Product model attribute identification method and system, storage medium and electronic equipment

Abstract

The invention provides a product model attribute identification method and system, a storage medium and electronic equipment. The method comprises the following steps: receiving product information input by a user; the product information comprises a plurality of product attribute categories; the product attribute category comprises a product name attribute, a product model attribute and a product quantity attribute; each product attribute category corresponds to a plurality of product attribute data; obtaining the possibility that each piece of product attribute data under each product attribute category is the product model attribute data; and based on the possibility that all the obtained product attribute data are the product model attribute data, obtaining an identification result of the product model attribute from the plurality of product attribute categories. According to the method and the device, the product model attribute can be automatically and accurately identified in the plurality of product attribute categories in the product information under different product model endowing rules, so that the corresponding product model attribute data is acquired, and the cost of manual identification is reduced.

Description

Technical Field

[0001] The present application belongs to the technical field of electronic products, and in particular, relates to a method, system, storage medium and electronic device for identifying product model attributes. Background Art

[0002] The domestic and international electronic products industry covers a wide range, involving a large number of manufacturers, and each manufacturer usually adopts its own rules and standards when naming its products. Even for different series or categories of products produced by the same manufacturer, the model rules may be significantly different. Some manufacturers name the model according to the function, performance or market positioning of the product, while others may assign the product model through production batch, internal coding or other custom rules. Therefore, there is no unified standard or specification in the industry that can accurately cover all product models. This leads to the fact that in actual work, identifying the model of electronic products often becomes a complex and challenging task. Usually, this kind of work can only be done through manual judgment. Although this method can cope with a variety of model forms, it is very time-consuming and labor-intensive, and due to the subjectivity of human judgment, it is difficult to guarantee complete accuracy. At the same time, manual judgment also places high demands on the operator's experience and understanding of the industry.

[0003] In addition, another common method is to match the product models that have been entered into the database. This method can improve the efficiency of model recognition to a certain extent, but it depends on the update frequency and comprehensiveness of the database. If the database is not updated in time, or new products of some manufacturers have not yet been entered, then even the database matching method may face the situation of being unable to recognize or misjudgment. To make it more complicated, new products and new technologies are emerging in the electronics industry, and the rules for assigning product models are constantly evolving and changing. Therefore, existing databases often cannot fully cover the models of all the latest products. In this case, relying solely on the database may lead to major omissions, especially after the release of new products, the timely update of models is particularly important.

[0004] In summary, whether it is manual recognition or database matching, the existing electronic product model recognition methods have limitations, especially in the face of the rapidly changing electronic product market, this problem is particularly prominent. Summary of the invention

[0005] The purpose of this application is to provide a method, system, storage medium and electronic device for identifying product model attributes, so as to solve the technical problem of how to automatically identify the model attributes of electronic products.

[0006] In a first aspect, the present application provides a method for identifying a product model attribute, the method comprising:

[0007] Receive product information input by a user; the product information includes several product attribute categories; the product attribute categories include product name attributes, product model attributes, product parameter attributes and product quantity attributes; each product attribute category corresponds to several product attribute data;

[0008] Obtaining the possibility that each of the product attribute data under each of the product attribute categories is the product model attribute data;

[0009] Based on the possibility that all the acquired product attribute data are the product model attribute data, the recognition results of the product model attributes are acquired in several product attribute categories.

[0010] In an implementation of the first aspect, the possibility of obtaining each of the product attribute data under each of the product attribute categories as the product model attribute includes:

[0011] Using the trained product model attribute recognition model, determining whether each of the product attribute data under the product attribute category contains empty information or non-ASCII characters;

[0012] If the empty information or the non-ASCII characters are contained, the trained product model attribute recognition model determines that the corresponding product attribute data is not the product model attribute data;

[0013] If the empty information and the non-ASCII characters are not contained, the trained product model attribute recognition model determines the possibility that the corresponding product attribute data is the product model attribute data.

[0014] In an implementation of the first aspect, obtaining an identification result of a product model attribute from among the plurality of product attribute types includes:

[0015] For each of the product attribute categories, obtain a corresponding average possibility result; the average possibility result is the average value of the possibility that the product attribute data under each of the product attribute categories is the product model attribute data;

[0016] The product attribute category with the highest average likelihood result is determined as the product model attribute.

[0017] In an implementation of the first aspect, the method further includes:

[0018] Acquire a model attribute recognition data set; the model attribute recognition data set includes a model attribute recognition training set and a model attribute recognition test set;

[0019] The predetermined original model is trained based on the model attribute recognition training set and the model attribute recognition test set to obtain the trained product model attribute recognition model; wherein,

[0020] The predetermined original model is trained for multiple rounds on the model attribute recognition training set; if the training loss value of multiple consecutive rounds does not exceed the loss threshold, the product model attribute recognition model is obtained;

[0021] The product model attribute recognition model is tested on the model attribute recognition test set to obtain the trained product model attribute recognition model.

