Front-end form data verification method and device

By building a form data knowledge graph and verification model, the problem of preset rules limitations in front-end form data verification is solved, and more efficient and flexible data verification is achieved, adapting to the Internet's efficient and flexible needs for front-end forms.

CN118535565BActive Publication Date: 2024-11-12紫金诚征信有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411010593.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-26
Publication Date
2024-11-12
Estimated Expiration
2044-07-26

AI Technical Summary

Technical Problem

In front-end form data verification, due to preset rules limitations, the complexity and diversity of user input cannot be fully captured, resulting in data verification being inflexible enough to meet the Internet's requirements for efficiency and flexibility.

Method used

By obtaining historical form data, determining the relationship between entities and entities, building a graph structure for graph convolution operations, optimizing graph neural network, generating a form data knowledge graph, and building a data verification model based on this, receiving the form data input by the user for verification. Update data to meet verification rules.

Benefits of technology

Improves the efficiency and flexibility of data verification of front-end forms, reduces the time cost of manually updating code, and enhances the ability to capture complex data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118535565B_ABST
    Figure CN118535565B_ABST
Patent Text Reader

Abstract

An embodiment of the present application provides a front-end form data verification method and device, the method comprising: determining entities and entity relationships corresponding to historical form data, inputting a graph structure constructed based on the entities and entity relationships into a convolutional layer of a preset graph neural network for graph convolution operations, determining a corresponding optimal graph neural network, constructing a corresponding form data knowledge graph based on graph features output by the optimal graph neural network, extracting form features in the form data knowledge graph, determining a corresponding data verification model, receiving form data input by a user, and determining whether the form data complies with rules in the preset rule engine based on a preset rule engine. If not, inputting the form data into a data verification model, determining a corresponding verification result, determining updated form data based on the verification result, and performing form page rendering operations based on the updated form data. The present application can improve the efficiency and flexibility of front-end forms based on enhanced data verification technology.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of front-end technology, and in particular to a front-end form data verification method and device. Background Art

[0002] A front-end form is an interface element used to collect user input on a web page or application. With the popularity of the Internet, people began to look for more convenient ways to interact with websites. Forms emerged as the main way to collect user information. They usually contain various input fields, such as text boxes, drop-down lists, radio buttons, check boxes, etc., through which users can enter or select information. Forms may also contain buttons, such as submit buttons, which users can click to send form data to the server for processing.

[0003] During the front-end form generation process, before the user submits the form, the system will verify the user's input data according to the preset validation rules. Data validation is an important step to ensure the accuracy and completeness of user input data. In traditional data validation, the system usually relies on preset rules, such as mandatory field checking, data type checking, format matching, etc.

[0004] However, due to the limitations of preset rules, forms are only limited to preset rules for data validation, which cannot fully capture the complexity and diversity of user input. At this time, the data validation rules need to be updated, but manual code updates take too much time and cannot meet the current Internet's requirements for front-end form flexibility and efficiency. Summary of the invention

[0005] In response to the problems in the prior art, the present application provides a front-end form data verification method and device, which can improve the efficiency and flexibility of the front-end form based on enhanced data verification technology.

[0006] In order to solve at least one of the above problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a front-end form data verification method, comprising:

[0008] Acquire historical form data, determine corresponding first form data according to the historical form data and a named entity recognition rule, determine corresponding second form data according to the first form data and a relationship extraction rule, and determine entities and inter-entity relationships corresponding to the historical form data according to the second form data and an entity connection rule;

[0009] Input the graph structure constructed according to the entity and the relationship between the entities into the convolution layer of the preset graph neural network for graph convolution operation, input the feature vector corresponding to the entity obtained through the graph convolution operation into the fully connected layer of the preset graph neural network for nonlinear transformation operation, determine the corresponding nonlinear feature vector, input the nonlinear feature vector into the task specific layer of the preset graph neural network, determine the corresponding loss function according to the task features output by the task specific layer and the preset real features, adjust the parameters of the preset graph neural network according to the minimum loss function, determine the corresponding optimal graph neural network, and construct the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network;

[0010] Extracting form features from the form data knowledge graph, inputting the form features into a preset data verification model for model training, and determining a corresponding data verification model according to a minimum loss function after the model training;

[0011] Receive form data input by the user, and determine whether the form data complies with the rules in the preset rule engine according to a preset rule engine. If so, perform a form page rendering operation according to the form data; if not, input the form data into the data validation model to determine the corresponding validation result, determine the updated form data according to the validation result, and perform a form page rendering operation according to the updated form data.

[0012] Further, the acquiring of historical form data, determining corresponding first form data according to the historical form data and a named entity recognition rule, and determining corresponding second form data according to the first form data and a relationship extraction rule, includes:

[0013] Obtaining historical form data, and determining corresponding standard form data according to the historical form data, preset character encoding rules, and preset data cleaning rules;

[0014] Inputting the standard form data into a preset deep conditional random field model to perform a named entity recognition operation, and determining first form data containing entity features output by the preset deep conditional random field model;

[0015] A relationship extraction operation is performed according to the first form data and a preset relationship extraction rule to determine corresponding second form data including entity features and inter-entity relationship features.

[0016] Further, determining the entity and the relationship between entities corresponding to the historical form data according to the second form data and the entity connection rule includes:

[0017] Generate a corresponding entity vector representation according to the entity features in the second form data and a preset word embedding model, perform a cosine similarity calculation operation on the entity vector representation, and determine the corresponding entity vector similarity;

[0018] Performing a similarity matching operation on a preset form knowledge base according to the entity vector similarity and a preset similarity threshold to determine a corresponding matching entity;

[0019] A merging operation is performed on the matching entities, and entity features in the second form data are updated according to the merged entity obtained after the merging operation, so as to determine entities corresponding to the historical form data and relationships between entities.

[0020] Furthermore, the step of inputting the graph structure constructed according to the entities and the relationships between the entities into a convolutional layer of a preset graph neural network to perform a graph convolution operation includes:

[0021] Inputting a graph structure constructed according to the entities and the relationships between the entities into a convolutional layer of a preset graph neural network to determine features of each of the entities in the graph structure and features of neighboring entities adjacent to the entities;

[0022] A feature vector corresponding to the entity is determined according to the neighbor entity features and a preset aggregation function.

[0023] Furthermore, the step of inputting the feature vector corresponding to the entity obtained by the graph convolution operation into the fully connected layer of the preset graph neural network for nonlinear transformation operation to determine the corresponding nonlinear feature vector includes:

[0024] Inputting the feature vector corresponding to the entity obtained by the graph convolution operation into a fully connected layer of a preset graph neural network, and performing a linear transformation operation on the feature vector according to a preset weight matrix;

[0025] A nonlinear activation function adding operation is performed on the linear feature vector obtained after the linear transformation operation to determine the corresponding nonlinear feature vector.

