Front-end form data maintenance method and device
The front-end form performance prediction and resource optimization are carried out through the time series model, which solves the problems of insufficient form adaptability and performance degradation under traditional maintenance methods, and achieves the improvement of efficiency and flexibility.
Patent Information
- Application Number
- CN202411010614.8
- 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
Traditional front-end form performance maintenance methods require manual writing of a large amount of code, which cannot meet the Internet's requirements for front-end form flexibility and efficiency. Moreover, form adaptability and performance degradation as business needs change.
By collecting historical form usage data, using preset time series models (such as LSTM) for feature extraction and performance prediction, if performance deteriorates, lazy loading technology will be used to optimize and adjust resources to improve the efficiency and flexibility of the form.
It has achieved improved efficiency and flexibility of front-end forms based on predictive performance optimization, reducing the need for manual intervention and code modification, and adapting to changes in business needs.
Smart Images

Figure CN118535566B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of front-end technology, and in particular to a front-end form data maintenance 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] The performance of a front-end form refers to the time it takes for the form to adjust resource allocation and respond. As business needs change, the completed form structure and data storage requirements may also change, resulting in insufficient form adaptability and degraded performance. The traditional front-end form performance maintenance method requires front-end developers to manually write a large amount of code and frequently modify the code based on changes in these requirements.
[0004] However, with the rapid development of the Internet era, the demand for forms is also changing rapidly. At this time, manually making changes to each form requirement is time-consuming, labor-intensive and inefficient, 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 maintenance method and device, which can improve the efficiency and flexibility of the front-end form based on predictive performance optimization.
[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 maintenance method, comprising:
[0008] Collect historical form usage data, perform feature evaluation operations on the historical form usage data according to preset feature correlation analysis rules, and determine corresponding form features;
[0009] Input the form features into the first long short-term memory layer of the preset time series model for preliminary feature extraction, input the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model for deep feature extraction, determine the corresponding deep features, determine the corresponding loss function according to the deep features, adjust the weight of the preset time series model according to the minimum loss function and the back propagation algorithm, and determine the corresponding time series model, wherein the first long short-term memory layer and the second long short-term memory layer are connected by a residual connection;
[0010] Receive the user's modification operation on the front-end form, determine the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, and judge whether the form performance has declined according to the form performance prediction result and a preset form performance baseline. If declined, use lazy loading technology to perform resource optimization and adjustment operations on the front-end form.
[0011] Furthermore, the collecting of historical form usage data, performing feature evaluation operations on the historical form usage data according to preset feature correlation analysis rules, and determining corresponding form features include:
[0012] Collect historical form data, perform data cleaning operations on the historical form data, and determine corresponding clean form data;
[0013] A feature evaluation operation is performed on the clean form data according to a preset correlation coefficient calculation technology and a feature selection algorithm to determine corresponding form features.
[0014] Furthermore, the performing of feature evaluation operation on the clean form data according to the preset correlation coefficient calculation technology and feature selection algorithm to determine the corresponding form features includes:
[0015] Performing a correlation coefficient calculation operation on the clean form data according to a preset Spearman rank correlation coefficient rule to determine corresponding correlation features;
[0016] The corresponding form features are determined according to the relevant features and the recursive feature elimination algorithm.
[0017] Furthermore, the step of inputting the form features into the first long short-term memory layer of a preset time series model for preliminary feature extraction includes:
[0018] Inputting the form feature into the input gate of the first long short-term memory layer of the preset time series model to determine the corresponding input feature;
[0019] Inputting the input feature into the forget gate of the first long short-term memory layer to determine the corresponding retained feature;
[0020] The retained features are input into the output gate of the first long short-term memory layer to determine corresponding preliminary features.
[0021] Furthermore, the step of inputting the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model to perform a deep feature extraction operation, determining the corresponding deep features, and determining the corresponding loss function according to the deep features includes:
[0022] Inputting the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model to perform a deep feature extraction operation to determine the corresponding deep features;
[0023] The deep features are input into the output layer of a preset time series model, and the corresponding loss function is determined according to the predicted features output by the output layer and the preset real features.