[0022] In an implementation of the first aspect, obtaining the model identification data set includes:

[0023] Acquire the product attribute training data from a database;

[0024] Label the product attribute training data according to the product attribute category to obtain a first data table;

[0025] Create a second data table corresponding to the first data table, the second data table is in DataFrame data format; in the second data table, the product model attribute category is marked as 1, and the non-product model attribute category is marked as 0;

[0026] The first data table and the second data table are merged and randomized to obtain the model identification data set.

[0027] In an implementation of the first aspect, performing multiple rounds of training on the predetermined original model on the model attribute recognition training set includes:

[0028] Vectorizing the product model attribute training data in the model attribute recognition training set to obtain a number of model vectors;

[0029] The predetermined original model is trained based on a plurality of the model vectors to obtain the product model attribute recognition model.

[0030] In an implementation of the first aspect, vectorizing the product model attribute training data in the model attribute recognition training set includes:

[0031] Select adjacent character strings of length [1, a] in the product model attribute training data, where a∈(1, b], b is the character length of the product model attribute training data;

[0032] Obtaining the number of occurrences of the adjacent character strings in the training data corresponding to the product model attributes;

[0033] Compare the number of occurrences with the number of all adjacent character strings of length [1, a] in the corresponding product model attribute training data to obtain a first result;

[0034] Obtaining the inverse of the quantity ratio of all the product model attribute training data containing the adjacent character strings in the model attribute recognition training set as a second result;

[0035] A model vector is obtained based on the first result and the second result.

[0036] In a second aspect, the present application provides a system for identifying product model attributes, the system comprising:

[0037] A receiving module, used to receive product demand information input by a user; the product demand information includes several product attribute categories; the product attribute categories include product name attributes, product model attributes, product parameter attributes and product quantity attributes; each of the product attribute categories corresponds to several product attribute data;

[0038] An acquisition module, used for acquiring the possibility that each of the product attribute data under each of the product attribute categories is the product model attribute;

[0039] The identification module is used to obtain the identification result of the product model attribute in several product attribute types based on the possibility that all the obtained product attribute data are the product model attributes.

[0040] In a third aspect, the present application provides an electronic device, comprising: a processor and a memory; the memory is used to store a computer program; the processor is used to execute the computer program stored in the memory, so that the electronic device performs the above-mentioned method for identifying product model attributes.

[0041] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by an electronic device, implements the above-mentioned method for identifying product model attributes.

[0042] As described above, the product model attribute identification method, system, storage medium and electronic device described in the present application have the following beneficial effects: they can automatically and accurately identify product model attributes in multiple product attribute categories in product information under different product model assignment rules, and then obtain corresponding product model attribute data, thereby reducing the cost of manual identification. By applying this application, electronic product sellers can automatically identify the electronic product models that purchasing users need to purchase, which is conducive to reducing the sales cost of electronic products and can accurately provide purchasing users with corresponding electronic products. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1A Shown is a schematic diagram of an application scenario of the method for identifying product model attributes described in this application.

[0044] Figure 1B Shown are structural diagrams of the client-cloud interaction scenarios in these implementations.

[0045] Figure 2 Shown is a flow chart of a method for identifying product model attributes described in an embodiment of the present application in one embodiment.

[0046] Figure 3 Shown is a flow chart of a method for identifying product model attributes described in an embodiment of the present application in one embodiment.

[0047] Figure 4 Shown is a flow chart of a method for identifying product model attributes described in an embodiment of the present application in one embodiment.

[0048] Figure 5 Shown is a flow chart of a method for identifying product model attributes described in an embodiment of the present application in one embodiment.

[0049] Figure 6 Shown is a flow chart of a method for identifying product model attributes described in an embodiment of the present application in one embodiment.

[0050] Figure 7 Shown is a schematic diagram of the structure of a system for identifying product model attributes in an embodiment of the present application.

[0051] Figure 8 Shown is a schematic structural diagram of an electronic device of the present application in one embodiment.

[0052] Component number description

[0053] 1. Identification device for product model attributes

[0054] 11 Processor

[0055] 12 Database

[0056] 2-end-cloud interactive system

[0057] 20 Terminal

[0058] 21 Cloud Server

[0059] 4 Product model attribute identification system

[0060] 41 Receiver Module

[0061] 42 Get Module

[0062] 43 Identification Module

[0063] 51 Processing Units

[0064] 52 Memory

[0065] 521 Memory

[0066] 522 Cache memory

[0067] 523 Storage System

[0068] 524 Programs / Utilities

[0069] 5241 Program Module

[0070] 53 Bus

[0071] 54 I / O interfaces

[0072] 55 Network Adapter

[0073] Steps S1 to S3

[0074] Steps S21 to S23

[0075] Steps S4-S5

[0076] Steps S41 to S44

[0077] Steps S51 to S55 DETAILED DESCRIPTION

[0078] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0079] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present application, and thus the drawings only show components related to the present application rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed at will, and the component layout may also be more complicated.

