Website registration method and device, electronic equipment and storage medium

Through the attribute fusion deep learning model, the inconsistency between registration information and specifications is analyzed, and the registration information is automatically corrected, which solves the problems of inefficiency and process interruption in traditional website registration, and realizes an efficient and automated registration process.

CN118965321AActive Publication Date: 2024-11-15CHINA MOBILE GROUP ZHEJIANG +2

Patent Information

Application Number
CN202411247008.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-06
Publication Date
2024-11-15
Estimated Expiration
2044-09-06

AI Technical Summary

Technical Problem

Traditional website registration methods require users to manually fill in a large amount of information, resulting in low registration efficiency and may lead to registration failure and interruption of the process due to inconsistent information.

Method used

Using attribute fusion deep learning model, automatically corrects registration information by analyzing the inconsistency between registration information and registration specifications, including inputting registration prompt information into the model to obtain target prompt types and target registration specifications, and correcting information based on this to realize the website. register.

Benefits of technology

It improves registration efficiency, solves the registration failure problem caused by inconsistency in information, realizes an automated registration process, and reduces manual intervention.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118965321A_ABST
    Figure CN118965321A_ABST
Patent Text Reader

Abstract

The invention provides a website registration method and device, electronic equipment and a storage medium, and relates to the technical field of communication. The method comprises the following steps: during website registration, firstly, respectively adding registration information in a plurality of information input boxes based on respective types of the plurality of information input boxes of a registration webpage of a website, and sending each piece of registration information to a website server; receiving registration prompt information sent by the website server; under the condition that the registration prompt information is used for representing that at least one piece of registration information is inconsistent with the corresponding registration specification, inputting the registration prompt information into the attribute fusion deep learning model to obtain a target prompt type and a target registration specification, and then correcting the target registration information based on the target prompt type and the target registration specification to obtain the target registration information. And carrying out website registration based on the registration information consistent with the registration specification and the corrected target registration information. By adopting the technical scheme provided by the invention, the registration efficiency of the website can be effectively improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a website registration method, device, electronic device and storage medium. Background Art

[0002] The Internet is a huge network covering hundreds of millions of users worldwide. It connects people together and enables information sharing, communication and collaboration. Websites are the most important part of the Internet. They provide users with various information and services through web pages.

[0003] With the continuous development of Internet technology, the content and form of websites are becoming increasingly rich and diverse. From early static web pages to today's dynamic interactive web pages, whether it is getting news, shopping, socializing or working, it is inseparable from the support of websites. In order to use the services provided by these websites, users need to register on the website, which not only makes it easier for users to obtain personalized services, but also helps the website manage and maintain user information.

[0004] Traditional website registration methods usually require users to manually fill in a large amount of registration information, which results in low registration efficiency. Summary of the invention

[0005] The present application provides a website registration method, device, electronic device and storage medium, which are used to solve the problem in the prior art that users need to manually fill in a large amount of registration information, resulting in low registration efficiency.

[0006] The present application provides a website registration method, which is applied to an electronic device, and the method includes: Based on the types of the multiple information input boxes on the registration webpage of the website, respectively add registration information in the multiple information input boxes, and send each of the registration information to the website server; Receiving registration prompt information sent by the website server, wherein the registration prompt information is determined based on the registration information and the registration specification corresponding to each information input box; In a case where the registration prompt information is used to represent that at least one of the registration information is inconsistent with the corresponding registration specification, the registration prompt information is input into the attribute fusion deep learning model to obtain a target prompt type and a target registration specification, wherein the target prompt type is a type corresponding to the target registration information inconsistent with the registration specification, and the target registration specification is a registration specification corresponding to the target registration information; Based on the target prompt type and the target registration specification, the target registration information is modified, and website registration is performed based on the registration information consistent with the registration specification and the modified target registration information.

[0007] According to a website registration method provided by the present application, the registration prompt information includes HTML text, style attributes of nodes in the website and the hierarchical structure of the nodes; The inputting the registration prompt information into the attribute fusion deep learning model to obtain the target prompt type and the target registration specification corresponding to the target registration information inconsistent with the registration specification includes: Inputting the HTML text into a first embedding layer in the attribute fusion deep learning model to obtain a first text feature vector; and inputting the first text feature vector into an activation layer to obtain the target registration specification; Inputting the style attribute of the node into the second embedding layer in the attribute fusion deep learning model to obtain a string feature vector; Inputting the hierarchical structure of the nodes into the third embedding layer in the attribute fusion deep learning model to obtain a string feature matrix; The first text feature vector, the character string feature vector and the character string feature matrix are input into a post-processing module in the attribute fusion deep learning model to obtain the target prompt type.

[0008] According to a website registration method provided by the present application, the style attribute of the node includes a numerical type and a category type; The step of inputting the style attribute of the node into the second embedding layer in the attribute fusion deep learning model to obtain a string feature vector includes: Inputting the style attributes of the node into the second embedding layer in the attribute fusion deep learning model, and for numerical style attributes, normalizing the numerical style attributes to obtain normalized style attributes; For a category-type style attribute, encoding the category-type style attribute to obtain an encoded style attribute; The normalized style attribute and the encoded style attribute are concatenated to obtain the character string feature vector.

[0009] According to a website registration method provided by the present application, the hierarchical structure of the nodes is input into the third embedding layer in the attribute fusion deep learning model to obtain a string feature matrix, including: Inputting the hierarchical structure of the nodes into the third embedding layer in the attribute fusion deep learning model to determine the character string corresponding to each node; Based on the string set consisting of the strings corresponding to the nodes, determining the character-level feature matrix corresponding to the hierarchical structure of the nodes; and determining the word-level feature matrix corresponding to the hierarchical structure of the nodes based on the string set; The character-level feature matrix and the word-level feature matrix are weightedly concatenated to obtain the string feature matrix.

[0010] According to a website registration method provided by the present application, the character-level feature matrix corresponding to the hierarchical structure of the nodes is determined based on a character string set consisting of the character strings corresponding to the nodes, including: Performing a word segmentation operation on each string in a string set consisting of strings corresponding to each node to obtain a word segmentation result set; Generate a character-level vector corresponding to the string set according to the frequency of occurrence of the character elements in the word segmentation result set; Adjusting the dimensions of all vectors in the character-level vector to obtain a vector set, wherein all vectors in the vector set have the same dimension; Encoding all vectors in the vector set, and performing dimensionality reduction processing on the encoded matrix to obtain a dense matrix; Feature extraction is performed on the dense matrix to obtain the character-level feature matrix.