[0026] Furthermore, the step of inputting the nonlinear feature vector into the task-specific layer of the preset graph neural network and determining the corresponding loss function according to the task features output by the task-specific layer and the preset real features includes:

[0027] Inputting the nonlinear transformation feature vector into the classification layer of the preset graph neural network to determine the corresponding classification feature;

[0028] The corresponding category cross entropy loss function is determined according to the classification features and the preset real features.

[0029] Furthermore, the step of inputting the nonlinear transformation feature vector into the task-specific layer of the preset graph neural network and determining the corresponding loss function according to the task features output by the task-specific layer and the preset real features further includes:

[0030] Inputting the nonlinear transformation feature vector into the embedding layer of the preset graph neural network to determine the corresponding embedding feature;

[0031] A corresponding binary cross entropy loss function is determined according to the embedded features and the preset real features.

[0032] In a second aspect, the present application provides a front-end form data verification device, comprising:

[0033] An entity feature association module, used to obtain historical form data, determine corresponding first form data according to the historical form data and a named entity recognition rule, determine corresponding second form data according to the first form data and a relationship extraction rule, and determine entities and inter-entity relationships corresponding to the historical form data according to the second form data and an entity connection rule;

[0034] A knowledge graph construction module, used for inputting the graph structure constructed according to the entities and the relationships between the entities into the convolution layer of the preset graph neural network for graph convolution operation, inputting the feature vector corresponding to the entity obtained through the graph convolution operation into the fully connected layer of the preset graph neural network for nonlinear transformation operation, determining the corresponding nonlinear feature vector, inputting the nonlinear feature vector into the task specific layer of the preset graph neural network, determining the corresponding loss function according to the task features output by the task specific layer and the preset real features, adjusting the parameters of the preset graph neural network according to the minimum loss function, determining the corresponding optimal graph neural network, and constructing the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network;

[0035] A data verification model training module is used to extract form features from the form data knowledge graph, input the form features into a preset data verification model for model training, and determine the corresponding data verification model according to the minimum loss function after the model training;

[0036] The data validation module is used to receive form data input by the user, and determine whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If so, the form page rendering operation is performed according to the form data; if not, the form data is input into the data validation model, the corresponding validation result is determined, the updated form data is determined according to the validation result, and the form page rendering operation is performed according to the updated form data.

[0037] In a third aspect, the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the front-end form data verification method when executing the program.

[0038] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the front-end form data verification method when executed by a processor.

[0039] In a fifth aspect, the present application provides a computer program product, including a computer program / instruction, which implements the steps of the front-end form data verification method when executed by a processor.

[0040] It can be seen from the above technical scheme that the present application provides a front-end form data verification method and device, which determines the entities and entity relationships corresponding to the historical form data, inputs the graph structure constructed according to the entities and entity relationships into the convolution layer of the preset graph neural network for graph convolution operation, determines the corresponding optimal graph neural network, constructs the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network, extracts the form features in the form data knowledge graph, determines the corresponding data verification model, receives the form data input by the user, and determines whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If not, the form data is input into the data verification model, determines the corresponding verification result, determines the updated form data according to the verification result, and performs form page rendering operation according to the updated form data, thereby improving the efficiency and flexibility of the front-end form based on the enhanced data verification technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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.

[0042] Figure 1 This is one of the flow charts of the front-end form data verification method in the embodiment of the present application;

[0043] Figure 2 This is a second flow chart of the front-end form data verification method in an embodiment of the present application;

[0044] Figure 3 This is a third flow chart of the front-end form data verification method in an embodiment of the present application;

[0045] Figure 4This is a fourth flow chart of the front-end form data verification method in an embodiment of the present application;

[0046] Figure 5 This is a fifth flow chart of the front-end form data verification method in an embodiment of the present application;

[0047] Figure 6 This is a sixth flow chart of the front-end form data verification method in an embodiment of the present application;

[0048] Figure 7 This is the seventh flow chart of the front-end form data verification method in the embodiment of the present application;

[0049] Figure 8 It is a structural diagram of the front-end form data verification device in the embodiment of the present application;

[0050] Fig. 9 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application.

[0051] Reference numerals:

[0052] Electronic device 9600, central processing unit 9100, memory 9140, communication module 9110, input unit 9120, audio processor 9130, display 9160, power supply 9170, buffer memory 9141, application / function storage unit 9142, data storage unit 9143, driver program storage unit 9144, antenna 9111, speaker 9131, microphone 9132. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0054] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0055] Considering that during the front-end form generation process, before the user submits the form, the system will validate the user's input data according to the preset validation rules. However, due to the limitations of the preset rules, the form is only limited to the preset rules when validating the data, and cannot fully capture the complexity and diversity of user input. At this time, the data validation rules need to be updated, but manually updating the code consumes too much time and cost, and cannot meet the current Internet's requirements for the flexibility and efficiency of front-end forms. The present application provides a front-end form data verification method and device, which determines the entities and the relationships between entities corresponding to the historical form data, inputs the graph structure constructed according to the entities and the relationships between entities into the convolution layer of the preset graph neural network for graph convolution operation, determines the corresponding optimal graph neural network, constructs the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network, extracts the form features in the form data knowledge graph, determines the corresponding data verification model, receives the form data input by the user, and determines whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If not, the form data is input into the data verification model, determines the corresponding verification result, determines the updated form data according to the verification result, and performs the form page rendering operation according to the updated form data, thereby improving the efficiency and flexibility of the front-end form based on the enhanced data verification technology.

[0056] In order to improve the efficiency and flexibility of the front-end form based on the enhanced data verification technology, the present application provides an embodiment of a front-end form data verification method, see Figure 1 , the front-end form data verification method specifically includes the following contents:

[0057] Step S101: acquiring historical form data, determining corresponding first form data according to the historical form data and a named entity recognition rule, determining corresponding second form data according to the first form data and a relationship extraction rule, and determining entities and inter-entity relationships corresponding to the historical form data according to the second form data and an entity connection rule;

[0058] Optionally, in this embodiment, obtaining historical form data and determining entities and relationships between entities based on historical form data information helps to build a structured, semantically rich knowledge base, thereby supporting various intelligent applications, such as intelligent question answering, recommendation systems, data analysis, etc. Based on the above purpose of building a knowledge graph, the core technologies for defining entities, attributes, and relationships between entities and the process of relationships include but are not limited to: at least one of named entity recognition, relationship extraction, and entity connection.

[0059] 1. Named Entity Recognition

[0060] The goal of named entity recognition (NER) is to identify and classify named entities from unstructured text data. Named entities usually refer to words or phrases with specific meanings in text, such as names, places, organizations, times, numbers, etc.