[0024] Furthermore, after inputting the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model to perform a deep feature extraction operation, determining the corresponding deep features, and determining the corresponding loss function according to the deep features, the method includes:
[0025] Determining a corresponding regularization term according to the parameters of the loss function and a preset regularization coefficient;
[0026] The loss function is updated according to the regularization term, and the minimum loss function is determined according to the updated loss function and a preset gradient descent algorithm.
[0027] Furthermore, judging whether the form performance has declined according to the form performance prediction result and the preset form performance baseline, and if so, using lazy loading technology to perform resource optimization and adjustment operations on the front-end form, includes:
[0028] Determine whether the form performance has decreased according to the form performance prediction result and the preset form performance baseline, and if so, perform a scroll event monitoring operation to determine the corresponding optimal resource loading time;
[0029] A resource optimization adjustment operation is performed according to the optimal resource loading time.
[0030] In a second aspect, the present application provides a front-end form data maintenance device, comprising:
[0031] A form feature extraction module is used to collect historical form usage data, perform feature evaluation operations on the historical form usage data according to preset feature correlation analysis rules, and determine corresponding form features;
[0032] A model training module, used for inputting the form features into a first long short-term memory layer of a preset time series model for preliminary feature extraction, inputting the preliminary features output by the first long short-term memory layer into a second long short-term memory layer of the preset time series model for deep feature extraction, determining corresponding deep features, determining corresponding loss functions according to the deep features, adjusting the weights of the preset time series model according to the minimum loss function and a back propagation algorithm, and determining a corresponding time series model, wherein the first long short-term memory layer and the second long short-term memory layer are connected via a residual connection;
[0033] The prediction and maintenance optimization module is used to receive the user's modification operation on the front-end form, determine the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, and judge whether the form performance has declined according to the form performance prediction result and the preset form performance baseline. If it has declined, use lazy loading technology to perform resource optimization and adjustment operations on the front-end form.
[0034] 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 maintenance method when executing the program.
[0035] 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 maintenance method when executed by a processor.
[0036] 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 maintenance method when executed by a processor.
[0037] It can be seen from the above technical scheme that the present application provides a front-end form data maintenance method and device, which collects historical form usage data, determines the corresponding form features, inputs the form features into the first long short-term memory layer of the preset time series model for preliminary feature extraction, inputs the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model for deep feature extraction, determines the corresponding deep features, determines the corresponding time series model according to the deep features, receives the user's modification operation on the front-end form, determines the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, determines whether the form performance has decreased according to the form performance prediction result and the preset form performance baseline, and if it has decreased, uses lazy loading technology to perform resource optimization and adjustment operations on the front-end form, thereby improving the efficiency and flexibility of the front-end form based on predictive performance optimization. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] 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.
[0039] Figure 1 This is one of the flow charts of the front-end form data maintenance method in the embodiment of the present application;
[0040] Figure 2 This is the second flow chart of the front-end form data maintenance method in the embodiment of the present application;
[0041] Figure 3 This is a third flow chart of the front-end form data maintenance method in the embodiment of the present application;
[0042] Figure 4 This is a fourth flow chart of the front-end form data maintenance method in an embodiment of the present application;
[0043] Figure 5 This is a fifth flow chart of the front-end form data maintenance method in the embodiment of the present application;
[0044] Figure 6 This is a sixth flow chart of the front-end form data maintenance method in the embodiment of the present application;
[0045] Figure 7 This is the seventh flow chart of the front-end form data maintenance method in the embodiment of the present application;
[0046] Figure 8 It is a structural diagram of the front-end form data maintenance device in the embodiment of the present application;
[0047] Fig. 9 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application.
[0048] Reference numerals:
[0049] 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
[0050] 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.
[0051] 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.