[0011] According to a website registration method provided by the present application, the word-level feature matrix corresponding to the hierarchical structure of the node is determined based on the character string set, including: In the string set consisting of the strings corresponding to the nodes, each string is segmented to obtain a segmentation matrix, and the lengths of all segmentation results in the segmentation matrix are the same; Performing dimensionality reduction processing on the website word vector matrix corresponding to the word segmentation matrix to obtain a low-dimensional word vector matrix; Feature extraction is performed on the low-dimensional word vector matrix to obtain the word-level feature matrix.

[0012] According to a website registration method provided by the present application, the target registration information is modified based on the target prompt type and the target registration specification, including: Convert the label sequences of all nodes in the website into corresponding index vectors, and fuse and concatenate the index vectors with the HTML text to obtain a fused sequence; Inputting the fused sequence into a large language model to obtain a correction strategy for the target registration information; The target registration information is modified based on the target prompt type, the target registration specification and the modification strategy.

[0013] According to a website registration method provided by the present application, the type of each of the multiple information input boxes on the registration webpage of the website, and adding registration information in the multiple information input boxes respectively, include: For each of the information input boxes, in response to a click operation on the information input box, in the case of outputting a modal dialog box, determining a target solution; The modal dialog box is processed based on the target solution, and the registration information is added to the information input box when the modal dialog box has been processed.

[0014] The present application also provides a website registration device, which is applied to an electronic device, and the device includes: A first processing unit, configured to add registration information to each of the plurality of information input boxes on the registration webpage of the website based on the respective types of the plurality of information input boxes; A sending unit, used to send each of the registration information to a website server; A receiving unit, configured to receive registration prompt information sent by the website server, wherein the registration prompt information is determined based on the registration information and the registration specification corresponding to each information input box; A second processing unit is used for inputting the registration prompt information into an attribute fusion deep learning model to obtain a target prompt type and a target registration specification, when the registration prompt information is used to represent that there is at least one registration information that is inconsistent with a corresponding registration specification, wherein the target prompt type is a type corresponding to the target registration information that is inconsistent with the registration specification, and the target registration specification is a registration specification corresponding to the target registration information; The registration unit is used to modify the target registration information based on the target prompt type and the target registration specification, and perform website registration based on the registration information consistent with the registration specification and the modified target registration information.

[0015] The present application also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the website registration method as described above is implemented.

[0016] The present application also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the website registration method as described in any one of the above is implemented.

[0017] The present application also provides a computer program product, including a computer program, wherein when the computer program is executed by a processor, the website registration method as described in any one of the above is implemented.

[0018] The website registration method, device, electronic device and storage medium provided by the present application can first add registration information to multiple information input boxes based on the types of each of the multiple information input boxes on the registration webpage of the website, and send each registration information to the website server; and receive the registration prompt information sent by the website server; when the registration prompt information is used to characterize that there is at least one registration information inconsistent with the corresponding registration specification, the registration prompt information is input into the attribute fusion deep learning model to obtain the target prompt type and the target registration specification, and then based on the target prompt type and the target registration specification, the target registration information is corrected, and the website registration is performed based on the registration information consistent with the registration specification and the corrected target registration information. In this way, with the help of the attribute fusion deep learning model, it is possible to automatically correct the target registration information based on the target prompt type and the target registration specification when there is at least one registration information inconsistent with the corresponding registration specification, which not only solves the problem that users need to manually fill in a large amount of registration information in the prior art, resulting in low registration efficiency, but also solves the problem of process interruption due to unsuccessful registration prompts on the website, thereby effectively improving registration efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present application or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 A flowchart of a website registration method provided in an embodiment of the present application.

[0021] Figure 2 A schematic diagram of a processing framework for causing a modal dialog box to appear when an information input box is clicked is provided in an embodiment of the present application.

[0022] Figure 3 A flowchart for obtaining a target prompt type and a target registration specification is provided in an embodiment of the present application.

[0023] Figure 4 A schematic diagram of the structure of an attribute fusion deep learning model provided in an embodiment of the present application.

[0024] Figure 5 A schematic diagram of the structure of a website registration device provided in an embodiment of the present application.

[0025] Figure 6 A schematic diagram of the physical structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0027] In the embodiments of the present application, "at least one" means one or more, and "more" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. In the text description of the present application, the character " / " generally indicates that the associated objects before and after are in an "or" relationship.

[0028] The technical solution provided by the embodiment of the present application can be adapted to the website registration scenario. In order to use the services provided by the website, users need to register on the website, which not only facilitates users to obtain personalized services, but also helps the website manage and maintain user information.

[0029] Traditional website registration methods usually require users to manually fill in a large amount of registration information, which results in low registration efficiency.

[0030] In order to realize automatic registration of a website and thus improve registration efficiency, an embodiment of the present application provides a website registration method. Below, the website registration method provided by the present application will be described in detail through the following specific embodiments. It can be understood that the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0031] Figure 1 A schematic diagram of a website registration method provided in an embodiment of the present application, which is applied to an electronic device, for example, see Figure 1 As shown, the website registration method may include: S101 . Based on the types of the multiple information input boxes on the registration webpage of the website, registration information is added to the multiple information input boxes respectively, and each registration information is sent to the website server.

[0032] For example, multiple information input boxes may include a user name input box, a password input box, a verification code input box or other input boxes, etc.; correspondingly, the type of information input box may include a user name prompt type, a password prompt type, a verification code prompt type or other prompt types, etc., which can be set according to actual needs.

[0033] In order to realize automatic registration on the website, the electronic device can add registration information in multiple information input boxes respectively based on the types of the multiple information input boxes. For example, when the type of the information input box is a username prompt type, the user can enter the username to be registered in the information input box; when the type of the information input box is a password prompt type, the user can enter the password in the information input box; when the type of the information input box is a verification code prompt type, the user can enter the verification code in the information input box.

[0034] For example, in an embodiment of the present application, when an electronic device enters a username and password required for registration, the username and password can be automatically generated based on a pre-built user information database, which is usually built by analyzing registration prompts of multiple websites to meet the registration requirements of most websites.