[0061] Optionally, in this embodiment, features that are helpful for identifying entities need to be designed, such as part-of-speech tagging results, word prefixes and suffixes, word form changes, etc. Use labeled data sets to train conditional random field models, which contain text and its corresponding entity labels. Apply the trained conditional random field model to new text to identify and classify entities.

[0062] For example, suppose you want to identify named entities in a text:

[0063] Text: "Apple Inc. is a technology company based in Cupertino, California, founded on April 1, 1976 by Steve Jobs, Steve Wozniak and Ronald Wayne."

[0064] The recognition results are as follows:

[0065] Organization: Apple Inc.

[0066] Location: Cupertino, California

[0067] Names: Steve Jobs, Steve Wozniak, Ronald Wayne

[0068] Date: April 1, 1976

[0069] It is understandable that NER is a key step in building multiple NLP applications such as knowledge graphs, information extraction, and question-answering systems, and its performance directly affects the results of downstream tasks.

[0070] 2. Relationship Extraction

[0071] Through relationship extraction rules, the corresponding entity relationships are learned from the preset knowledge base in a semi-automatic manner, and industry experts are organized to reinforce relationship annotations based on relationship rules, and the identified relationships are screened and verified, so that subsequent models can predict relationships more accurately.

[0072] 3. Entity Connection

[0073] Entity linking, which involves matching or linking entities mentioned in the text with specific entities in the knowledge base. This process helps resolve ambiguity in the text and ensures consistency and accuracy of entities.

[0074] Optionally, in this embodiment, for each entity identified by the entity recognition operation, a set of candidate entities is generated in the pre-built form knowledge base, and the most suitable candidate entity is selected as the link target. This step is the key to entity linking and needs to resolve the ambiguity of entities in the text. The final entity link is determined by matching the entities in the text with the entities in the knowledge base.

[0075] Optionally, in this embodiment, the step of selecting the most suitable candidate entity as the link target is implemented through a word embedding model. The word embedding model is selected to generate a vector representation of entity features, and the cosine similarity between entity vectors is calculated to evaluate the similarity between them. According to the calculated cosine similarity and a preset similarity threshold, the entity vector is matched with the entity vector in a preset form knowledge base to determine the most similar entity. The matched entities are merged, which involves the integration and deduplication of information. Based on the merged entity information, the entity features in the second form data are updated to ensure the consistency and accuracy of the data.

[0076] To illustrate, consider the following text: "Yesterday, I visited the Statue of Liberty in New York."

[0077] The entity recognition operation identified "New York" and "Statue of Liberty" as potential entities.

[0078] For "New York", candidate entities are generated such as New York City, New York State, etc.

[0079] For “Statue of Liberty”, candidate entities are generated such as the Statue of Liberty in New York, the Statue of Liberty in France, etc.

[0080] According to the context information, it is determined that "New York" refers to New York City, USA, and "Statue of Liberty" refers to the Statue of Liberty in New York. "New York" in the text is linked to the entity of New York City in the knowledge base, and "Statue of Liberty" is linked to the entity of the Statue of Liberty in New York. In this way, entity connection and word sense disambiguation are completed. Entity linking is a key step in applications such as knowledge graph construction, information extraction, and question-answering systems, which helps to improve the accuracy and richness of information.

[0081] It can be understood that the above steps perform entity recognition of unstructured text through named entity recognition, construct entity relationships through relationship extraction, and compare vector similarities and disambiguate entities in the form knowledge base through entity connection. The entities and entity relationships obtained in this way lay a solid foundation for the accuracy of the subsequent knowledge graph construction.

[0082] Step S102: inputting the graph structure constructed according to the entity and the relationship between the entities into the convolution layer of the preset graph neural network for graph convolution operation, inputting the feature vector corresponding to the entity obtained through the graph convolution operation into the fully connected layer of the preset graph neural network for nonlinear transformation operation, determining the corresponding nonlinear feature vector, inputting the nonlinear feature vector into the task specific layer of the preset graph neural network, determining the corresponding loss function according to the task features output by the task specific layer and the preset real features, adjusting the parameters of the preset graph neural network according to the minimum loss function, determining the corresponding optimal graph neural network, and constructing the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network;

[0083] Optionally, in this embodiment, a graph structure is constructed based on the entities and the relationships between entities obtained after processing in the above step S101, and the obtained graph structure is input into a preset graph neural network to construct a knowledge graph. The knowledge graph obtained based on this can learn more feature representations and increase the prediction performance of the knowledge graph.

[0084] Optionally, in this embodiment, a Graph Neural Network (GNN) is a deep learning model designed to process graph-structured data. They are particularly suitable for knowledge graphs because knowledge graphs are essentially graph representations of entities and their relationships. GNNs infer possible relationships between entities by learning the embedded representations of nodes (entities) in the graph and the connection patterns between nodes.

[0085] Optionally, in this embodiment, the graph convolution operation is to use the entities of the graph structure as nodes and the relationships as edges, and initialize the feature vector for each node (entity) in the graph, using a random vector or a vector based on entity attributes. The graph convolution operation and relationship embedding are applied to aggregate the information of neighbor nodes, each node collects the information of its neighbor nodes, and updates its own feature representation to obtain an updated entity feature vector, wherein the relationship embedding operation can be used together with the node features to improve the information aggregation process.

[0086] Optionally, in this embodiment, in the above-mentioned graph convolution step, the relationship embedding operation must first determine all possible relationship types in the knowledge graph and assign a unique identifier to each relationship. Then randomly initialize an embedding vector (a vector in a low-dimensional space) for each relationship. Finally, the embedded representation of the relationship is learned according to the preset word embedding model (BERT) or GNN network, including using the relationship embedding vector as a weight to weight the features of neighboring nodes. This step can significantly improve the aggregation process. When aggregating neighbor node features, it is not a simple mathematical operation (such as summing or averaging) on ​​the neighbor features, but the neighbor features are weighted according to the relationship type between the nodes. Relationship embeddings serve as weights to guide the flow of information.

[0087] Optionally, in this embodiment, in the above graph convolution step, different aggregation strategies are used to aggregate the information of neighbor nodes, including a summation strategy and an average strategy (this step will consider the weight of the relationship embedding vector).

[0088] Sum aggregation: For each node, calculate the sum of the feature vectors of all its direct neighbors, combine the aggregated neighbor features with the node's own features, and update the node's feature representation.

[0089] Average aggregation: For each node, the average of the feature vectors of all its direct neighbors is calculated, and the average neighbor features are combined with the node's own features to update the node's feature representation.