[0052] Considering that with the rapid development of the Internet era, the demand for front-end forms is changing rapidly. With the change of business needs, the completed form structure and data storage requirements may also change, resulting in insufficient form adaptability and performance degradation. The traditional front-end form performance maintenance method requires front-end developers to manually write a large amount of code. However, it is time-consuming, labor-intensive and inefficient to manually change each form requirement, which cannot meet the requirements of the current Internet for the flexibility and efficiency of front-end forms. The present application provides a front-end form data maintenance method and device, which collects historical form usage data, determines the corresponding form features, inputs the form features into the first long short-term memory layer of the preset time series model for preliminary feature extraction, inputs the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model for deep feature extraction, determines the corresponding deep features, determines the corresponding time series model according to the deep features, receives the user's modification operation on the front-end form, determines the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, and judges whether the form performance is reduced according to the form performance prediction result and the preset form performance baseline. If it is reduced, the front-end form is optimized and adjusted by using lazy loading technology, thereby improving the efficiency and flexibility of the front-end form based on predictive performance optimization.
[0053] In order to improve the efficiency and flexibility of the front-end form based on predictive performance optimization, the present application provides an embodiment of a front-end form data maintenance method, see Figure 1 , the front-end form data maintenance method specifically includes the following contents:
[0054] Step S101: collecting historical form usage data, performing feature evaluation operations on the historical form usage data according to preset feature correlation analysis rules, and determining corresponding form features;
[0055] Optionally, in this embodiment, historical form usage data includes but is not limited to data related to form performance and data on form structure changes, wherein form performance related data includes but is not limited to page loading time, response time, resource loading time, element rendering time, CPU and memory usage, etc., and form structure change data includes but is not limited to newly added functions, modified fields, user interaction patterns, user click events, etc. It can be understood that the combination of these data helps to discover which usage patterns will cause the performance of the front-end form to degrade during the user's interaction with the front-end form, which is convenient for helping the model to distinguish and identify these patterns during model training later, complete predictive resource allocation adjustments, and optimize the form, for example: formulate a dynamic resource adjustment strategy based on the model prediction results, and for non-critical resources, lazy loading technology can be used, that is, loading only when the user scrolls to the corresponding position or when needed.
[0056] Optionally, in this embodiment, the collected data is cleaned to remove outliers and noise, including removing invalid or erroneous records, filling missing values, and then standardizing or normalizing the data to make features of different magnitudes comparable, which is a key step in data processing and helps the accuracy of the data set. The cleaned data is subjected to feature evaluation to extract features that are helpful for model understanding, such as resource size, number of requests, user behavior patterns, etc.
[0057] Optionally, in this embodiment, feature evaluation uses feature importance evaluation, correlation analysis and other methods to select the features that have the greatest impact on the prediction target. Feature correlation analysis is an important step in data preprocessing and feature selection, which is used to evaluate the relationship between different features and the target variable, as well as the relationship between features.
[0058] First, feature correlation analysis requires data exploration to gain a preliminary understanding of the distribution of the data set, including the type of feature (numeric or categorical), range, missing values, etc.
[0059] Secondly, data visualization is required to intuitively display feature distribution and outliers through scatter plots, box plots, violin plots and other visualization methods, and then apply statistical tests, such as t-test or ANOVA, to evaluate whether the differences between different features and the target variable are significant.
[0060] Then, a correlation coefficient calculation is performed to calculate the correlation coefficient between the features, such as Spearman's rank correlation coefficient (suitable for non-linear relationships). A feature selection algorithm, Recursive Feature Elimination (RFE), is applied to identify the most influential features.
[0061] Finally, based on the results of the correlation analysis, feature transformation, feature combination, or creation of new features are performed to enhance the predictive ability of the model.
[0062] For example, in Example 1, suppose we are analyzing a data set and our goal is to predict housing prices. Our data set contains multiple features, such as house area, number of rooms, location, year of construction, etc.
[0063] Data Exploration: We first examine the data type and distribution of the features.
[0064] Data Visualization: By plotting scatter plots and box plots, we can see that house size and number of rooms have a strong linear relationship with house price.
[0065] Statistical test: Use t-test to determine if there is a significant difference in house prices in different areas.
[0066] Correlation coefficient calculation: The correlation coefficients between all characteristics and house prices were calculated, and it was found that the correlation between the construction year and house prices was low.