[0035] After adding registration information in multiple information input boxes respectively, the added registration information can be sent to the registration server, so that the registration server determines whether it meets the corresponding registration specifications based on the received registration information, and sends registration prompt information to the electronic device based on the judgment result.

[0036] S102: receiving registration prompt information sent by the website server, where the registration prompt information is determined based on the registration specifications corresponding to each registration information and each information input box.

[0037] For example, the registration prompt information can be used to indicate that the registration is successful, that is, each piece of registration information is consistent with the corresponding registration specification and meets the registration requirements; or, the registration prompt information can also be used to indicate that there is at least one piece of registration information that is inconsistent with the corresponding registration specification among multiple pieces of registration information. In this case, the electronic device can determine the target prompt type and the target registration specification by means of the attribute fusion deep learning model, that is, execute the following S103: S103: When the registration prompt information is used to represent that there is at least one registration information that is inconsistent with the corresponding registration specification, the registration prompt information is input into the attribute fusion deep learning model to obtain a target prompt type and a target registration specification.

[0038] The target prompt type is a type corresponding to the target registration information that is inconsistent with the registration specification, and the target registration specification is a registration specification corresponding to the target registration information.

[0039] For example, in an embodiment of the present application, the target prompt type may be a user name prompt type, a password prompt type, a verification code prompt type, or other prompt types, etc., which may be specifically set according to actual needs.

[0040] For example, the attribute fusion deep learning model may be a model based on deep learning, which may be specifically set according to actual needs.

[0041] For example, if the target prompt type output by the attribute fusion deep learning model is a username prompt type, it means that the target registration information inconsistent with the registration specification is the username, and the corresponding target registration specification is the registration specification corresponding to the username information input box, for example, a string composed of letters and numbers; if the target prompt type output by the attribute fusion deep learning model is a password prompt type, it means that the target registration information inconsistent with the registration specification is a password, and the corresponding target registration specification is the registration specification corresponding to the password information input box, for example, a string composed of uppercase letters, lowercase letters and numbers with a length of not less than 8; if the target prompt type output by the attribute fusion deep learning model is a verification code prompt type, it means that the target registration information inconsistent with the registration specification is a verification code, and the corresponding target registration specification is the registration specification corresponding to the verification code information input box, for example, a string composed of 6 digits.

[0042] After the target prompt type and target registration specification are determined with the help of the attribute fusion deep learning model, the target registration information that is inconsistent with the registration specification can be corrected in a targeted manner based on the target prompt type and the target registration specification, that is, the following S104 is executed to realize automatic registration of the website, which solves the problem in the prior art that users need to manually fill in a large amount of registration information, resulting in low registration efficiency, thereby effectively improving the registration efficiency.

[0043] S104: based on the target prompt type and the target registration specification, modify the target registration information, and perform website registration based on the registration information consistent with the registration specification and the modified target registration information.

[0044] For example, if the target prompt type output by the attribute fusion deep learning model is a username prompt type, then based on the target prompt type, the target registration information that is inconsistent with the registration specification can be determined as the username, and combined with its corresponding username registration specification, for example, a string of letters and numbers, the added username is corrected to obtain the corrected username, and then combined with the registration information consistent with the registration specification, such as a password and verification code, to jointly register on the website.

[0045] For example, if the target prompt type output by the attribute fusion deep learning model is a password prompt type, the target registration information that is inconsistent with the registration specifications can be determined as a password based on the target prompt type, and combined with the registration specifications corresponding to the corresponding password information input box, for example, a character string composed of uppercase letters, lowercase letters and numbers with a length of not less than 8, the added password is corrected to obtain the corrected password, and then combined with the registration information consistent with the registration specifications, such as the user name and verification code, to jointly register on the website.

[0046] It can be seen that in the embodiment of the present application, when registering on the website, the registration information can be added to the multiple information input boxes based on the types of the multiple information input boxes on the registration webpage of the website, and each registration information is sent to the website server; and the registration prompt information sent by the website server is received; when the registration prompt information is used to characterize that there is at least one registration information that is inconsistent with the corresponding registration specification, the registration prompt information is input into the attribute fusion deep learning model to obtain the target prompt type and the target registration specification, and then based on the target prompt type and the target registration specification, the target registration information is corrected, and the website registration is performed based on the registration information consistent with the registration specification and the corrected target registration information. In this way, with the help of the attribute fusion deep learning model, it is possible to automatically correct the target registration information based on the target prompt type and the target registration specification when there is at least one registration information that is inconsistent with the corresponding registration specification, which not only solves the problem in the prior art that users need to manually fill in a large amount of registration information, resulting in low registration efficiency, but also solves the problem of process interruption due to unsuccessful registration prompts on the website, thereby effectively improving registration efficiency.

[0047] Based on the above Figure 1 In the embodiment shown, for example, in the above S101, based on the respective types of multiple information input boxes of the registration webpage of the website, when adding registration information to the multiple information input boxes respectively, it is considered that many websites (including illegal websites) usually display activities and discounts on the webpage to guide new users to register and encourage existing users to continue to participate. This information generally appears in the form of a modal dialog box. In view of the forced interaction feature of this modal dialogue, in the embodiment of the present application, a modal exception automatic correction branch process can be added, specifically including: for each information input box, in response to a click operation on the information input box, when a modal dialog box is output, a target solution is determined; the modal dialog box is processed based on the target solution, and when the modal dialog box has been processed, the registration information is added to the information input box, so that it can be automatically started without human intervention, which can minimize the impact of exceptions on the website automation process.

[0048] For example, the target solution may be an exception capture and processing solution for elements outside the information input box, which may be specifically set according to actual needs.

