain the feature vector of each node. Then, the feature integration weight corresponding to the second node of each target node is determined, and the feature vector of the second node of the corresponding amount is integrated into the target node according to the corresponding weight to obtain the node feature of the target node. Then, the node feature of the product unit node is extracted, and the corresponding target product unit node is determined, thereby obtaining the target product name node to determine the target product name. Then, the analysis model can be adjusted based on the labeled product name and target product name corresponding to the product information. The trained analysis model can determine the corresponding product name based on the product information to be analyzed.

[0116] Based on the above embodiments, this embodiment also provides a data processing apparatus, referring to... Figure 9 Specifically, it can include the following modules:

[0117] The interactive page provides module 902, which is used to provide an interactive page to obtain product information.

[0118] The knowledge graph generation module 904 is used to determine the corresponding product knowledge graph based on the product information. The product knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes.

[0119] The product name generation module 906 is used to determine the node features of the product unit node of the product based on the product knowledge graph, and to determine the target product name node corresponding to the node features of the product unit node, so as to determine the corresponding product name.

[0120] The product name distribution module 908 is used to provide feedback on the product name.

[0121] In summary, in this embodiment, the server can provide an interactive page to the terminal. This page may include an information upload control. The terminal user can trigger the upload control to upload various product information, which is then transmitted to the server. Upon receiving the product information, the server can construct a corresponding product knowledge graph. This knowledge graph includes nodes and connecting edges between them. Each node includes at least one of the following: product unit nodes, product category nodes, product name nodes, and product attribute nodes. After determining the product knowledge graph, the server can incorporate features such as product name, product category, and product attribute into the product unit nodes to determine their node characteristics. Based on these characteristics, the server can then determine the corresponding target product name node to identify the product name. Once the server determines the product name, it can send it back to the terminal for display on the interactive page.

[0122] This application also provides a non-volatile readable storage medium storing one or more modules (programs). When these modules are applied to a device, they enable the device to execute the instructions for the method steps in this application.

[0123] This application provides one or more machine-readable media storing instructions that, when executed by one or more processors, cause an electronic device to perform one or more of the methods described in the above embodiments. In this application, the electronic device includes devices such as servers and terminal devices.

[0124] Embodiments of this disclosure can be implemented as an apparatus with any suitable hardware, firmware, software, or any combination thereof, configured as desired, and the apparatus may include electronic devices such as servers (clusters) and terminals. Figure 10An exemplary apparatus 1000 is schematically shown that can be used to implement the various embodiments described in this application.

[0125] In one embodiment, Figure 10 An exemplary device 1000 is shown, which includes one or more processors 1002, a control module (chipset) 1004 coupled to at least one of the processors 1002, a memory 1006 coupled to the control module 1004, a non-volatile memory (NVM) / storage device 1008 coupled to the control module 1004, one or more input / output devices 1010 coupled to the control module 1004, and a network interface 1012 coupled to the control module 1004.

[0126] Processor 1002 may include one or more single-core or multi-core processors, and processor 1002 may include any combination of general-purpose processors or special-purpose processors (e.g., graphics processors, application processors, baseband processors, etc.). In some embodiments, device 1000 can serve as a server, terminal, or other device as described in the embodiments of this application.

[0127] In some embodiments, apparatus 1000 may include one or more computer-readable media (e.g., memory 1006 or NVM / storage device 1008) having instructions 1014 and one or more processors 1002 that are combined with the one or more computer-readable media and configured to execute instructions 1014 to implement modules and thereby perform the actions described in this disclosure.

[0128] In one embodiment, the control module 1004 may include any suitable interface controller to provide any suitable interface to at least one of the processors 1002 and / or any suitable device or component communicating with the control module 1004.

[0129] The control module 1004 may include a memory controller module to provide an interface to the memory 1006. The memory controller module may be a hardware module, a software module, and / or a firmware module.

[0130] Memory 1006 may be used, for example, to load and store data and / or instructions 1014 for device 1000. In one embodiment, memory 1006 may include any suitable volatile memory, such as suitable DRAM. In some embodiments, memory 1006 may include double data rate type quad synchronous dynamic random access memory (DDR4 SDRAM).