[0067] Heatmap: Generate a correlation coefficient heatmap to quickly identify highly correlated feature pairs, such as house size and number of rooms.
[0068] Feature Selection Algorithm: Use RFE to determine the most important features for house price prediction.
[0069] Principal Component Analysis: PCA is applied to reduce the feature dimension and extract the principal components most relevant to housing prices.
[0070] Model-based feature importance: Train a random forest model and use feature importance scores to assess the contribution of each feature.
[0071] Iterative optimization: Based on the results of feature importance, the most important features are selected for model training, and the feature set is continuously optimized iteratively.
[0072] Through feature correlation analysis, we can better understand the relationship between data features and target variables, provide valuable information for model training, and improve the prediction accuracy of the model.
[0073] Example 2: If one feature is "number of form elements" and the other is "form loading time", then the distribution of "number of form elements" and "form loading time" and the relationship between the features are understood through visual scatter plots and descriptive statistics. The Spearman rank correlation coefficient is used to determine whether the features have a significant correlation with form performance, and the correlation coefficient between the two features is calculated. If the correlation coefficient is close to +1 or -1, it indicates that there is a strong linear relationship between them; if it is close to 0, it indicates that there is no obvious linear relationship between them. In this way, we can identify key features that have a significant impact on form performance.
[0074] It is understandable that model training through identified features, especially adding attention mechanisms to these key features, helps the model accurately identify key factors that affect performance and predict form performance degradation in real time when users modify the front-end form for these key factors, and take subsequent measures to improve performance.
[0075] Step S102: input the form features into the first long short-term memory layer of the preset time series model to perform a preliminary feature extraction operation, input the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model to perform a deep feature extraction operation, determine the corresponding deep features, determine the corresponding loss function according to the deep features, adjust the weight of the preset time series model according to the minimum loss function and the back propagation algorithm, and determine the corresponding time series model, wherein the first long short-term memory layer and the second long short-term memory layer are connected by a residual connection;
[0076] Optionally, in this embodiment, the preset time series model is an LSTM model. The LSTM model is suitable for scenarios that require prediction and can make relatively accurate predictions. Based on this, in order to increase the accuracy of the prediction, a neural network model including two functional layers and an output layer is established. The functional layers can be constructed with both layers being LSTM layers, or the first layer being LSTM and the second layer being GRU.
[0077] Among them, when both layers are LSTM layers, the first LSTM layer is responsible for extracting preliminary feature representations from the input data, which can capture the temporal dependencies in the data, and these features are then passed to the next layer of the network. In order to solve the gradient disappearance problem, the second LSTM layer uses a bidirectional LSTM to further extract higher-level features while solving the gradient disappearance problem or utilizing bidirectional information.
[0078] Another implementation to solve the vanishing gradient problem is to add residual connections between LSTM layers which can help the gradient flow and reduce the vanishing gradient problem during training.
[0079] Optionally, in this embodiment, when the first layer is LSTM and the second layer is GRU, the second LSTM layer is replaced by a GRU layer, which has fewer parameters but can provide similar functions.
[0080] Optionally, in this embodiment, the output layer (fully connected layer) is usually located at the end of the model to convert the features extracted by the LSTM layer into the final prediction results. For example, in a regression problem, it can output a continuous numerical prediction.
[0081] Optionally, in this embodiment, the form features after feature analysis are input into the first LSTM layer. Through the forget gate and input gate of the first LSTM layer, the LSTM unit decides which information to discard and which information to update and store. Through the output gate, the LSTM unit generates an output that reflects the hidden state of the current time step and can be used for downstream tasks. The parameters learned by the first LSTM layer during training include the weights and biases of the forget gate, input gate, and output gate. The configuration of the first LSTM layer can be adjusted according to task requirements, including the number of neurons, the depth of the layer, the options for return sequences, etc.
[0082] Optionally, in this embodiment, the specific steps of learning the weights and bias parameters of the forget gate, input gate, and output gate in the first LSTM layer are:
[0083] Initialization parameters: Before training begins, all weights and bias parameters in the LSTM layer are initialized randomly or through some preset strategy.