[0049] For example, see Figure 2 As shown, Figure 2 A schematic diagram of a processing framework for a modal dialog box that appears when a click operation is performed on an information input box is provided for an embodiment of the present application. When a click operation is performed on the information input box, it can be first determined whether the click operation is successful; in the case of a successful click, by comparing the Document Object Model (DOM) structure before and after the click operation, it is determined whether the registration page has changed to determine whether the click operation is effective. Among them, DOM displays HTML documents in a tree structure, including elements, attributes, and text nodes, reflecting all page changes. Through depth-first traversal, the HTML text is parsed and all nodes are saved in the form of [node label, node object, [child node list]]. Through breadth-first traversal, the changed nodes can be compared and identified level by level to determine whether there are new additions or deletions. If the nodes are added or deleted, it is determined that the registration page has changed. When the registration page changes, it indicates that the click operation is implemented; when the click is unsuccessful, it is determined that a modal dialog box is output to interrupt the website's automated registration process, and there is feedback. The feedback is further analyzed and the target solution is automatically determined. When the target solution successfully handles the modal dialog box, the information input box is clicked again, and the click operation is implemented; when the target solution successfully handles the modal dialog box, the entire page is analyzed and the target solution is re-determined, and the information input box is clicked again, and the click operation is implemented; when the registration page does not change, it is determined that a modal dialog box is output to interrupt the website's automated registration process, and there is no feedback. The entire page is further analyzed and the target solution is re-determined, and the information input box is clicked again, and the click operation is implemented. In this way, by adding modal exceptions to automatically correct branch processes, the problem of forcing users to make some form of response and being unable to perform other operations before responding, which leads to process interruption, is solved. It can start automatically without human intervention, which can minimize the impact of exceptions on the website's automated process.

[0050] For example, in the embodiments of the present application, the click operation identifier used to trigger a specific function on the registration page can be defined as a "button", and the buttons described below all belong to this category. Buttons can be divided into three categories: image buttons, whose functions are triggered by clicking on image elements in the web page; non-image buttons, which work in a similar way to the former; and submit buttons, which have various presentation methods, but all realize the form submission function, which is different from the above two types of buttons.

[0051] For example, the image button retrieval logic follows a three-stage process: Step 1, execute The first step is to filter the tag elements; the second step is to extract the source file path and image class name and filter based on keywords; the third step is to parse and output the corresponding path of the determined element.

[0052] Based on the above Figure 1 In the embodiment shown, for example, in the embodiment of the present application, the registration prompt information may include HyperText Markup Language (HTML) text, style attributes of nodes in the website, and the hierarchical structure of the nodes.

[0053] For example, the style attributes of a node may include the location of the node, the size of the graphic, the pop-up window display mode, etc.; the hierarchical structure of a node may include the labels and attribute values ​​of all child nodes in the node, as well as the tree structure of the entire node.

[0054] In the case where the registration prompt information may include HTML text, style attributes of nodes in the website, and the hierarchical structure of nodes, in the above S103, the registration prompt information is input into the attribute fusion deep learning model to obtain the target prompt type and the target registration specification. For details, see the following Figure 2 The embodiment shown.

[0055] Figure 3 A schematic diagram of a process for obtaining a target prompt type and a target registration specification provided in an embodiment of the present application, the method may include: S301, inputting HTML text into the first embedding layer in the attribute fusion deep learning model to obtain a first text feature vector; and inputting the first text feature vector into the activation layer to obtain a target registration specification.

[0056] For example, in the embodiments of the present application, see Figure 4 As shown, Figure 4 A structural schematic diagram of an attribute fusion deep learning model provided in an embodiment of the present application, wherein the attribute fusion deep learning model may include a first embedding layer, a second embedding layer, a third embedding layer, an activation layer, and a post-processing module.

[0057] When obtaining the target registration specification through HTML text, the HTML text can be input into the first embedding layer in the attribute fusion deep learning model, the HTML text can be converted into a first text feature vector through the first embedding layer, and the converted first text feature vector can be input into the activation layer to obtain the target registration specification. In this way, the target registration specification corresponding to the target registration information inconsistent with the registration specification can be obtained based on the HTML text.

[0058] Next, in combination with the following S302-S304, the target prompt type corresponding to the target registration information that is inconsistent with the registration specification can be obtained, so that the target registration information can be automatically corrected based on the target prompt type and the target registration specification. This not only solves the problem in the prior art that users need to manually fill in a large amount of registration information, resulting in low registration efficiency, but also solves the problem of process interruption due to website prompts of unsuccessful registration, thereby effectively improving registration efficiency.

[0059] S302: Input the style attributes of the node into the second embedding layer in the attribute fusion deep learning model to obtain a string feature vector.

[0060] By way of example, in an embodiment of the present application, the style attributes of a node include numerical and categorical types. When the style attributes of the node are input into the second embedding layer in the attribute fusion deep learning model to obtain a string feature vector, the style attributes of the node can be first input into the second embedding layer in the attribute fusion deep learning model. For numerical style attributes, the numerical style attributes are normalized to obtain normalized style attributes; for categorical style attributes, the categorical style attributes are encoded to obtain encoded style attributes; and the normalized style attributes and the encoded style attributes are concatenated to obtain a string feature vector.

[0061] For example, for numerical style attributes, such as node location or size, when normalizing the numerical style attributes, refer to the following formula 1: Formula 1

[0062] in, Represents the normalized style attributes. Represents a numeric style attribute. Represents the minimum style attributes, Represents the maximum style attribute.

[0063] For example, for categorical style attributes, such as display mode, element transformation attributes, etc., the categorical style attributes can be one-hot encoded to obtain encoded style attributes, and the normalized style attributes and the encoded style attributes can be concatenated to obtain a string feature vector.

[0064] Assume that the numeric style attribute can be recorded as , the category-type style attribute can be recorded as ,in, ,and , then concatenate the normalized style attributes and the encoded style attributes to obtain the string feature vector: [ ] -> . Indicates A numeric style attribute. Indicates A class-type style attribute.

[0065] S303. Input the hierarchical structure of the nodes into the third embedding layer in the attribute fusion deep learning model to obtain a string feature matrix.

[0066] By way of example, in an embodiment of the present application, when the hierarchical structure of the nodes is input into the third embedding layer in the attribute fusion deep learning model to obtain a string feature matrix, the hierarchical structure of the nodes can be first input into the third embedding layer in the attribute fusion deep learning model to determine the string corresponding to each node, where the string corresponding to the node includes the label and attribute value of the node; and based on the string set consisting of the strings corresponding to each node, the character-level feature matrix corresponding to the hierarchical structure of the nodes is determined; and based on the string set, the word-level feature matrix corresponding to the hierarchical structure of the nodes is determined; and then the character-level feature matrix and the word-level feature matrix are weightedly concatenated to obtain a string feature matrix.