[0080] In addition, in this application, descriptions such as "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in this field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0081] At present, with the development of Internet technology, e-commerce malls have become the first choice for many users to purchase electronic components. An e-commerce mall is an online platform dedicated to selling electronic products. It usually includes websites, mobile applications and social media integration, providing a variety of electronic products. For users who need to buy a large number of electronic products, they often use Excel spreadsheets to organize the information of the electronic products they want to buy, including product names, product models, product parameters and product quantities. After receiving the purchase order sent by the user, the e-commerce mall will pick and ship the corresponding goods to complete the sales process. However, during the picking process, the seller often needs to manually identify the product model in the purchase order or identify it through database comparison. However, since each manufacturer usually adopts its own rules and standards when naming its electronic products. Even if the same manufacturer produces different series or categories of products, the product model rules may be significantly different. This adds many obstacles for electronic product sellers in identifying product models.

[0082] In order to at least solve the above-mentioned problems, an embodiment of the present application provides a method for identifying product model attributes, which can automatically and accurately identify product model attributes in product information and corresponding product model attribute data, thereby reducing the cost of manual identification.

[0083] The method for identifying product model attributes provided in the embodiment of the present application can be applied to Figure 1A The identification device of the product model attribute shown. Figure 1A As shown, the device 1 for identifying product model attributes includes a processor 11 and a database 12. The method for identifying product model attributes provided in the embodiment of the present application can be applied to the processor 11.

[0084] in, Figure 1A The processor 11 in the embodiment may be a single processor or a processor cluster or a cloud computing center composed of multiple processors, and the specific details are not limited here. Figure 1A Only one processor 11 and one database 12 are shown, but it should be understood that Figure 1AThe examples are only used to understand the present solution, and the specific numbers of processors 11 and databases 12 should be flexibly determined based on actual conditions.

[0085] In some other embodiments, the device 1 for identifying product model attributes may not include the database 12, but only include a processor 11 with a storage function. The method for identifying product model attributes provided in the embodiment of the present application can be applied to the processor 11. The processor 11 with a storage function may include a tablet computer, a laptop computer, a PDA, a mobile phone, a personal computer (PC), etc., which is not limited here.

[0086] In still other embodiments, the method for identifying product model attributes described in the present application may be applied to a terminal-cloud interaction scenario. Figure 1B The following is a schematic diagram of the structure of the client-cloud interaction scenario in these implementations. Figure 1B As shown, the terminal-cloud interaction system 2 includes a terminal 20 and a cloud server 21. The terminal 20 and the cloud server 21 can communicate with each other, and the communication method is not limited to wired or wireless.

[0087] Among them, the terminal 20 can be mobile or fixed, for example, the terminal 20 can be a wireless terminal or a wired terminal. The wireless terminal can refer to a device with wireless transceiver function, which can be deployed indoors, outdoors and in industrial workshops. The terminal 20 can be a mobile phone, a tablet computer, a laptop computer, etc., which is not limited here. The cloud server 21 can include one or more servers, or one or more processing nodes, or one or more virtual machines running on the server. The cloud server 21 can also be called a server cluster, a management platform, an identification center for product model attributes, etc., which is not limited in the embodiments of the present application.

[0088] The technical solutions in the embodiments of the present application will be described in detail below in conjunction with the drawings in the embodiments of the present application.

[0089] The following embodiments of the present application provide a method for identifying a product model attribute, which can be performed by Figure 1A The processor 11 or Figure 1B It is implemented by the cloud server 21 shown. Figure 2 The flowchart of the method for identifying the product model attribute according to the embodiment of the present application is shown as follows: Figure 2 As shown, the method for identifying product model attributes includes steps S1 to S3.

[0090] S1. Receive product information input by a user; the product information includes several product attribute categories; the product attribute categories include product name attributes, product model attributes, product parameter attributes and product quantity attributes; each of the product attribute categories corresponds to several product attribute data.

[0091] In some embodiments, the product information input by the user is presented as an Excel table, in which each column represents a different product attribute category, and each column has a corresponding number of product attribute data. Among them, the product attribute categories include product name attributes, product model attributes, product parameter attributes, and product quantity attributes. That is, the Excel table input by the user as a whole contains the product information that the user wants to purchase, including product name, product model, product parameters, and product quantity.