[0084] Forward propagation: Input data is fed into the LSTM layer, and the LSTM unit processes the data at each time step. The forget gate, input gate, and output gate calculate the gating signal based on the current input and the hidden state of the previous time step.
[0085] Update cell state and hidden state: Update cell state and hidden state using gating signal.
[0086] Loss function calculation: The difference between the model's output and the true label is calculated using the loss function.
[0087] Backpropagation: The gradient of the loss function with respect to the network parameters is calculated using the backpropagation algorithm. This involves the chain rule, where the gradient of each parameter is calculated step by step from the output layer to the input layer.
[0088] Parameter Update: Update the weights and biases in the LSTM layer based on the calculated gradients using gradient descent or its variants like Adam, RMSprop, etc.
[0089] It can be understood that in this way, the weights and bias parameters of the forget gate, input gate, and output gate in the first LSTM layer are continuously adjusted during the training process to minimize the loss function, so that the model can better capture the patterns and dependencies in the data. The trained first LSTM layer will pass the extracted preliminary features to the next layer of the network.
[0090] Optionally, in this embodiment, the preliminary features output by the first long short-term memory layer (LSTM layer) are input into the second long short-term memory layer of the preset time series model for deep feature extraction operation to determine the corresponding deep features, and the second LSTM layer receives the output from the first LSTM layer as input. If the first LSTM layer is set to return sequences (return_sequences=True), its output is a three-dimensional array that can be directly used as input for the second LSTM layer. The second LSTM layer processes the input data and executes its own forgetting, input, and output gating mechanisms to extract deeper features that may not be fully extracted in the first LSTM layer, such as longer-distance dependencies. The output of the second LSTM layer can be combined with the output of the first LSTM layer to provide a richer feature representation.
[0091] It is understandable that through this deep feature extraction process, the model is able to learn more abstract and advanced feature representations, which are very useful for solving complex sequence prediction and pattern recognition tasks.
[0092] Optionally, in this embodiment, the second LSTM layer may use a bidirectional LSTM to simultaneously process forward and reverse context information in the second LSTM layer, providing a more comprehensive sequence understanding.
[0093] Optionally, in this embodiment, the second LSTM layer can also be replaced by a GRU layer, which has fewer parameters but can provide similar functions.
[0094] Optionally, in this embodiment, a residual connection may be added between the first LSTM layer and the second LSTM layer to help gradient flow and alleviate the gradient vanishing or exploding problem in the deep network.
[0095] Optionally, in this embodiment, an attention mechanism can be introduced to enable the second LSTM layer to focus on the key parts of the input sequence to enhance the interpretability of the model.
[0096] It can be understood that the above steps, for adding or changing the second LSTM layer, are intended to enhance the expressiveness and adaptability of the model, and can be used alone or in combination without affecting the implementation of the embodiments of the present invention.
[0097] Optionally, in this embodiment, after obtaining the output of the second LSTM layer, the obtained deep features are input to the output layer (fully connected layer), converted into a suitable task format, and the final predicted values are obtained through the output layer. These predicted values are calculated based on the current parameters of the model and the input data, and then the target value or true label corresponding to each sample is determined. These target values are the known correct results in the training data set. The difference between the predicted value and the target value is calculated using the selected loss function. The result of the loss function is a scalar value that represents the error size under the current model parameters.
[0098] Optionally, in this embodiment, in order to control the complexity of the model and prevent overfitting, a regularization term is added to the obtained loss function to improve the generalization ability of the model.
[0099] In the optional embodiment, the regularization term uses L2 regularization, and the steps are:
[0100] Define the regularization coefficient: Define a coefficient for the regularization term (usually expressed as λ or α), which controls the contribution of the regularization term to the total loss. (This coefficient is a hyperparameter and the optimal value needs to be determined through cross-validation)
[0101] Construct regularization term: According to the selected regularization type, construct the regularization term. For L2 regularization, it is the sum of the squares of the model parameters.
[0102] Modify the loss function: Add the regularization term to the original loss function. The new loss function is usually expressed as the sum of the original loss function and the regularization term.