[0067] Typically, a node hierarchy assigns a unique identifier to each node in its tree structure, and creates a vector for each node, which is a string consisting of the node's label and attribute value. Therefore, the node hierarchy is input into the third embedding layer in the attribute fusion deep learning model to determine the string corresponding to each node. It is assumed that the attribute value of node 1 can be expressed as a string. and The string corresponding to node 1 can be written as ; Assume that the attribute value of node 2 can be used as a string The string corresponding to node 2 can be written as ; Assume that the attribute value of node 3 can be used as a string and The string corresponding to node 3 can be written as .

[0068] After respectively determining the string corresponding to each node, the string corresponding to each node can be converted into a corresponding string set based on the hierarchical structure of the node. For example, in the embodiment of the present application, the string set can be recorded as U.

[0069] By way of example, in an embodiment of the present application, when determining a character-level feature matrix corresponding to a hierarchical structure of a node based on a string set consisting of strings corresponding to each node, it is possible to first perform a word segmentation operation on each string in the string set consisting of strings corresponding to each node to obtain a word segmentation result set; generate a character-level vector corresponding to the string set according to the frequency of occurrence of character elements in the word segmentation result set; adjust the dimensions of all vectors in the character-level vector to obtain a vector set, in which all vectors in the vector set have the same dimensions; encode all vectors in the vector set, and perform dimensionality reduction processing on the encoded matrix to obtain a dense matrix; and perform feature extraction on the dense matrix to obtain a character-level feature matrix.

[0070] For example, when determining the character-level feature matrix based on the string set consisting of the strings corresponding to each node, all tags in the string set U can be segmented by characters first, and the segmentation result set is recorded as V, which represents all the character sets that have appeared in the tag set. The frequency of occurrence of the character elements in the segmentation result set V is assigned with numbers 1, 2, ..., n, and the character-level vector corresponding to the string set is generated to form a character-level vector, where the value of n can be set according to actual needs. Secondly, padding operations are performed on all vectors in the character-level vector, the dimensions of all tag vectors are fixed, and 0 is added to the front of the vector for vectors with insufficient length, and the excess part of the vector is intercepted for vectors with excessive length, and finally a vector set W is formed, as shown in the following formula 2, where the dimensions of all vectors in the vector set W are the same.

[0071] W={w1, w2, w3,..., wi} Formula 2

[0072] Among them, w1 represents the first vector and wi represents the i-th vector.

[0073] In addition, all vectors in the vector set are one-hot encoded to form a one-hot matrix G. In order to construct a dense low-dimensional feature matrix, the matrix G is reduced in dimension to obtain a dense matrix S, as shown in the following formula 3: Formula 3

[0074] in, is a dimension reduction matrix used to Dimensional matrix Perform dimensionality reduction to reduce its dimension to dimension to obtain a dense matrix .

[0075] The dense matrix S is subjected to feature extraction through one-dimensional convolution of TextCNN to obtain the character-level feature matrix CV, as shown in the following formula 4: CV = {cv1, cv2, cv3, ..., cvi} Formula 4

[0076] Among them, cv1 represents the first character-level feature, and cvi represents the i-th character-level feature.

[0077] By way of example, in an embodiment of the present application, when determining a word-level feature matrix corresponding to a hierarchical structure of a node based on a string set, each string in the string set consisting of the strings corresponding to each node can be first segmented to obtain a segmentation matrix, in which all segmentation results in the segmentation matrix have the same length; a dimensionality reduction process is performed on a website word vector matrix corresponding to the segmentation matrix to obtain a low-dimensional word vector matrix; and feature extraction is performed on the low-dimensional word vector matrix to obtain a word-level feature matrix.

[0078] For example, when determining the word-level feature matrix corresponding to the hierarchical structure of the node based on the string set, the strings in the string set U can be first segmented according to the corpus of the BERT pre-trained model, and a maximum segmentation length is set. If the length exceeds the maximum length, the excess part is directly deleted, and if the length is less than the maximum length, it is automatically supplemented with 0 when it is subsequently input into the BERT model to obtain a segmentation matrix T.

[0079] T={t1, t2, t3,..., ti} Formula 5

[0080] Among them, t1 represents the first participle, and ti represents the i-th participle.

[0081] Secondly, each row ti in the word segmentation matrix T is converted into an idSequence to obtain an ids matrix, and then 12 hidden layers are obtained through the pre-trained BERT model. The last four rows are summed to obtain a 768-dimensional website word vector matrix Word for prompt information, as shown in the following formula 6: Word={word1, word2, word3,..., wordi} Formula 6

[0082] Among them, word1 represents the first website word vector, and wordi represents the i-th website word vector.

[0083] Then, a fully connected layer is used to reduce the dimensionality of the high-dimensional website word vector matrix Word to obtain a low-dimensional word vector matrix R, as shown in the following formula 7: Formula 7

[0084] in, Represents the high-dimensional website word vector matrix Word, is a dimension reduction matrix used to The high-dimensional website word vector matrix Word is reduced to dimension to obtain a low-dimensional word vector matrix R.

[0085] The low-dimensional word vector matrix R is subjected to feature extraction through the TextCNN one-dimensional convolution to obtain the word-level feature matrix WV, as shown in the following formula 8: WV={wv1, wv2, wv3, ..., wvi} Formula 8

[0086] Among them, wv1 represents the first word-level feature, and wvi represents the i-th word-level feature.

[0087] Then, the character-level feature matrix CV and the word-level feature matrix WV are weightedly concatenated to obtain the string feature matrix.

[0088] In combination with the above description, after respectively obtaining the first text feature vector, the character string feature vector and the character string feature matrix, the following S304 may be executed: S304: Input the first text feature vector, the character string feature vector and the character string feature matrix into a post-processing module in the attribute fusion deep learning model to obtain a target prompt type.

[0089] It should be noted that, in the embodiment of the present application, there is no order between the operation of inputting the first text feature vector into the activation layer to obtain the target registration specification in the above S301 and the operation of obtaining the target prompt type in the above S302-S304. Here, the embodiment of the present application only uses the example of inputting the first text feature vector into the activation layer to obtain the target registration specification, and then executing the above S202-S304 to obtain the target prompt type, but it does not mean that the embodiment of the present application is limited to this.

[0090] See above Figure 4 As shown, for example, in an embodiment of the present application, the post-processing module may include a convolutional layer, a maximum pooling layer, a regularization layer, and an activation layer connected in series in sequence, which can be specifically configured according to actual needs.