[0092] S2. Obtain the possibility that each product attribute data under each product attribute category is the product model attribute data.

[0093] Specifically, the product information input by the user is judged, and it is analyzed which product attribute data in the product information is product model attribute data.

[0094] In some embodiments, after receiving the Excel spreadsheet sent by the user, the seller will identify the spreadsheet, mainly to determine which column in the Excel spreadsheet is the product model attribute, and the product attribute data corresponding to the column is the product model attribute data. In this way, the specific product model attribute data to be sold can be determined, and the product is prepared for sale based on this.

[0095] Figure 3 The flowchart of the method for identifying the product model attribute according to the embodiment of the present application is shown as follows: Figure 3 As shown, the method for identifying product model attributes includes steps S21 to S23.

[0096] S21. Using the trained product model attribute recognition model, determine whether each product attribute data under the product attribute category contains empty information or non-ASCII characters.

[0097] S22. If the empty information or the non-ASCII characters are contained, the trained product model attribute recognition model determines that the corresponding product attribute data is not the product model attribute data.

[0098] S23: If the empty information and the non-ASCII characters are not contained, the trained product model attribute recognition model determines the possibility that the corresponding product attribute data is the product model attribute data.

[0099] Specifically, the trained product model attribute recognition model is used to analyze which product attribute data in the product information are product model attribute data. If the trained product model attribute recognition model determines that the product attribute data contains empty information or non-ASCII characters, then the possibility of determining that the product attribute data is the product model attribute data is 0; if the trained product model attribute recognition model determines that the product attribute data does not contain empty information and non-ASCII characters, then the possibility of determining that the corresponding product attribute data is the product model attribute data is 0.

[0100] In some embodiments, when a user sends product information in an Excel table, each column in the table represents a different product attribute type. At this time, the trained product model attribute recognition model provided in the embodiment of the present application judges all cells (product attribute data) in the Excel table. If the cell is empty or contains non-ASCII characters, the possibility that the cell is filled with product model attribute data is 0; if the cell is not empty and does not contain non-ASCII characters, the data in the cell may be product model attribute data.

[0101] It should be noted that when the product attribute data does not contain empty information and non-ASCII characters, the trained product model attribute recognition model will automatically obtain the probability of the product model attribute data according to the previous training process and training data set. That is, when the model determines that the product attribute data is closer to the product model format, it will give a higher probability. If it does not contain empty information and non-ASCII characters, but the characters are short or have other expressions, it will give a lower probability.

[0102] It should be noted that the method for identifying product model attributes provided in the embodiment of the present application also includes a method for training the product model attribute identification model. Figure 4 The flowchart of the training method of the product model attribute recognition model described in the embodiment of the present application is shown as follows: Figure 4 As shown, the training method of the product model attribute recognition model includes steps S4 to S5.

[0103] S4. Obtain a model attribute recognition data set; the model attribute recognition data set includes a model attribute recognition training set and a model attribute recognition test set.

[0104] Specifically, Figure 5 The flowchart of the training method of the product model attribute recognition model described in the embodiment of the present application is shown as follows: Figure 5 As shown, the training method of the product model attribute recognition model includes steps S41 to S44.

[0105] S41. Acquire the product attribute training data from a database.

[0106] In some embodiments, original data related to product attributes, including detailed information such as product model, name, specifications, etc., are obtained from a MongoDB database to serve as basic training data for training the original model.

[0107] S42: Label the product attribute training data according to the product attribute category to obtain a first data table.

[0108] In some embodiments, the product attribute training data obtained from the database is classified and labeled according to the product attribute category to generate a first data table. Each product is labeled accordingly according to its attribute (e.g., model, name), forming a table containing the original product attribute data and its category label.

[0109] S43. Create a second data table corresponding to the first data table, wherein the second data table is in DataFrame data format; in the second data table, the product model attribute category is marked as 1, and the non-product model attribute category is marked as 0.

[0110] In some embodiments, a second data table corresponding to the first data table is generated and is in the format of DataFrame. DataFrame is a common structured data table format and can usually be processed using tools such as pandas. The second data table is used to perform further marking operations.

[0111] Furthermore, in the second data table, the product model attribute category is marked as 1, and other non-product model attributes are marked as 0. This is to transform the original classification problem into a binary classification problem, which is convenient for subsequent model training. In this case, identifying whether the product attribute data is product model attribute data becomes the target task.

[0112] S44: merge and randomize the first data table and the second data table to obtain the model identification data set.