[0103] Back propagation: During model training, the gradient of the new loss function with respect to the model parameters is calculated through the back propagation algorithm. (This process takes into account both the original loss and the regularization term to find the best model parameters that can not only minimize the prediction error but also meet the regularization requirements, thereby improving the generalization ability of the model)
[0104] Parameter update: Update the model parameters using gradient descent or its variants. The regularization term affects the direction and step size of the parameter update.
[0105] Adjust the regularization coefficient: According to the performance of the model on the validation set, adjust the regularization coefficient. Multiple attempts are required to find the optimal regularization strength.
[0106] Monitor the training process: Monitor the loss and validation performance during the training process to ensure that the model is not overfitting while maintaining a good fit.
[0107] It is understandable that through the above regularization method, the regularization term helps to limit the complexity of the model and improve the model's ability to generalize to new data. The loss function corresponding to the optimal regularization strength is determined as the minimum loss function, and the model parameters are updated inversely with this loss function to obtain the best time series model, which can solve complex sequence prediction and pattern recognition tasks.
[0108] Step S103: Receive the user's modification operation on the front-end form, determine the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, and judge whether the form performance has declined according to the form performance prediction result and the preset form performance baseline. If declined, use lazy loading technology to perform resource optimization and adjustment operations on the front-end form.
[0109] Optionally, in this embodiment, the system receives form configuration information entered by the user through a graphical interface or configuration file. This information includes but is not limited to field name, field type, field validation rules, form layout, etc. The user can customize this configuration information according to business needs.
[0110] After obtaining the information added by the user, the time series model trained in step S102 will perform form performance prediction based on the information added by the user and generate a prediction result.
[0111] Optionally, in this embodiment, the preset performance baseline is a performance baseline established using historical data, which is used to determine normal and abnormal performance indicators. The prediction results output by the time series model are compared with the performance baseline indicators to determine whether they are normal. If it is predicted that the form performance will deteriorate due to the user's addition operation, the lazy loading technology is applied in advance for this change to improve the performance of the front-end form.
[0112] Optionally, in this embodiment, lazy loading is an optimization technology that is mainly used to improve the loading speed and performance of web pages or applications. Its core idea is to delay the loading of resources that are not immediately needed until the user needs them. In this way, the amount of data at the initial loading time can be reduced, the response speed of the page can be accelerated, and the user experience can be improved.
[0113] In the context of front-end forms, lazy loading technology can be applied to the following aspects:
[0114] Lazy loading of fields: For forms containing a large number of fields, you can only load the fields that the user currently needs to fill in, and load other fields when the user scrolls to the corresponding position.
[0115] Asynchronous data loading: If the form needs to obtain data from the server (such as drop-down list options), you can use lazy loading technology to request data only when the user needs to fill in the relevant fields.
[0116] Lazy loading of images and multimedia resources: For multimedia resources such as images or videos contained in the form, they can be loaded only when the user scrolls to their location to reduce the burden of initial page loading.
[0117] Lazy loading of scripts and style sheets: For non-critical JavaScript scripts or CSS style sheets, you can use lazy loading to load them only when needed.
[0118] Component lazy loading: In a single-page application (SPA), you can use route-level lazy loading, which loads components of a page only when a user visits that specific page.
[0119] Optionally, in this embodiment, the lazy loading technology determines when to load resources by monitoring scrolling events.
[0120] The following are some specific application scenarios:
[0121] Images and videos: Load images or videos only when the user scrolls to a specific part of the page.
[0122] JavaScript and CSS files: For scripts and style sheets used below the fold, you can delay loading.
[0123] Web fonts: Fonts can be loaded when the page is scrolled to the text containing the font.
[0124] Components and Modules: In a single-page application (SPA), different components or modules can be loaded on demand.
[0125] It is understandable that in terms of performance prediction and baseline judgment, if the form performance is monitored to be degraded, lazy loading technology can be used for optimization. By analyzing user behavior and resource usage, it is determined which resources can be lazy loaded, thereby reducing the initial loading time of the page and improving the response speed and overall performance of the form.