[0090] Optionally, in this embodiment, the relationship embedding vector is continuously updated through learning tasks (such as link prediction, knowledge graph completion, etc.) to adjust the embedding vector to reduce the prediction error.

[0091] It can be understood that by using the relation triplets (head entity, relation, tail entity) in the knowledge graph to train the GNN model, the GNN model can learn how to predict the tail entity based on the embedding and relation of the entity.

[0092] Optionally, in this embodiment, after the graph convolution layer obtains the updated entity feature vector, a fully connected layer is used to perform further linear transformation on the entity feature vector, and a ReLU activation function is applied to introduce nonlinearity to obtain nonlinear features to help the model learn complex feature interactions.

[0093] Optionally, in this embodiment, the above-mentioned nonlinear features are added to a specific task layer for task prediction according to specific application tasks. In this way, for different types of task modes, only the specific task layer needs to be replaced, and there is no need to retrain the model.

[0094] Optionally, in this embodiment, the specific task layer may be at least one of a classification layer, a regression layer, and a link prediction layer.

[0095] For classification tasks, a fully connected layer is added to process the features output by the previous layer to generate a fixed-length feature vector for the classification task. The loss function for multi-class classification tasks uses the category cross entropy loss.

[0096] For regression tasks, a fully connected layer is added to extract features and perform linear transformation, and the mean square error loss is selected as the loss function.

[0097] For prediction tasks, an embedding layer is added to generate embedding vectors for entity relationships in the knowledge graph. The function uses binary cross entropy loss, which is suitable for interpreting prediction results as probabilities.

[0098] Optionally, in this embodiment, the task features output by the above-mentioned specific task layer are compared with the real features for loss calculation, and the loss function is selected according to the task type. The parameters of the preset graph neural network model (GNN) are updated through the back propagation algorithm according to the minimum loss function to obtain the optimal graph neural network, and the knowledge graph is updated according to the output of the optimal graph neural network model, for example, by adding newly discovered relationships or entities.

[0099] It can be understood that the knowledge graph is a structured semantic knowledge base that stores and represents knowledge by organizing entities in the real world (such as people, places, organizations, etc.) and their relationships (such as "is", "belongs to", "is located in", etc.) in the form of a graph. The knowledge graph constructed through the above steps has the entity prediction function, laying the foundation for the subsequent training of the form data verification model.

[0100] Step S103: extracting form features from the form data knowledge graph, inputting the form features into a preset data verification model for model training, and determining a corresponding data verification model according to a minimum loss function after the model training;

[0101] It can be understood that this step is to use the constructed knowledge graph to train the preset neural network model, so that the trained neural network model can perform data verification based on the non-compliant data entered by the user into the front-end form and predict compliant data for form filling.

[0102] For example, a form requires users to fill in

[0103] Name: "Zhang San"

[0104] ID number: "120XXXXXXXXXXXXXX"

[0105] The user fills in the name column of the form: "Zhang San 120XXXXXXXXXXXXXX"

[0106] At this point, the data validation model can split the above non-compliant data according to the results of the knowledge graph prediction, and automatically fill in the split compliant data into the form.

[0107] Step S104: Receive form data input by the user, and determine whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If so, perform a form page rendering operation according to the form data; if not, input the form data into the data validation model to determine the corresponding validation result, determine the updated form data according to the validation result, and perform a form page rendering operation according to the updated form data.

[0108] Optionally, in this embodiment, the preset rule engine is a set of field validation rule templates pre-set by the form developer. Before the user submits the form, the system will validate the data entered by the user according to the preset validation rules. These validation rules may include mandatory item checks, data type checks, length restrictions, etc.

[0109] Optionally, in this embodiment, if the data input by the user meets the preset validation rules, the validation is passed, and the data input by the user is directly rendered on the front-end page.

[0110] Optionally, in this embodiment, if the data input by the user does not meet the pre-set verification rules and the verification fails, the user data is input into the data verification model trained in step S103 to obtain the verification result output by the data verification model, and the front-end page is rendered according to the verification result.

[0111] Optionally, in this embodiment, when the verification does not comply with the rules, the model can also issue an alarm to remind the user that the input data does not comply with the verification rules. At this time, the user can manually check the data output by the model. If the user clicks to agree, the model verification passes. If the user clicks to disagree and makes manual modifications, the user's modifications are recorded and the knowledge graph is updated to enhance the model's predictive ability.

[0112] This example demonstrates how this embodiment constructs a knowledge graph for prediction, and trains a data validation model based on the knowledge graph for data validation in front-end forms. When encountering data that does not conform to the rules, there is no need to manually modify the rules, which increases the efficiency and flexibility of the front-end forms.

[0113] From the above description, it can be seen that the front-end form data verification method provided in the embodiment of the present application can determine the entities and entity relationships corresponding to the historical form data, input the graph structure constructed according to the entities and entity relationships into the convolution layer of the preset graph neural network for graph convolution operation, determine the corresponding optimal graph neural network, construct the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network, extract the form features in the form data knowledge graph, determine the corresponding data verification model, receive the form data input by the user, and determine whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If not, the form data is input into the data verification model, the corresponding verification result is determined, the updated form data is determined according to the verification result, and the form page rendering operation is performed according to the updated form data, thereby improving the efficiency and flexibility of the front-end form based on the enhanced data verification technology.

[0114] In one embodiment of the front-end form data verification method of the present application, see Figure 2 , and can also include the following:

[0115] Step S201: Acquire historical form data, and determine corresponding standard form data according to the historical form data, preset character encoding rules, and preset data cleaning rules;

[0116] Step S202: inputting the standard form data into a preset deep conditional random field model to perform a named entity recognition operation, and determining first form data containing entity features output by the preset deep conditional random field model;

[0117] Step S203: performing a relationship extraction operation according to the first form data and a preset relationship extraction rule to determine corresponding second form data including entity features and inter-entity relationship features.

[0118] Optionally, in this embodiment, historical form data is obtained, and corresponding standard form data is determined based on the historical form data, preset character encoding rules, and preset data cleaning rules, so as to clean and standardize the text, convert it into a unified character encoding, remove irrelevant formats and noise, and lay the foundation for subsequent entity naming.

[0119] Optionally, in this embodiment, the preset deep conditional random field model is a model that combines deep learning with conditional random field (CRF), which is mainly used for sequence prediction tasks, especially when the input data has a complex structure and rich contextual information. DCRF improves the model's data modeling ability by combining the representation ability of deep neural networks with the sequence modeling ability of CRF.

[0120] Optionally, in this embodiment, features that are helpful for identifying entities need to be designed, such as part-of-speech tagging results, word prefixes and suffixes, word form changes, etc. Use labeled data sets to train conditional random field models, which contain text and its corresponding entity labels. Apply the trained conditional random field model to new text to identify and classify entities.