[0113] In some embodiments, the first data table and the second data table are merged together to form a complete data set containing original product attributes and binary classification labels. Afterwards, the data is randomized (shuffled) to disrupt the order of the data to ensure the randomness of the data during model training, avoid deviations in the learning order, and improve the generalization ability of the model. Finally, one quarter of the data set is used as the model attribute recognition test set, and the rest is used as the model attribute recognition training set.

[0114] S5. Train a predetermined original model based on the model attribute recognition training set and the model attribute recognition test set to obtain the trained product model attribute recognition model.

[0115] Specifically, after obtaining the model attribute recognition training set and the model attribute recognition test set, the predetermined original model is trained for multiple rounds on the model attribute recognition training set. If the training loss value of multiple consecutive rounds does not exceed the loss threshold, the product model attribute recognition model is obtained, and then the product model attribute recognition model is tested on the model attribute recognition test set to obtain the trained product model attribute recognition model.

[0116] In some embodiments, when training the model, it is necessary to vectorize the product model attribute training data in the model attribute recognition training set to obtain a number of model vectors. Then, the predetermined original model is trained based on the model vectors to obtain the product model attribute recognition model.

[0117] Figure 6 The flowchart of the vectorization process described in the embodiment of the present application is shown as follows: Figure 6 As shown, vectorizing the product model attribute training data in the model attribute recognition training set includes steps S51 to S55.

[0118] S51. Select adjacent character strings of length [1, a] in the product model attribute training data, where a∈(1, b], and b is the character length of the product model attribute training data.

[0119] Specifically, the product model attribute training data is set to d, and then an adjacent character string t of length [1, a] is selected in d. Where a∈(1, b], b is the character length of the product model attribute training data. That is, if d is xingh, the length b is 5, and then an adjacent character string t of length [1, a] is selected in "xingh", and the length of the adjacent character string t is at most consistent with d.

[0120] It should be noted that the exhaustive principle is adopted to select adjacent character strings of length [1, a] in the product model attribute training data, that is, adjacent character strings satisfying the length [1, a] need to be selected.

[0121] In some embodiments, an adjacent character string t of length [1,3] is selected from d "xingh" of length 5. Then the adjacent character string t includes "x", "i", "n", "g", "h", "xi", "in", "ng", "gh", "xin", "ing", and "ngh".

[0122] S52: Obtain the number of occurrences of the adjacent character strings in the training data corresponding to the product model attribute.

[0123] Specifically, obtain the number of times f the adjacent string t appears in the model attribute training data d t,d .

[0124] S53: Compare the number of occurrences with the number of adjacent character strings of length [1, a] in the corresponding product model attribute training data to obtain a first result.

[0125] Specifically, the first result can be expressed as:

[0126]

[0127] For example, in the above embodiment, the adjacent character string t "x" appears once in "xingh", f t,d is 1, that is, the number of all adjacent strings of length [1, 3] is 12, and the first result is 1 / 12.

[0128] S54. Obtain the inverse of the quantity ratio of all the product model attribute training data containing the adjacent character strings in the model attribute recognition training set as a second result.

[0129] Specifically, find the inverse of the proportion of the number of models containing t in the training set D:

[0130]

[0131] S55. Obtain a model vector based on the first result and the second result.

[0132] Specifically, multiply the first result tf(t,d) and the second result idf(t,D):

[0133] tfidf(t,d,D)=tf(t,d)·idf(t,D) (3)

[0134] In this way, each string of adjacent characters is converted into numbers, and each product model attribute data is converted into a vector of the adjacent characters it contains, that is, a model vector.

[0135] In some embodiments, after the product model attribute training data is converted into a vector of adjacent characters (model vector) through vectorization processing, the multi-layer perceptron classifier (Multi-Layer Perceptron Classifier) ​​of scikit-learn is used to train a neural network on the obtained model vector. If the training loss value for 10 consecutive rounds does not exceed 0.0001, the training is stopped and the product model attribute recognition model is obtained. Afterwards, the product model attribute recognition model is tested on the model attribute recognition test set to obtain the trained product model attribute recognition model.

[0136] In some embodiments, during the training process of the model, the Adam algorithm can be used to dynamically adjust the learning rate so that the parameter update during the training process is both stable and efficient. By using the exponentially weighted moving average method, the Adam algorithm can effectively cope with the situation where the gradient changes greatly and adaptively adjust the learning rate of each parameter. Specifically, for each round of iteration, the following operations are performed:

[0137] (1) Initialization, that is, setting hyperparameters: learning rate α = 0.0001, momentum decay coefficient β1 is 0.9, mean square decay coefficient β2 is 0.999, and smoothing term ∈ is 10 -8 .

[0138] (2) Initialize the first-order momentum estimate and the second-order momentum estimate, i.e., the initial bias correction term: initialize the initial value m0 of the first-order momentum and the initial value v0 of the second-order momentum to 0.