[0131] In one embodiment, the control module 1004 may include one or more input / output controllers to provide interfaces to the NVM / storage device 1008 and (one or more) input / output devices 1010.

[0132] For example, NVM / storage device 1008 may be used to store data and / or instructions 1014. NVM / storage device 1008 may include any suitable non-volatile memory (e.g., flash memory) and / or may include any suitable (one or more) non-volatile storage devices (e.g., one or more hard disk drives (HDDs), one or more optical disc drives (CDs), and / or one or more digital universal optical disc (DVD) drives).

[0133] NVM / storage device 1008 may include storage resources that are part of a device on which device 1000 is mounted, or that can be accessed by the device without being part of the device. For example, NVM / storage device 1008 may be accessed via a network via one or more input / output devices 1010.

[0134] One or more input / output devices 1010 may provide an interface for device 1000 to communicate with any other suitable device. Input / output devices 1010 may include communication components, audio components, sensor components, etc. Network interface 1012 may provide an interface for device 1000 to communicate via one or more networks. Device 1000 may wirelessly communicate with one or more components of a wireless network according to any of one or more wireless network standards and / or protocols, such as accessing wireless networks based on communication standards, such as WiFi, 2G, 3G, 4G, 5G, etc., or combinations thereof.

[0135] In one embodiment, at least one of the processors 1002 may be logically packaged with one or more controllers (e.g., memory controller modules) of the control module 1004. In one embodiment, at least one of the processors 1002 may be logically packaged with one or more controllers of the control module 1004 to form a system-in-package (SiP). In one embodiment, at least one of the processors 1002 may be integrated with the logic of one or more controllers of the control module 1004 on the same die. In one embodiment, at least one of the processors 1002 may be integrated with the logic of one or more controllers of the control module 1004 on the same die to form a system-on-a-chip (SoC).

[0136] In various embodiments, device 1000 may be, but is not limited to, a terminal device such as a server, desktop computing device, or mobile computing device (e.g., laptop computing device, handheld computing device, tablet computer, netbook, etc.). In various embodiments, device 1000 may have more or fewer components and / or different architectures. For example, in some embodiments, device 1000 includes one or more cameras, a keyboard, a liquid crystal display (LCD) screen (including a touch screen display), a non-volatile memory port, multiple antennas, a graphics chip, an application-specific integrated circuit (ASIC), and a speaker.

[0137] The detection device can use a main control chip as a processor or control module, and sensor data, position information, etc. can be stored in a memory or NVM / storage device. The sensor group can be used as an input / output device, and the communication interface can include a network interface.

[0138] This application also provides an electronic device, including: a processor; and a memory storing executable code thereon, which, when executed, causes the processor to perform one or more methods as described in this application.

[0139] This application also provides one or more machine-readable media having executable code stored thereon, which, when executed, causes a processor to perform one or more of the methods described in this application.

[0140] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0141] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0142] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0143] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0144] These computer program instructions may also be loaded onto a computer or other programmable data processing terminal equipment to cause a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable terminal equipment, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0145] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0146] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0147] The above provides a detailed description of a data processing method, a data processing device, an electronic device, and a storage medium provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A data processing method, characterized in that, The method includes: Obtain product information for the product to be identified; Based on the product information, a product knowledge graph of the product to be identified is determined. The product knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. The knowledge graph is input into a pre-trained analysis model, which outputs the product name of the product to be identified. The analysis model is trained based on a knowledge graph constructed from labeled product information, and the nodes of the knowledge graph include product name nodes. The analysis model performs the following steps: Determine the first node and the second node related to the product unit node, where the first node is a node connected to the first node and the second node is a node connected to the first node. Based on the feature information of the first node and the feature information of the second node, the node features of the commodity unit node are determined; Based on the node characteristics of the product unit node, determine the target product unit node; If the similarity between the node features of the target product unit node and the node features of the product unit meets the preset conditions, the corresponding product name of the target product unit node is determined.