[0121] Optionally, in this embodiment, the relationship extraction rules learn the corresponding entity relationships from the preset knowledge base in a semi-automatic manner, and organize industry experts to reinforce the relationship annotations according to the relationship rules, and screen and verify the identified relationships, so that subsequent models can predict the relationships more accurately.

[0122] Through step S203, this embodiment obtains a form data including entity features and inter-entity relationship features, which lays a solid foundation for subsequent entity connection disambiguation and construction of new entities and inter-entity relationships.

[0123] In one embodiment of the front-end form data verification method of the present application, see Figure 3 , and can also include the following:

[0124] Step S301: Generate a corresponding entity vector representation according to the entity features in the second form data and a preset word embedding model, perform a cosine similarity calculation operation on the entity vector representation, and determine the corresponding entity vector similarity;

[0125] Step S302: performing a similarity matching operation on a preset form knowledge base according to the entity vector similarity and a preset similarity threshold, and determining a corresponding matching entity;

[0126] Step S303: performing a merging operation on the matching entities, and updating entity features in the second form data according to the merged entity obtained after the merging operation, to determine entities corresponding to the historical form data and relationships between entities.

[0127] Optionally, in this embodiment, the entity features identified by the named entity recognition technology are converted into a vector pattern through a pre-trained word embedding model (BERT), and the cosine similarity between the entity vectors is calculated to evaluate the similarity between them. Cosine similarity is calculated by measuring the similarity of two vectors in direction. According to the calculated cosine similarity and a preset similarity threshold, the entity vector is matched with the entity vector in a preset form knowledge base to determine the most similar entity. The matched entities are merged, which involves the integration and deduplication of information. Based on the merged entity information, the entity features in the second form data are updated to ensure the consistency and accuracy of the data. Finally, the entities corresponding to the historical form data and the relationship between them are determined based on the updated data.

[0128] Optionally, in this embodiment, the word embedding model may also adopt the CNN model itself, which does not affect the implementation of the embodiment of the present invention.

[0129] Optionally, in this embodiment, the similarity threshold is a numerical value used to determine to what extent the vector representations of two entities are considered similar. If the similarity score between the two entity vectors exceeds this threshold, they are considered similar or matched.

[0130] Through step S303, this embodiment realizes entity connection through word embedding, obtains disambiguated entities and inter-entity relationships, and lays a solid foundation for the subsequent construction of the knowledge graph.

[0131] In one embodiment of the front-end form data verification method of the present application, see Figure 4 , and can also include the following:

[0132] Step S401: inputting a graph structure constructed according to the entities and the relationships between the entities into a convolutional layer of a preset graph neural network to determine features of each of the entities in the graph structure and features of neighboring entities adjacent to the entity;

[0133] Step S402: determining a feature vector corresponding to the entity according to the neighbor entity features and a preset aggregation function.

[0134] Optionally, in this embodiment, a Graph Neural Network (GNN) is a deep learning model designed to process graph-structured data. They are particularly suitable for knowledge graphs because knowledge graphs are essentially graph representations of entities and their relationships. GNNs infer possible relationships between entities by learning the embedded representations of nodes (entities) in the graph and the connection patterns between nodes.

[0135] Optionally, in this embodiment, the graph convolution operation is to use the entities of the graph structure as nodes and the relationships as edges, and initialize the feature vector for each node (entity) in the graph, using a random vector or a vector based on entity attributes. The graph convolution operation is applied to aggregate the information of neighboring nodes. Each node collects the information of its neighboring nodes and updates its own feature representation to obtain an updated entity feature vector.

[0136] Optionally, in this embodiment, the graph convolution operation can be used in conjunction with the relationship embedding operation and node features to improve the information aggregation process.

[0137] Optionally, in this embodiment, in the above graph convolution step, there are two options for the aggregation function, including:

[0138] Sum aggregation: For each node, calculate the sum of the feature vectors of all its direct neighbors, combine the aggregated neighbor features with the node's own features, and update the node's feature representation.

[0139] Average aggregation: For each node, the average of the feature vectors of all its direct neighbors is calculated, and the average neighbor features are combined with the node's own features to update the node's feature representation.

[0140] Through step S402, this embodiment successfully obtains the feature vector corresponding to the updated entity, which lays a foundation for subsequent further vector transformation and learning of complex features.

[0141] In one embodiment of the front-end form data verification method of the present application, see Figure 5 , and can also include the following:

[0142] Step S501: inputting the feature vector corresponding to the entity obtained by the graph convolution operation into the fully connected layer of the preset graph neural network, and performing a linear transformation operation on the feature vector according to a preset weight matrix;

[0143] Step S502: performing a nonlinear activation function adding operation on the linear feature vector obtained after the linear transformation operation to determine a corresponding nonlinear feature vector.

[0144] Optionally, in this embodiment, after the graph convolution layer obtains the updated entity feature vector, a fully connected layer is used to perform further linear transformation on the entity feature vector, and a ReLU activation function is applied to introduce nonlinearity to obtain nonlinear features to help the model learn complex feature interactions.

[0145] Through step S502, this embodiment successfully introduces nonlinear features, which adds richness to the subsequent prediction results of specific task features.

[0146] In one embodiment of the front-end form data verification method of the present application, see Figure 6 , and can also include the following:

[0147] Step S601: inputting the nonlinear transformation feature vector into the classification layer of the preset graph neural network to determine the corresponding classification feature;

[0148] Step S602: Determine a corresponding category cross entropy loss function according to the classification features and preset real features.

[0149] Optionally, in this embodiment, the above-mentioned nonlinear features are added to a specific task layer for task prediction according to specific application tasks. In this way, for different types of task modes, only the specific task layer needs to be replaced, and there is no need to retrain the model.

[0150] Optionally, in this embodiment, the specific task layer may be at least one of a classification layer, a regression layer, and a link prediction layer.

[0151] For classification tasks, a fully connected layer is added to process the features output by the previous layer to generate a fixed-length feature vector for the classification task. The loss function for multi-class classification tasks uses the category cross entropy loss.

[0152] Through step S602, this embodiment successfully predicts the classification task and obtains a knowledge graph that can be used for the classification task.

[0153] In one embodiment of the front-end form data verification method of the present application, see Figure 7 , and can also include the following:

[0154] Step S701: inputting the nonlinear transformation feature vector into the embedding layer of the preset graph neural network to determine the corresponding embedding feature;

[0155] Step S702: Determine a corresponding binary cross entropy loss function according to the embedded features and the preset real features.

[0156] Optionally, in this embodiment, the above-mentioned nonlinear features are added to a specific task layer for task prediction according to specific application tasks. In this way, for different types of task modes, only the specific task layer needs to be replaced, and there is no need to retrain the model.