[0091] When the first text feature vector, string feature vector and string feature matrix are input into the post-processing module in the attribute fusion deep learning model, the first text feature vector, string feature vector and string feature matrix can be firstly subjected to global pooling and full connection layer feature by feature; secondly, the activation function is connected to generate feature weight scores to achieve the effect of feature by feature weighting, and then the global self-attention module is performed to obtain better global information. Then, the local pattern in the vector is captured by the convolution layer, the dimension of the feature vector is reduced by the pooling layer, and the dropout layer is used to avoid overfitting of the model; finally, the target prompt type is output.

[0092] When there is at least one registration information that is inconsistent with the corresponding registration specification, after determining the target prompt type and the target registration specification with the help of the attribute fusion deep learning model, the target registration information can be corrected based on the target prompt type and the target registration specification, that is, executing the above S104.

[0093] For example, in the above S104, when the target registration information is corrected based on the target prompt type and the target registration specification, the label sequences of all nodes in the website can be first converted into corresponding index vectors, and the index vectors are fused and spliced ​​with the HTML text to obtain a fused sequence; the fused sequence is input into the large language model to obtain a correction strategy for the target registration information; and then the target registration information is corrected based on the target prompt type, the target registration specification and the correction strategy. This not only solves the problem in the prior art that users need to manually fill in a large amount of registration information, resulting in low registration efficiency, but also solves the problem of process interruption due to unsuccessful registration prompts from the website, thereby effectively improving registration efficiency.

[0094] For example, when converting the label sequences of all nodes in the website into corresponding index vectors, a label dictionary for registration information prompts can be pre-built, in which all information prompt related tags and five special identifiers [SEP], [LEFT], [RIGHT], [UNK], [PAD] in the current HTML specification are screened, totaling 64. In this way, the label dictionary can be combined to convert and map the label sequences of all nodes into corresponding index vectors, and then the index vectors are fused and spliced ​​with the HTML text to obtain a fused sequence, which is ready to be input into a large language model to output a correction strategy for the target registration information through the large language model.

[0095] For example, in the embodiment of the present application, in order to enable the large language model to fully understand the vertical scenario of website registration, a prompt paradigm enhancement method generated by a combination of multiple concepts can be used to replace the prompt type with multiple groups of conceptual words or phrases with the same meaning. For example, "character length" is replaced with "password length", "user name length", "within characters", etc., to enrich the semantics of the target, so that the pre-trained large language model can learn more relevant knowledge. Given a prompt template and type word set V (character length, special characters, uppercase letters, numbers, occupied, spaces, etc.), for each node's fusion sequence and concept word combination {concept1, concept2, ..., concept𝑘}, where k represents the kth concept word. First, use the given prompt template to merge the node's fusion sequence Combine with each concept word and map it into prompt input ; Each prompt input is then fed into a separate large language model In the example, we get the hidden layer representation of [MASK] , and then the probability of the registration prompt type at [MASK] is:

[0096] in, express exist The probability of yes In the pre-trained large language model After calculating the mask prediction probability, the softmax function is used to normalize the probability.

[0097] Finally, the activation layer of the large language model outputs the correction strategy for the target registration information. The probability of label words obtained by each concept word through prompt learning is different for each concept word, and the knowledge learned is also different. Therefore, they need to be fused to obtain the overall type probability:

[0098] Through the above-mentioned large language model, the correction strategy of the target registration information, that is, the adjustment direction of the target registration information, can be determined, so that on the basis of the target prompt type and the target registration specification, the target registration information can be corrected in combination with the correction strategy of the target registration information, and registration can be retried until the registration is successful. This not only solves the problem in the prior art that users need to manually fill in a large amount of registration information, resulting in low registration efficiency, but also solves the problem of process interruption due to website prompts of unsuccessful registration, thereby effectively improving registration efficiency.

[0099] The website registration device provided by the present application is described below. The website registration device described below and the website registration method described above can be referenced to each other.

[0100] Figure 5 A schematic diagram of a website registration device provided in an embodiment of the present application, which is applied to an electronic device, for example, see Figure 5 As shown, the website registration device 50 may include: The first processing unit 501 is used to add registration information to each of the multiple information input boxes on the registration webpage of the website based on the types of the multiple information input boxes. A sending unit 502, configured to send each of the registration information to a website server; The receiving unit 503 is used to receive the registration prompt information sent by the website server, where the registration prompt information is determined based on the registration information and the registration specification corresponding to each information input box; The second processing unit 504 is used to input the registration prompt information into the attribute fusion deep learning model to obtain a target prompt type and a target registration specification, when the registration prompt information is used to represent that there is at least one registration information that is inconsistent with the corresponding registration specification, the target prompt type is a type corresponding to the target registration information that is inconsistent with the registration specification, and the target registration specification is a registration specification corresponding to the target registration information; The registration unit 505 is used to modify the target registration information based on the target prompt type and the target registration specification, and perform website registration based on the registration information consistent with the registration specification and the modified target registration information.

[0101] For example, in an embodiment of the present application, the registration prompt information includes HTML text, style attributes of nodes in the website, and the hierarchical structure of the nodes; The second processing unit 504 is used to input the registration prompt information into the attribute fusion deep learning model to obtain the target prompt type and the target registration specification corresponding to the target registration information inconsistent with the registration specification, including: Inputting the HTML text into a first embedding layer in the attribute fusion deep learning model to obtain a first text feature vector; and inputting the first text feature vector into an activation layer to obtain the target registration specification; Inputting the style attribute of the node into the second embedding layer in the attribute fusion deep learning model to obtain a string feature vector; Inputting the hierarchical structure of the nodes into the third embedding layer in the attribute fusion deep learning model to obtain a string feature matrix; The first text feature vector, the character string feature vector and the character string feature matrix are input into a post-processing module in the attribute fusion deep learning model to obtain the target prompt type.

[0102] For example, in the embodiment of the present application, the style attribute of the node includes a numerical type and a category type; The second processing unit 504 is used to input the style attribute of the node into the second embedding layer in the attribute fusion deep learning model to obtain a string feature vector, including: Inputting the style attributes of the node into the second embedding layer in the attribute fusion deep learning model, and for numerical style attributes, normalizing the numerical style attributes to obtain normalized style attributes; For a category-type style attribute, encoding the category-type style attribute to obtain an encoded style attribute; The normalized style attribute and the encoded style attribute are concatenated to obtain the character string feature vector.