2. The method according to claim 1, characterized in that, Based on the feature information of the nodes, the feature information of the first node, and the feature information of the second node, the node features of each node are determined, including: The feature information of the corresponding first node is incorporated into each target node to obtain the feature vector of each target node; The feature integration weights are determined based on the feature similarity between the feature vector of the target node and the feature vector of the second node of the target node. Based on the aforementioned feature integration weights, the feature vector of the second node is integrated into the feature vector of the target node to form the node features of the target node.

3. The method according to claim 1, characterized in that, The product name is determined after analyzing product information using a trained analysis model, and the method further includes: Based on the labeled product information, a knowledge graph is determined. The knowledge graph includes nodes and connecting edges between nodes. The nodes also include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. Based on the knowledge graph, an analysis model is trained. The analysis model is used to determine the node characteristics of the product unit node and the target product unit node corresponding to the node characteristics of the product unit node, so as to determine the corresponding target product name.

4. A data processing method, characterized in that, The method includes: Based on the labeled product information, a knowledge graph is established. The knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of the following: product name node, product unit node, product category node, product name node, and product attribute node. The knowledge graph is input into a pre-trained analysis model, which outputs the product name of the target product. The analysis model performs the following steps: determining a first node and a second node related to the product unit node, where the first node is a node connected to the first node and the second node is a node connected to the first node; determining the node features of the product unit node based on the feature information of the first node and the feature information of the second node; determining the target product unit node based on the node features of the product unit node; and determining the corresponding product name of the target product unit node if the similarity between the node features of the target product unit node and the node features of the product unit meets a preset condition. Based on the labeled product name and target product name corresponding to the product information, the analysis model is adjusted to determine the product name of the product to be identified based on the trained analysis model.

5. The method according to claim 4, characterized in that, Based on the feature information of the nodes, the feature information of the first node, and the feature information of the second node, the node features of each node are determined, including: The feature information of the corresponding first node is incorporated into each target node to obtain the feature vector of each target node; The feature integration weights are determined based on the feature similarity between the feature vector of the target node and the feature vector of the second node of the target node. Based on the aforementioned feature integration weights, the feature vector of the second node is integrated into the feature vector of the target node to form the node features of the target node.

6. A data processing method, characterized in that, The method includes: Provide an interactive page to obtain product information of the product to be identified; Based on the product information, a product knowledge graph of the product to be identified is determined. The product knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. The knowledge graph is input into a pre-trained analysis model, which outputs the product name of the product to be identified. The analysis model is trained based on a knowledge graph constructed from labeled product information, and the nodes of the knowledge graph include product name nodes. The analysis model performs the following steps: Determine the first node and the second node related to the product unit node, where the first node is a node connected to the first node and the second node is a node connected to the first node. Based on the feature information of the first node and the feature information of the second node, the node features of the commodity unit node are determined; Based on the node characteristics of the product unit node, determine the target product unit node; If the similarity between the node features of the target product unit node and the node features of the product unit meets the preset conditions, the corresponding product name of the target product unit node is determined. Please provide the product name.

7. A data processing apparatus, characterized in that, The device includes: The knowledge graph acquisition module is used to acquire product information of the product to be identified, and determine the product knowledge graph of the product to be identified based on the product information. The product knowledge graph includes nodes and connecting edges between nodes. The nodes include at least one of product unit nodes, product category nodes, product name nodes, and product attribute nodes. The product name acquisition module is used to input the knowledge graph into a pre-trained analysis model, and output the product name of the product to be identified through the analysis model. The analysis model is trained based on a knowledge graph constructed from labeled product information, and the nodes of the knowledge graph include product name nodes. The analysis model performs the following steps: determining a first node and a second node related to the product unit node, where the first node is a node connected to the first node, and the second node is a node connected to the first node; determining the node features of the product unit node based on the feature information of the first node and the feature information of the second node; determining the target product unit node based on the node features of the product unit node; and determining the corresponding product name of the target product unit node if the similarity between the node features of the target product unit node and the node features of the product unit meets a preset condition.

8. An electronic device, characterized in that, include: processor; and A memory having executable code stored thereon, which, when executed, causes the processor to perform the method as described in one or more of claims 1-6.

9. One or more machine-readable media having executable code stored thereon, which, when executed, causes a processor to perform the method as described in one or more of claims 1-6.

Images