[0126] This example shows how this embodiment analyzes user behavior and resource usage, predicts which resources may cause performance bottlenecks, and optimizes them in advance.
[0127] From the above description, it can be seen that the front-end form data maintenance method provided in the embodiment of the present application can determine the corresponding form features by collecting historical form usage data, input the form features into the first long short-term memory layer of the preset time series model for preliminary feature extraction operations, input the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model for deep feature extraction operations, determine the corresponding deep features, determine the corresponding time series model according to the deep features, receive the user's modification operation on the front-end form, determine the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, judge whether the form performance has decreased according to the form performance prediction result and the preset form performance baseline, and if it has decreased, use lazy loading technology to perform resource optimization and adjustment operations on the front-end form, thereby improving the efficiency and flexibility of the front-end form based on predictive performance optimization.
[0128] In one embodiment of the front-end form data maintenance method of the present application, see Figure 2 , and can also include the following:
[0129] Step S201: collecting historical form data, performing data cleaning operations on the historical form data, and determining corresponding clean form data;
[0130] Step S202: performing a feature evaluation operation on the clean form data according to a preset correlation coefficient calculation technique and a feature selection algorithm to determine corresponding form features.
[0131] Optionally, in this embodiment, historical form usage data includes but is not limited to data related to form performance and data on form structure changes, wherein data related to form performance includes but is not limited to page loading time, response time, resource loading time, element rendering time, CPU and memory usage, etc., and data on form structure changes includes but is not limited to newly added functions, modified fields, user interaction modes, user click events, etc.
[0132] Optionally, in this embodiment, the collected data is cleaned to remove outliers and noise, including removing invalid or erroneous records, filling missing values, and then standardizing or normalizing the data to make features of different magnitudes comparable, which is a key step in data processing and helps the accuracy of the data set. The cleaned data is subjected to feature evaluation to extract features that are helpful for model understanding, such as resource size, number of requests, user behavior patterns, etc.
[0133] Optionally, in this embodiment, the purpose of combining the preset correlation coefficient calculation technology and the feature selection algorithm is to perform feature evaluation, select the most important features, and subsequently introduce an attention mechanism to cooperate with model training to increase the accuracy of model prediction.
[0134] Optionally, in this embodiment, a correlation coefficient is calculated, and the correlation coefficient is used to quantify the strength and direction of the linear relationship between two variables. The following are the steps for calculating the correlation coefficient:
[0135] Data preparation: Prepare two datasets of the variables whose relationship needs to be analyzed.
[0136] Calculate Correlation Coefficient: Calculates the Spearman rank correlation coefficient (suitable for non-linear relationships).
[0137] Interpreting the Correlation Coefficient: Interpret the value of the correlation coefficient, where -1 indicates a perfect negative correlation, 0 indicates no correlation, and 1 indicates a perfect positive correlation.
[0138] Assess significance: Use a t-test or z-test to determine the significance of the correlation coefficient to determine whether the relationship between the variables is likely to be caused by random variation alone. Calculate the P value, which is the probability of observing the current or more extreme result if the null hypothesis is true.
[0139] Visualization: Display the relationship between variables through visualization methods such as scatter plots.
[0140] To illustrate, suppose we are analyzing two variables X and Y to determine if there is a statistically significant relationship between them. We calculate the Spearman rank correlation coefficient and assume that we get a correlation coefficient of 0.85, indicating that there is a strong positive linear relationship between X and Y. We check the significance P value. If the P value is less than 0.05, we can consider the relationship to be statistically significant.
[0141] By combining statistical tests and correlation coefficient calculations, a more complete understanding of the relationships between data features and variables can be obtained.
[0142] Optionally, in this embodiment, a feature selection algorithm, Recursive Feature Elimination (RFE), is used to identify the most influential features.
[0143] By combining correlation coefficient calculation and feature selection algorithm, we can better understand the relationship between data features and target variables, provide valuable information for model training, and improve the prediction accuracy of the model.
[0144] Through step S202, this embodiment performs feature evaluation on historical data to obtain evaluated form features, laying a foundation for subsequent model training.