[0139] (3) Perform the following operations in each iteration:

[0140] -Calculate the gradient: according to the current parameter θ t-1 , calculate the gradient g of the objective function f t

[0141]

[0142] - Update the first-order momentum estimate (exponentially weighted moving average of momentum)

[0143] m t =β1·m t-1 +(1-β1)g t (5)

[0144] Among them, m t represents the weighted average of the gradients.

[0145] -Update the second-order momentum estimate (exponentially weighted moving average of squared gradients)

[0146]

[0147] Among them, v t Represents the weighted average of the squared gradients.

[0148] -Bias correction for momentum and mean square error (because the initial value of momentum is 0, it may cause estimation bias in the first few iterations)

[0149]

[0150] in, and are bias-corrected momentum and mean square error estimates.

[0151] - Update the parameters θ t : Parameter update is done by the modified gradient estimate m t and the mean square error of the gradient Then multiply it by the learning rate α to adjust it.

[0152]

[0153] If the training loss value for 10 consecutive rounds does not exceed 0.0001, the training is stopped and the product model attribute recognition model is obtained. In some embodiments, the trained product model attribute recognition model can be expressed as:

[0154]

[0155] Where x is the input vector of the neural network, E is the set of synapses (connections between neurons), and v t.r refers to the rth neuron in the tth layer, a t,r refers to its input, o t,r refers to its output, v t+1.j refers to the jth neuron in the t+1th layer, w refers to the weight function (the input is the element of E), and σ refers to the activation function of the neuron.

[0156] Each neuron is connected to all neurons in the previous layer, and the output of the previous neuron is multiplied and added by the weight of the synapse connecting it. Therefore, the output of the jth neuron in the t+1th layer can be expressed as:

[0157] o t+1,j (x) = σ(a t+1,j (x)) (11)

[0158] Among them, the activation function of the neuron is expressed as:

[0159]

[0160] It should be noted that, for the product model attribute recognition model, the more layers of the neural network, the more accurate the recognition result. However, too many layers will lead to too long calculation time and may also cause overfitting. The more neurons in each layer, the more accurate the recognition effect, but the greater the computing resources required, and the number of neurons in each layer can actually be the same or different. For this application, after a large amount of data training, it was found that when the neural network of the product model attribute recognition model contains 2 layers, each layer contains 50 neurons, which is the best in balancing training time, required resources and recognition effect.

[0161] In some embodiments, the trained product model attribute recognition model is exported and saved in ONNX (Open Neural Networks Exchange) format to obtain the possibility that each product attribute data under each product attribute category is the product model attribute data.

[0162] S3. Based on the possibility that all the acquired product attribute data are the product model attribute data, obtain the recognition results of the product model attribute in several product attribute categories.

[0163] Specifically, for each of the product attribute categories, a corresponding average possibility result is obtained; the average possibility result is the average value of the possibility that the product attribute data under each of the product attribute categories is the product model attribute data.

[0164] Specifically, the product attribute category with the highest average likelihood result is determined as the product model attribute.

[0165] In some embodiments, the possibility that all product attribute data in the product information is the product model attribute data can be obtained through step S2. Since the product attribute data corresponds to different attribute categories, the average value of the possibility that the product attribute data under each category is the product model attribute data can be calculated. When the average value is the highest, it means that the corresponding category should be the product model attribute category.

[0166] For example, a user inputs product information in the form of an Excel table, where each column of the table corresponds to a different product attribute category. At this time, the seller cannot determine which column is the product model attribute, so it is necessary to identify the data one by one to determine the product model required by the user. Then, through step S2, the possibility of whether all cells in the Excel table are product model attribute data can be calculated, and the average probability of each column can be calculated. When the average value is the highest, it means that the column should be a product model attribute category, and all data under the column is product model attribute data. In this way, the seller can automatically determine the product model attribute in the product information sent by the user, and obtain all the corresponding product model attribute data.

[0167] The protection scope of the method for identifying product model attributes described in the embodiment of the present application is not limited to the execution order of the steps listed in the present embodiment. All solutions implemented by adding, reducing or replacing steps in the prior art based on the principles of the present application are included in the protection scope of the present application.

[0168] The embodiments of the present application also provide a system for identifying product model attributes, which can implement the method for identifying product model attributes described in the present application. However, the implementation device of the system for identifying product model attributes described in the present application includes but is not limited to the structure of the system for identifying product model attributes listed in the present embodiment. All structural deformations and replacements of the prior art made according to the principles of the present application are included in the protection scope of the present application.

[0169] Figure 7 The structure diagram of the product model attribute recognition system described in the embodiment of the present application is shown as follows: Figure 7 As shown, the product model attribute identification system 4 includes a receiving module 41 , an acquisition module 42 and an identification module 43 .