[0157] Optionally, in this embodiment, the specific task layer may be at least one of a classification layer, a regression layer, and a link prediction layer.

[0158] For prediction tasks, an embedding layer is added to generate embedding vectors for entity relationships in the knowledge graph. The function uses binary cross entropy loss, which is suitable for interpreting prediction results as probabilities.

[0159] Through step S702, this embodiment successfully constructs a knowledge graph for prediction tasks, laying a foundation for subsequent training data verification models to predict form data input by users.

[0160] In order to improve the efficiency and flexibility of the front-end form based on the enhanced data verification technology, the present application provides an embodiment of a front-end form data verification device for implementing all or part of the content of the front-end form data verification method, see Figure 8 , the front-end form data verification device specifically includes the following contents:

[0161] The entity feature association module 10 is used to obtain historical form data, determine the corresponding first form data according to the historical form data and the named entity recognition rule, determine the corresponding second form data according to the first form data and the relationship extraction rule, and determine the entity corresponding to the historical form data and the relationship between entities according to the second form data and the entity connection rule;

[0162] The knowledge graph construction module 20 is used to input the graph structure constructed according to the entity and the relationship between the entities into the convolution layer of the preset graph neural network for graph convolution operation, input the feature vector corresponding to the entity obtained through the graph convolution operation into the fully connected layer of the preset graph neural network for nonlinear transformation operation, determine the corresponding nonlinear feature vector, input the nonlinear feature vector into the task specific layer of the preset graph neural network, determine the corresponding loss function according to the task features output by the task specific layer and the preset real features, adjust the parameters of the preset graph neural network according to the minimum loss function, determine the corresponding optimal graph neural network, and construct the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network;

[0163] A data verification model training module 30 is used to extract form features from the form data knowledge graph, input the form features into a preset data verification model for model training, and determine a corresponding data verification model according to a minimum loss function after the model training;

[0164] The data verification module 40 is used to receive form data input by the user, and determine whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If so, the form page rendering operation is performed according to the form data; if not, the form data is input into the data verification model, and the corresponding verification result is determined. The updated form data is determined according to the verification result, and the form page rendering operation is performed according to the updated form data.

[0165] From the above description, it can be seen that the front-end form data verification device provided by the embodiment of the present application can determine the entities and entity relationships corresponding to the historical form data, input the graph structure constructed according to the entities and entity relationships into the convolution layer of the preset graph neural network for graph convolution operation, determine the corresponding optimal graph neural network, construct the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network, extract the form features in the form data knowledge graph, determine the corresponding data verification model, receive the form data input by the user, and determine whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If not, the form data is input into the data verification model, the corresponding verification result is determined, the updated form data is determined according to the verification result, and the form page rendering operation is performed according to the updated form data, thereby improving the efficiency and flexibility of the front-end form based on the enhanced data verification technology.

[0166] From the hardware level, in order to improve the efficiency and flexibility of the front-end form based on the enhanced data verification technology, the present application provides an embodiment of an electronic device for implementing all or part of the content of the front-end form data verification method, and the electronic device specifically includes the following content:

[0167] Processor, memory, communication interface and bus; wherein the processor, memory and communication interface communicate with each other through the bus; the communication interface is used to realize information transmission between the front-end form data verification method and the core business system, user terminal and related database and other related devices; the logic controller can be a desktop computer, a tablet computer and a mobile terminal, etc., but the present embodiment is not limited thereto. In the present embodiment, the logic controller can be implemented with reference to the embodiment of the front-end form data verification method in the embodiment, and the embodiment of the front-end form data verification method, and the contents thereof are incorporated herein, and the repeated parts are not repeated.

[0168] It is understandable that the user terminal may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, a smart watch, a smart bracelet, etc.

[0169] In practical applications, part of the front-end form data verification method can be executed on the electronic device side as described above, or all operations can be completed in the client device. The specific selection can be based on the processing capability of the client device and the limitations of the user's usage scenario. This application does not limit this. If all operations are completed in the client device, the client device may also include a processor.

[0170] The client device may have a communication module (i.e., a communication unit) that can communicate with a remote server to achieve data transmission with the server. The server may include a server on the task scheduling center side, and other implementation scenarios may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, or a server cluster consisting of multiple servers, or a server structure of a distributed device.

[0171] Fig. 9 FIG. 9 is a schematic block diagram of the system structure of the electronic device 9600 according to an embodiment of the present application. Fig. 9 As shown, the electronic device 9600 may include a central processor 9100 and a memory 9140; the memory 9140 is coupled to the central processor 9100. It is worth noting that Fig. 9 is exemplary; other types of structures may also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0172] In one embodiment, the front-end form data verification method function may be integrated into the central processing unit 9100. The central processing unit 9100 may be configured to perform the following control:

[0173] Step S101: acquiring historical form data, determining corresponding first form data according to the historical form data and a named entity recognition rule, determining corresponding second form data according to the first form data and a relationship extraction rule, and determining entities and inter-entity relationships corresponding to the historical form data according to the second form data and an entity connection rule;

[0174] Step S102: inputting the graph structure constructed according to the entity and the relationship between the entities into the convolution layer of the preset graph neural network for graph convolution operation, inputting the feature vector corresponding to the entity obtained through the graph convolution operation into the fully connected layer of the preset graph neural network for nonlinear transformation operation, determining the corresponding nonlinear feature vector, inputting the nonlinear feature vector into the task specific layer of the preset graph neural network, determining the corresponding loss function according to the task features output by the task specific layer and the preset real features, adjusting the parameters of the preset graph neural network according to the minimum loss function, determining the corresponding optimal graph neural network, and constructing the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network;

[0175] Step S103: extracting form features from the form data knowledge graph, inputting the form features into a preset data verification model for model training, and determining a corresponding data verification model according to a minimum loss function after the model training.

[0176] Step S104: Receive form data input by the user, and determine whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If so, perform a form page rendering operation according to the form data; if not, input the form data into the data validation model to determine the corresponding validation result, determine the updated form data according to the validation result, and perform a form page rendering operation according to the updated form data.

[0177] From the above description, it can be seen that the electronic device provided in the embodiment of the present application determines the entities and the relationships between entities corresponding to the historical form data, inputs the graph structure constructed according to the entities and the relationships between entities into the convolution layer of the preset graph neural network for graph convolution operation, determines the corresponding optimal graph neural network, constructs the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network, extracts the form features in the form data knowledge graph, determines the corresponding data verification model, receives the form data input by the user, and determines whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If not, the form data is input into the data verification model, the corresponding verification result is determined, the updated form data is determined according to the verification result, and the form page rendering operation is performed according to the updated form data, thereby improving the efficiency and flexibility of the front-end form based on the enhanced data verification technology.