[0103] For example, in the embodiment of the present application, the second processing unit 504 is used to input the hierarchical structure of the node into the third embedding layer in the attribute fusion deep learning model to obtain a string feature matrix, including: Inputting the hierarchical structure of the nodes into the third embedding layer in the attribute fusion deep learning model to determine the character string corresponding to each node; Based on the string set consisting of the strings corresponding to the nodes, determining the character-level feature matrix corresponding to the hierarchical structure of the nodes; and determining the word-level feature matrix corresponding to the hierarchical structure of the nodes based on the string set; The character-level feature matrix and the word-level feature matrix are weightedly concatenated to obtain the string feature matrix.

[0104] For example, in the embodiment of the present application, the second processing unit 504 is used to determine the character-level feature matrix corresponding to the hierarchical structure of the node based on the string set consisting of the strings corresponding to the nodes, including: Performing a word segmentation operation on each string in a string set consisting of strings corresponding to each node to obtain a word segmentation result set; Generate a character-level vector corresponding to the string set according to the frequency of occurrence of the character elements in the word segmentation result set; Adjusting the dimensions of all vectors in the character-level vector to obtain a vector set, wherein all vectors in the vector set have the same dimension; Encoding all vectors in the vector set, and performing dimensionality reduction processing on the encoded matrix to obtain a dense matrix; Feature extraction is performed on the dense matrix to obtain the character-level feature matrix.

[0105] For example, in the embodiment of the present application, the second processing unit 504 is used to determine the word-level feature matrix corresponding to the hierarchical structure of the node based on the character string set, including: In the string set consisting of the strings corresponding to the nodes, each string is segmented to obtain a segmentation matrix, and the lengths of all segmentation results in the segmentation matrix are the same; Performing dimensionality reduction processing on the website word vector matrix corresponding to the word segmentation matrix to obtain a low-dimensional word vector matrix; Feature extraction is performed on the low-dimensional word vector matrix to obtain the word-level feature matrix.

[0106] For example, in the embodiment of the present application, the registration unit 505 is used to modify the target registration information based on the target prompt type and the target registration specification, including: Convert the label sequences of all nodes in the website into corresponding index vectors, and fuse and concatenate the index vectors with the HTML text to obtain a fused sequence; Inputting the fused sequence into a large language model to obtain a correction strategy for the target registration information; The target registration information is modified based on the target prompt type, the target registration specification and the modification strategy.

[0107] For example, in the embodiment of the present application, the first processing unit 501 is used to add registration information to each of the multiple information input boxes based on the types of the multiple information input boxes on the registration webpage of the website, including: For each of the information input boxes, in response to a click operation on the information input box, in the case of outputting a modal dialog box, determining a target solution; The modal dialog box is processed based on the target solution, and the registration information is added to the information input box when the modal dialog box has been processed.

[0108] The website registration device 50 provided in the embodiment of the present application can execute the technical solution of the website registration method in any of the above embodiments. Its implementation principle and beneficial effects are similar to those of the website registration method. Please refer to the implementation principle and beneficial effects of the website registration method, and no further details will be given here.

[0109] Figure 6 A schematic diagram of the physical structure of an electronic device provided in an embodiment of the present application, such as Figure 6As shown, the electronic device may include: a processor (processor) 610 , a communication interface (Communications Interface) 620 , a memory (memory) 630 and a communication bus 640 , wherein the processor 610 , the communication interface 620 , and the memory 630 communicate with each other through the communication bus 640 . The processor 610 can call the logic instructions in the memory 630 to execute the website registration method, which includes: based on the respective types of multiple information input boxes on the registration webpage of the website, adding registration information in the multiple information input boxes respectively, and sending each of the registration information to the website server; receiving registration prompt information sent by the website server, the registration prompt information is determined based on the registration specifications corresponding to each of the registration information and each of the information input boxes; when the registration prompt information is used to represent that there is at least one of the registration information inconsistent with the corresponding registration specification, inputting the registration prompt information into the attribute fusion deep learning model to obtain a target prompt type and a target registration specification, the target prompt type is a type corresponding to the target registration information inconsistent with the registration specification, and the target registration specification is a registration specification corresponding to the target registration information; based on the target prompt type and the target registration specification, the target registration information is corrected, and the website registration is performed based on the registration information consistent with the registration specification and the corrected target registration information.

[0110] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0111] On the other hand, the present application also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the website registration method provided by the above methods, which method includes: based on the types of multiple information input boxes on the registration webpage of the website, adding registration information in the multiple information input boxes respectively, and sending each of the registration information to the website server; receiving registration prompt information sent by the website server, wherein the registration prompt information is determined based on the registration specifications corresponding to each of the registration information and each of the information input boxes; when the registration prompt information is used to represent that there is at least one registration information that is inconsistent with the corresponding registration specification, the registration prompt information is input into the attribute fusion deep learning model to obtain a target prompt type and a target registration specification, wherein the target prompt type is a type corresponding to the target registration information that is inconsistent with the registration specification, and the target registration specification is a registration specification corresponding to the target registration information; based on the target prompt type and the target registration specification, the target registration information is corrected, and the website registration is performed based on the registration information consistent with the registration specification and the corrected target registration information.

[0112] On the other hand, the present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the website registration method provided by the above-mentioned methods, the method comprising: based on the types of multiple information input boxes of the registration webpage of the website, respectively adding registration information in the multiple information input boxes, and sending each of the registration information to the website server; receiving registration prompt information sent by the website server, the registration prompt information being determined based on the registration specifications corresponding to each of the registration information and each of the information input boxes; in a case where the registration prompt information is used to characterize that there is at least one registration information that is inconsistent with the corresponding registration specification, inputting the registration prompt information into an attribute fusion deep learning model to obtain a target prompt type and a target registration specification, the target prompt type being a type corresponding to the target registration information that is inconsistent with the registration specification, and the target registration specification being a registration specification corresponding to the target registration information; based on the target prompt type and the target registration specification, correcting the target registration information, and performing website registration based on the registration information consistent with the registration specification and the corrected target registration information.