[0170] The receiving module 41 is used to receive product demand information input by the user; the product demand information includes several product attribute categories; the product attribute categories include product name attributes, product model attributes, product parameter attributes and product quantity attributes; each of the product attribute categories corresponds to several product attribute data;

[0171] An acquisition module 42, used for acquiring the possibility that each of the product attribute data under each of the product attribute categories is the product model attribute;

[0172] The identification module 43 is used to obtain the identification result of the product model attribute in several product attribute types based on the possibility that all the obtained product attribute data are the product model attribute.

[0173] It should be noted that the structures and principles of the receiving module 41 , the obtaining module 42 and the identifying module 43 correspond one-to-one to the steps in the above-mentioned product model attribute identification method, so they are not described in detail here.

[0174] In the several embodiments provided in the present application, it should be understood that the disclosed system, device or method can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of modules / units is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules or units, which can be electrical, mechanical or other forms.

[0175] The modules / units described as separate components may or may not be physically separated, and the components displayed as modules / units may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules / units may be selected according to actual needs to achieve the purpose of the embodiments of the present application. For example, the functional modules / units in the various embodiments of the present application may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0176] Those of ordinary skill in the art should further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0177] The present application embodiment also provides a computer-readable storage medium. A person of ordinary skill in the art can understand that all or part of the steps in the method for implementing the above embodiment can be completed by a program to instruct the processor, and the program can be stored in a computer-readable storage medium, and the storage medium is a non-transitory medium, such as a random access memory, a read-only memory, a flash memory, a hard disk, a solid-state hard disk, a magnetic tape, a floppy disk, an optical disc, and any combination thereof. The above storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a digital video disc (DVD)), or a semiconductor medium (e.g., a solid-state hard disk (SSD)), etc.

[0178] The embodiment of the present application further provides an electronic device, wherein the electronic device includes a processor and a memory.

[0179] The memory is used to store computer programs.

[0180] The memory includes: ROM, RAM, disk, USB flash drive, memory card or CD and other media that can store program codes.

[0181] The processor is connected to the memory and is used to execute the computer program stored in the memory so that the electronic device executes the above-mentioned method for identifying product model attributes.

[0182] Preferably, the processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0183] like Figure 8As shown, the electronic device of the present application is in the form of a general computing device. The components of the electronic device may include but are not limited to: one or more processors or processing units 51, a memory 52, and a bus 53 connecting different system components (including the memory 52 and the processing unit 51).

[0184] Bus 53 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures. For example, these architectures include but are not limited to Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus and Peripheral Component Interconnect (PCI) bus.

[0185] Electronic devices typically include a variety of computer system readable media. These media can be any available media that can be accessed by the electronic device, including volatile and non-volatile media, removable and non-removable media.

[0186] The memory 52 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 521 and / or cache memory 522. The electronic device may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 523 may be used to read and write non-removable, non-volatile magnetic media ( Figure 8 not shown, usually called a "hard drive"). Although Figure 8 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical medium) may be provided. In these cases, each drive may be connected to the bus 53 via one or more data medium interfaces. The memory 52 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present application.

[0187] A program / utility 524 having a set (at least one) of program modules 5241 may be stored, for example, in the memory 52, such program modules 5241 including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules 5241 generally perform the functions and / or methods of the embodiments described herein.

[0188] The electronic device may also communicate with one or more external devices (e.g., keyboards, pointing devices, displays, etc.), one or more devices that enable a user to interact with the electronic device, and / or any device that enables the electronic device to communicate with one or more other computing devices (e.g., network cards, modems, etc.). Such communication may be performed via input / output (I / O) interface 54. Furthermore, the electronic device may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via network adapter 55. Figure 8 As shown, the network adapter 55 communicates with other modules of the electronic device via the bus 53. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the electronic device, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0189] The present application embodiment may also provide a computer program product, the computer program product including one or more computer instructions. When the computer instructions are loaded and executed on a computing device, the process or function described in the embodiment of the present application is generated in whole or in part. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from a website site, computer or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer or data center.

[0190] When the computer program product is executed by a computer, the computer executes the method described in the above method embodiment. The computer program product may be a software installation package, and when the above method is required, the computer program product may be downloaded and executed on a computer.

[0191] The present application provides a method, system, storage medium and electronic device for identifying product model attributes, which can automatically and accurately identify product model attributes in multiple product attribute categories in product information under different product model assignment rules, and then obtain corresponding product model attribute data, reducing the cost of manual identification. By using this application, electronic product sellers can automatically identify the electronic product models that purchasers need to purchase, which is conducive to reducing the sales cost of electronic products and accurately providing purchasers with corresponding electronic products.