[0178] In another embodiment, the front-end form data verification method can be configured separately from the central processing unit 9100. For example, the front-end form data verification method can be configured as a chip connected to the central processing unit 9100, and the function of the front-end form data verification method is realized through the control of the central processing unit.

[0179] like Fig. 9 As shown, the electronic device 9600 may also include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It is worth noting that the electronic device 9600 does not necessarily have to include Fig. 9 In addition, the electronic device 9600 may also include Fig. 9 For components not shown, reference may be made to the prior art.

[0180] like Fig. 9As shown, the central processing unit 9100 is sometimes also referred to as a controller or an operation control, and may include a microprocessor or other processor device and / or logic device, which receives input and controls the operation of various components of the electronic device 9600.

[0181] The memory 9140 may be, for example, one or more of a cache, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory or other suitable devices. The above-mentioned information related to the failure may be stored, and a program for executing the relevant information may also be stored. The CPU 9100 may execute the program stored in the memory 9140 to implement information storage or processing, etc.

[0182] The input unit 9120 provides input to the central processing unit 9100. The input unit 9120 is, for example, a key or a touch input device. The power supply 9170 is used to provide power to the electronic device 9600. The display 9160 is used to display display objects such as images and texts. The display may be, for example, an LCD display, but is not limited thereto.

[0183] The memory 9140 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It may also be a memory that saves information even when the power is off, can be selectively erased, and is provided with more data, examples of which are sometimes referred to as EPROMs, etc. The memory 9140 may also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142, which is used to store application programs and function programs or processes for executing the operation of the electronic device 9600 through the central processor 9100.

[0184] The memory 9140 may also include a data storage unit 9143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers for communication functions of the electronic device and / or for executing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0185] The communication module 9110 is a transmitter / receiver that sends and receives signals via the antenna 9111. The communication module 9110 (transmitter / receiver) is coupled to the central processor 9100 to provide input signals and receive output signals, which may be the same as the case of a conventional mobile communication terminal.

[0186] Based on different communication technologies, multiple communication modules 9110 may be provided in the same electronic device, such as a cellular network module, a Bluetooth module and / or a wireless LAN module, etc. The communication module (transmitter / receiver) 9110 is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide an audio output via the speaker 9131 and receive an audio input from the microphone 9132, thereby realizing a common telecommunication function. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. In addition, the audio processor 9130 is also coupled to the central processor 9100, so that recording can be performed on the local machine through the microphone 9132, and the sound stored on the local machine can be played through the speaker 9131.

[0187] The embodiments of the present application also provide a computer-readable storage medium capable of implementing all the steps of the front-end form data verification method in the above-mentioned embodiment, where the execution subject is a server or a client. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, all the steps of the front-end form data verification method in the above-mentioned embodiment, where the execution subject is a server or a client, are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0188] Step S101: acquiring historical form data, determining corresponding first form data according to the historical form data and a named entity recognition rule, determining corresponding second form data according to the first form data and a relationship extraction rule, and determining entities and inter-entity relationships corresponding to the historical form data according to the second form data and an entity connection rule;

[0189] Step S102: inputting the graph structure constructed according to the entity and the relationship between the entities into the convolution layer of the preset graph neural network for graph convolution operation, inputting the feature vector corresponding to the entity obtained through the graph convolution operation into the fully connected layer of the preset graph neural network for nonlinear transformation operation, determining the corresponding nonlinear feature vector, inputting the nonlinear feature vector into the task specific layer of the preset graph neural network, determining the corresponding loss function according to the task features output by the task specific layer and the preset real features, adjusting the parameters of the preset graph neural network according to the minimum loss function, determining the corresponding optimal graph neural network, and constructing the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network;

[0190] Step S103: extracting form features from the form data knowledge graph, inputting the form features into a preset data verification model for model training, and determining a corresponding data verification model according to a minimum loss function after the model training.

[0191] Step S104: Receive form data input by the user, and determine whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If so, perform a form page rendering operation according to the form data; if not, input the form data into the data validation model to determine the corresponding validation result, determine the updated form data according to the validation result, and perform a form page rendering operation according to the updated form data.

[0192] From the above description, it can be seen that the computer-readable storage medium provided in the embodiment of the present application determines the entities and the relationships between entities corresponding to the historical form data, inputs the graph structure constructed according to the entities and the relationships between entities into the convolution layer of the preset graph neural network for graph convolution operation, determines the corresponding optimal graph neural network, constructs the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network, extracts the form features in the form data knowledge graph, determines the corresponding data verification model, receives the form data input by the user, and determines whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If not, the form data is input into the data verification model, the corresponding verification result is determined, the updated form data is determined according to the verification result, and the form page rendering operation is performed according to the updated form data, thereby improving the efficiency and flexibility of the front-end form based on the enhanced data verification technology.

[0193] The embodiments of the present application also provide a computer program product capable of implementing all steps of the front-end form data verification method in the above embodiments, where the execution subject is a server or a client. When the computer program / instruction is executed by a processor, the steps of the front-end form data verification method are implemented. For example, the computer program / instruction implements the following steps:

[0194] Step S101: acquiring historical form data, determining corresponding first form data according to the historical form data and a named entity recognition rule, determining corresponding second form data according to the first form data and a relationship extraction rule, and determining entities and inter-entity relationships corresponding to the historical form data according to the second form data and an entity connection rule;

[0195] Step S102: inputting the graph structure constructed according to the entity and the relationship between the entities into the convolution layer of the preset graph neural network for graph convolution operation, inputting the feature vector corresponding to the entity obtained through the graph convolution operation into the fully connected layer of the preset graph neural network for nonlinear transformation operation, determining the corresponding nonlinear feature vector, inputting the nonlinear feature vector into the task specific layer of the preset graph neural network, determining the corresponding loss function according to the task features output by the task specific layer and the preset real features, adjusting the parameters of the preset graph neural network according to the minimum loss function, determining the corresponding optimal graph neural network, and constructing the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network;

[0196] Step S103: extracting form features from the form data knowledge graph, inputting the form features into a preset data verification model for model training, and determining a corresponding data verification model according to a minimum loss function after the model training.

[0197] Step S104: Receive form data input by the user, and determine whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If so, perform a form page rendering operation according to the form data; if not, input the form data into the data validation model to determine the corresponding validation result, determine the updated form data according to the validation result, and perform a form page rendering operation according to the updated form data.