[0113] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0114] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0115] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A website registration method, characterized in that: Applied to electronic equipment, the method comprises: Based on the types of the multiple information input boxes on the registration webpage of the website, respectively add registration information in the multiple information input boxes, and send each of the registration information to the website server; Receiving registration prompt information sent by the website server, wherein the registration prompt information is determined based on the registration information and the registration specification corresponding to each information input box; In a case where the registration prompt information is used to represent that at least one of the registration information is inconsistent with the corresponding registration specification, the registration prompt information is input into the attribute fusion deep learning model to obtain a target prompt type and a target registration specification, wherein the target prompt type is a type corresponding to the target registration information inconsistent with the registration specification, and the target registration specification is a registration specification corresponding to the target registration information; Based on the target prompt type and the target registration specification, the target registration information is modified, and website registration is performed based on the registration information consistent with the registration specification and the modified target registration information.

2. The website registration method according to claim 1, characterized in that: The registration prompt information includes HTML text, style attributes of nodes in the website and the hierarchical structure of the nodes; The step of inputting the registration prompt information into the attribute fusion deep learning model to obtain the target prompt type and the target registration specification includes: Inputting the HTML text into a first embedding layer in the attribute fusion deep learning model to obtain a first text feature vector; and inputting the first text feature vector into an activation layer to obtain the target registration specification; Inputting the style attribute of the node into the second embedding layer in the attribute fusion deep learning model to obtain a string feature vector; Inputting the hierarchical structure of the nodes into the third embedding layer in the attribute fusion deep learning model to obtain a string feature matrix; The first text feature vector, the character string feature vector and the character string feature matrix are input into a post-processing module in the attribute fusion deep learning model to obtain the target prompt type.

3. The website registration method according to claim 2, characterized in that: The style attributes of the node include numerical type and category type; The step of inputting the style attribute of the node into the second embedding layer in the attribute fusion deep learning model to obtain a string feature vector includes: Inputting the style attributes of the node into the second embedding layer in the attribute fusion deep learning model, and for numerical style attributes, normalizing the numerical style attributes to obtain normalized style attributes; For a category-type style attribute, encoding the category-type style attribute to obtain an encoded style attribute; The normalized style attribute and the encoded style attribute are concatenated to obtain the character string feature vector.

4. The website registration method according to claim 2, characterized in that: The step of inputting the hierarchical structure of the nodes into the third embedding layer in the attribute fusion deep learning model to obtain a string feature matrix includes: Inputting the hierarchical structure of the nodes into the third embedding layer in the attribute fusion deep learning model to determine the character string corresponding to each node; Based on the string set consisting of the strings corresponding to the nodes, determining the character-level feature matrix corresponding to the hierarchical structure of the nodes; and determining the word-level feature matrix corresponding to the hierarchical structure of the nodes based on the string set; The character-level feature matrix and the word-level feature matrix are weightedly concatenated to obtain the string feature matrix.

5. The website registration method according to claim 4, characterized in that: The step of determining a character-level feature matrix corresponding to the hierarchical structure of the nodes based on a set of character strings corresponding to the nodes includes: Performing a word segmentation operation on each string in a string set consisting of strings corresponding to each node to obtain a word segmentation result set; Generate a character-level vector corresponding to the string set according to the frequency of occurrence of the character elements in the word segmentation result set; Adjusting the dimensions of all vectors in the character-level vector to obtain a vector set, wherein all vectors in the vector set have the same dimension; Encoding all vectors in the vector set, and performing dimensionality reduction processing on the encoded matrix to obtain a dense matrix; Feature extraction is performed on the dense matrix to obtain the character-level feature matrix.

6. The website registration method according to claim 4, characterized in that: The step of determining the word-level feature matrix corresponding to the hierarchical structure of the nodes based on the character string set includes: In the string set consisting of the strings corresponding to the nodes, each string is segmented to obtain a segmentation matrix, and the lengths of all segmentation results in the segmentation matrix are the same; Performing dimensionality reduction processing on the website word vector matrix corresponding to the word segmentation matrix to obtain a low-dimensional word vector matrix; Feature extraction is performed on the low-dimensional word vector matrix to obtain the word-level feature matrix.

7. The website registration method according to any one of claims 2 to 6, characterized in that: The modifying the target registration information based on the target prompt type and the target registration specification includes: Convert the label sequences of all nodes in the website into corresponding index vectors, and fuse and concatenate the index vectors with the HTML text to obtain a fused sequence; Inputting the fused sequence into a large language model to obtain a correction strategy for the target registration information; The target registration information is modified based on the target prompt type, the target registration specification and the modification strategy.

8. The website registration method according to any one of claims 1 to 6, characterized in that: The types of the multiple information input boxes of the registration webpage of the website, respectively adding registration information in the multiple information input boxes, include: For each of the information input boxes, in response to a click operation on the information input box, in the case of outputting a modal dialog box, determining a target solution; The modal dialog box is processed based on the target solution, and the registration information is added to the information input box when the modal dialog box has been processed.

9. A website registration device, characterized in that: Applied to electronic equipment, the device comprises: A first processing unit, configured to add registration information to each of the plurality of information input boxes on the registration webpage of the website based on the respective types of the plurality of information input boxes; A sending unit, used to send each of the registration information to a website server; A receiving unit, configured to receive registration prompt information sent by the website server, wherein the registration prompt information is determined based on the registration information and the registration specification corresponding to each information input box; A second processing unit is used for inputting the registration prompt information into an attribute fusion deep learning model to obtain a target prompt type and a target registration specification, when the registration prompt information is used to represent that there is at least one registration information that is inconsistent with a corresponding registration specification, wherein the target prompt type is a type corresponding to the target registration information that is inconsistent with the registration specification, and the target registration specification is a registration specification corresponding to the target registration information; The registration unit is used to modify the target registration information based on the target prompt type and the target registration specification, and perform website registration based on the registration information consistent with the registration specification and the modified target registration information.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the website registration method according to any one of claims 1 to 8 is implemented.

11. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the website registration method according to any one of claims 1 to 8 is implemented.

12. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the website registration method according to any one of claims 1 to 8 is implemented.

Citation Information

Patent Citations

  • Dialogue communication type account registration method and device

    CN109784017A

  • Equipment registration method and device based on object model automatic matching

    CN112383891A

  • Application registration method and device and electronic equipment

    CN116974647A

  • A member registration method for network system

    CN1972307A

  • Method for inputting diesel power generation

    KR1020210123694A

Cited By

  • Webpage form control intelligent filling method and system based on page change monitoring

    CN120995997A

  • Website registration method and apparatus, electronic device, and storage medium

    WO2026051970A1