[0192] The descriptions of the processes or structures corresponding to the above-mentioned figures have different emphases. For parts that are not described in detail in a certain process or structure, please refer to the relevant descriptions of other processes or structures.

[0193] The above embodiments are merely illustrative of the principles and effects of the present application and are not intended to limit the present application. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed in the present application shall still be covered by the claims of the present application.

Claims

1. A method for identifying product model attributes, characterized in that: The method comprises: Receive product information input by a user; the product information includes several product attribute categories; the product attribute categories include product name attributes, product model attributes, product parameter attributes and product quantity attributes; each product attribute category corresponds to several product attribute data; Obtaining the possibility that each of the product attribute data under each of the product attribute categories is the product model attribute data; Based on the possibility that all the acquired product attribute data are the product model attribute data, the recognition results of the product model attributes are acquired in several product attribute categories.

2. The method for identifying product model attributes according to claim 1, characterized in that: Possibilities of obtaining each of the product attribute data under each of the product attribute categories as the product model attribute include: Using the trained product model attribute recognition model, determining whether each of the product attribute data under the product attribute category contains empty information or non-ASCII characters; If the empty information or the non-ASCII characters are contained, the trained product model attribute recognition model determines that the corresponding product attribute data is not the product model attribute data; If the empty information and the non-ASCII characters are not contained, the trained product model attribute recognition model determines the possibility that the corresponding product attribute data is the product model attribute data.

3. The method for identifying product model attributes according to claim 1, characterized in that: Obtaining the identification result of the product model attribute in the several product attribute types includes: For each of the product attribute categories, obtain a corresponding average possibility result; the average possibility result is the average value of the possibility that the product attribute data under each of the product attribute categories is the product model attribute data; The product attribute category with the highest average likelihood result is determined as the product model attribute.

4. The method for identifying product model attributes according to claim 2, characterized in that: The training method of the product model attribute recognition model includes: Acquire a model attribute recognition data set; the model attribute recognition data set includes a model attribute recognition training set and a model attribute recognition test set; The predetermined original model is trained based on the model attribute recognition training set and the model attribute recognition test set to obtain the trained product model attribute recognition model; wherein, The predetermined original model is trained for multiple rounds on the model attribute recognition training set; if the training loss value of multiple consecutive rounds does not exceed the loss threshold, the product model attribute recognition model is obtained; The product model attribute recognition model is tested on the model attribute recognition test set to obtain the trained product model attribute recognition model.

5. The method for identifying product model attributes according to claim 4, characterized in that: Obtaining the model recognition dataset includes: Acquire the product attribute training data from a database; Label the product attribute training data according to the product attribute category to obtain a first data table; Create a second data table corresponding to the first data table, where the second data table is in DataFrame data format; The product model attribute category in the second data table is marked as 1, and the non-product model attribute category is marked as 0; The first data table and the second data table are merged and randomized to obtain the model identification data set.

6. The method for identifying product model attributes according to claim 4, characterized in that: The predetermined original model is trained multiple times on the model attribute recognition training set, including: Vectorizing the product model attribute training data in the model attribute recognition training set to obtain a number of model vectors; The predetermined original model is trained based on a plurality of the model vectors to obtain the product model attribute recognition model.

7. The method for identifying product model attributes according to claim 6, characterized in that: Vectorizing the product model attribute training data in the model attribute recognition training set includes: Select adjacent character strings of length [1, a] in the product model attribute training data, where a∈(1, b], b is the character length of the product model attribute training data; Obtaining the number of occurrences of the adjacent character strings in the training data corresponding to the product model attributes; Compare the number of occurrences with the number of all adjacent character strings of length [1, a] in the corresponding product model attribute training data to obtain a first result; Obtaining the inverse of the quantity ratio of all the product model attribute training data containing the adjacent character strings in the model attribute recognition training set as a second result; A model vector is obtained based on the first result and the second result.

8. A product model attribute recognition system, characterized in that: The system comprises: A receiving module, used to receive product demand information input by a user; the product demand information includes several product attribute categories; the product attribute categories include product name attributes, product model attributes, product parameter attributes and product quantity attributes; each of the product attribute categories corresponds to several product attribute data; An acquisition module, used for acquiring the possibility that each of the product attribute data under each of the product attribute categories is the product model attribute; The identification module is used to obtain the identification result of the product model attribute in several product attribute types based on the possibility that all the obtained product attribute data are the product model attributes.

9. An electronic device, characterized in that: The electronic device comprises: a processor and a memory; The memory is used to store computer programs; The processor is used to execute the computer program stored in the memory, so that the electronic device executes the method for identifying product model attributes according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by an electronic device, the method for identifying the product model attribute described in any one of claims 1 to 7 is implemented.

Images