[0198] From the above description, it can be seen that the computer program product provided by the embodiment of the present application determines the entities and the relationships between entities corresponding to the historical form data, inputs the graph structure constructed according to the entities and the relationships between entities into the convolution layer of the preset graph neural network for graph convolution operation, determines the corresponding optimal graph neural network, constructs the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network, extracts the form features in the form data knowledge graph, determines the corresponding data verification model, receives the form data input by the user, and determines whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If not, the form data is input into the data verification model, the corresponding verification result is determined, the updated form data is determined according to the verification result, and the form page rendering operation is performed according to the updated form data, thereby improving the efficiency and flexibility of the front-end form based on the enhanced data verification technology.

[0199] It should be understood by those skilled in the art that embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0200] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (apparatus), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0201] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0202] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0203] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A front-end form data verification method, characterized in that: The method comprises: Acquire historical form data, determine corresponding first form data according to the historical form data and a named entity recognition rule, determine corresponding second form data according to the first form data and a relationship extraction rule, and determine entities and inter-entity relationships corresponding to the historical form data according to the second form data and an entity connection rule; Input the graph structure constructed according to the entity and the relationship between the entities into the convolution layer of the preset graph neural network for graph convolution operation, input the feature vector corresponding to the entity obtained through the graph convolution operation into the fully connected layer of the preset graph neural network for nonlinear transformation operation, determine the corresponding nonlinear feature vector, input the nonlinear feature vector into the task specific layer of the preset graph neural network, determine the corresponding loss function according to the task features output by the task specific layer and the preset real features, adjust the parameters of the preset graph neural network according to the minimum loss function, determine the corresponding optimal graph neural network, and construct the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network; Extracting form features from the form data knowledge graph, inputting the form features into a preset data verification model for model training, and determining a corresponding data verification model according to a minimum loss function after the model training; Receive form data input by the user, and determine whether the form data complies with the rules in the preset rule engine according to a preset rule engine. If so, perform a form page rendering operation according to the form data; if not, input the form data into the data validation model to determine the corresponding validation result, determine the updated form data according to the validation result, and perform a form page rendering operation according to the updated form data.

2. The front-end form data verification method according to claim 1, characterized in that: The acquiring of historical form data, determining corresponding first form data according to the historical form data and a named entity recognition rule, and determining corresponding second form data according to the first form data and a relationship extraction rule, includes: Obtaining historical form data, and determining corresponding standard form data according to the historical form data, preset character encoding rules, and preset data cleaning rules; Inputting the standard form data into a preset deep conditional random field model to perform a named entity recognition operation, and determining first form data containing entity features output by the preset deep conditional random field model; A relationship extraction operation is performed according to the first form data and a preset relationship extraction rule to determine corresponding second form data including entity features and inter-entity relationship features.

3. The front-end form data verification method according to claim 1, characterized in that: The determining the entity and the relationship between entities corresponding to the historical form data according to the second form data and the entity connection rule includes: Generate a corresponding entity vector representation according to the entity features in the second form data and a preset word embedding model, perform a cosine similarity calculation operation on the entity vector representation, and determine the corresponding entity vector similarity; Performing a similarity matching operation on a preset form knowledge base according to the entity vector similarity and a preset similarity threshold to determine a corresponding matching entity; A merging operation is performed on the matching entities, and entity features in the second form data are updated according to the merged entity obtained after the merging operation, so as to determine entities corresponding to the historical form data and relationships between entities.

4. The front-end form data verification method according to claim 1, characterized in that: The step of inputting the graph structure constructed according to the entities and the relationships between the entities into a convolutional layer of a preset graph neural network to perform a graph convolution operation includes: Inputting a graph structure constructed according to the entities and the relationships between the entities into a convolutional layer of a preset graph neural network to determine features of each of the entities in the graph structure and features of neighboring entities adjacent to the entities; A feature vector corresponding to the entity is determined according to the neighbor entity features and a preset aggregation function.

5. The front-end form data verification method according to claim 1, characterized in that: The step of inputting the feature vector corresponding to the entity obtained by the graph convolution operation into the fully connected layer of the preset graph neural network for nonlinear transformation operation to determine the corresponding nonlinear feature vector includes: Inputting the feature vector corresponding to the entity obtained by the graph convolution operation into a fully connected layer of a preset graph neural network, and performing a linear transformation operation on the feature vector according to a preset weight matrix; A nonlinear activation function adding operation is performed on the linear feature vector obtained after the linear transformation operation to determine the corresponding nonlinear feature vector.

6. The front-end form data verification method according to claim 1, characterized in that: The step of inputting the nonlinear feature vector into the task-specific layer of the preset graph neural network and determining the corresponding loss function according to the task features output by the task-specific layer and the preset real features comprises: Inputting the nonlinear feature vector into the classification layer of the preset graph neural network to determine the corresponding classification feature; The corresponding category cross entropy loss function is determined according to the classification features and the preset real features.

7. The front-end form data verification method according to claim 1, characterized in that: The step of inputting the nonlinear feature vector into the task-specific layer of the preset graph neural network and determining the corresponding loss function according to the task features output by the task-specific layer and the preset real features also includes: Inputting the nonlinear feature vector into the embedding layer of the preset graph neural network to determine the corresponding embedding feature; A corresponding binary cross entropy loss function is determined according to the embedded features and the preset real features.

8. A front-end form data verification device, characterized in that: The device comprises: An entity feature association module, used to obtain historical form data, determine corresponding first form data according to the historical form data and a named entity recognition rule, determine corresponding second form data according to the first form data and a relationship extraction rule, and determine entities and inter-entity relationships corresponding to the historical form data according to the second form data and an entity connection rule; A knowledge graph construction module, used for inputting the graph structure constructed according to the entities and the relationships between the entities into the convolution layer of the preset graph neural network for graph convolution operation, inputting the feature vector corresponding to the entity obtained through the graph convolution operation into the fully connected layer of the preset graph neural network for nonlinear transformation operation, determining the corresponding nonlinear feature vector, inputting the nonlinear feature vector into the task specific layer of the preset graph neural network, determining the corresponding loss function according to the task features output by the task specific layer and the preset real features, adjusting the parameters of the preset graph neural network according to the minimum loss function, determining the corresponding optimal graph neural network, and constructing the corresponding form data knowledge graph according to the graph features output by the optimal graph neural network; A data verification model training module is used to extract form features from the form data knowledge graph, input the form features into a preset data verification model for model training, and determine the corresponding data verification model according to the minimum loss function after the model training; The data validation module is used to receive form data input by the user, and determine whether the form data complies with the rules in the preset rule engine according to the preset rule engine. If so, the form page rendering operation is performed according to the form data; if not, the form data is input into the data validation model, the corresponding validation result is determined, the updated form data is determined according to the validation result, and the form page rendering operation is performed according to the updated form data.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the front-end form data verification method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the front-end form data verification method described in any one of claims 1 to 7 are implemented.