[0145] In one embodiment of the front-end form data maintenance method of the present application, see Figure 3 , and can also include the following:
[0146] Step S301: performing a correlation coefficient calculation operation on the clean form data according to a preset Spearman rank correlation coefficient rule to determine corresponding correlation features;
[0147] Step S302: determining corresponding form features according to the relevant features and the recursive feature elimination algorithm;
[0148] Optionally, in this embodiment, feature evaluation uses correlation analysis to select the features that have the greatest impact on the prediction target. Feature correlation analysis is an important step in data preprocessing and feature selection, which is used to evaluate the relationship between different features and the target variable, as well as the relationship between features.
[0149] First, feature correlation analysis requires data exploration to gain a preliminary understanding of the distribution of the data set, including the type of feature (numeric or categorical), range, missing values, etc.
[0150] Secondly, data visualization is required to intuitively display feature distribution and outliers through scatter plots, box plots, violin plots and other visualization methods, and then apply statistical tests, such as t-test or ANOVA, to evaluate whether the differences between different features and the target variable are significant.
[0151] Then, a correlation coefficient calculation is performed to calculate the correlation coefficient between the features, such as Spearman's rank correlation coefficient (suitable for non-linear relationships). A feature selection algorithm, Recursive Feature Elimination (RFE), is applied to identify the most influential features.
[0152] Through step S302, this embodiment obtains the evaluated form features, laying a foundation for subsequent model training.
[0153] In one embodiment of the front-end form data maintenance method of the present application, see Figure 4 , and can also include the following:
[0154] Step S401: inputting the form feature into the input gate of the first long short-term memory layer of the preset time series model to determine the corresponding input feature;
[0155] Step S402: inputting the input feature into the forget gate of the first long short-term memory layer to determine the corresponding retained feature;
[0156] Step S403: input the retained features into the output gate of the first long short-term memory layer to determine the corresponding preliminary features.
[0157] Optionally, in this embodiment, the form features after feature analysis are input into the first LSTM layer. Through the forget gate and input gate of the first LSTM layer, the LSTM unit decides which information to discard and which information to update and store. Through the output gate, the LSTM unit generates an output that reflects the hidden state of the current time step and can be used for downstream tasks. The parameters learned by the first LSTM layer during training include the weights and biases of the forget gate, input gate, and output gate. The configuration of the first LSTM layer can be adjusted according to task requirements, including the number of neurons, the depth of the layer, the options for return sequences, etc.
[0158] Optionally, in this embodiment, the specific steps of learning the weights and bias parameters of the forget gate, input gate, and output gate in the first LSTM layer are:
[0159] Initialization parameters: Before training begins, all weights and bias parameters in the LSTM layer are initialized randomly or through some preset strategy.
[0160] Forward propagation: Input data is fed into the LSTM layer, and the LSTM unit processes the data at each time step. The forget gate, input gate, and output gate calculate the gating signal based on the current input and the hidden state of the previous time step.
[0161] Update cell state and hidden state: Update cell state and hidden state using gating signal.
[0162] Loss function calculation: The difference between the model's output and the true label is calculated using the loss function.
[0163] Backpropagation: The gradient of the loss function with respect to the network parameters is calculated using the backpropagation algorithm. This involves the chain rule, where the gradient of each parameter is calculated step by step from the output layer to the input layer.
[0164] Parameter Update: Update the weights and biases in the LSTM layer based on the calculated gradients using gradient descent or its variants like Adam, RMSprop, etc.
[0165] It can be understood that in this way, the weights and bias parameters of the forget gate, input gate, and output gate in the first LSTM layer are continuously adjusted during the training process to minimize the loss function, so that the model can better capture the patterns and dependencies in the data. The trained first LSTM layer will pass the extracted preliminary features to the next layer of the network.
[0166] Through step S403, this embodiment successfully trains the first LSTM layer, which can be used to extract preliminary features and lay the foundation for subsequent extraction of deep features.
[0167] In one embodiment of the front-end form data maintenance method of the present application, see Figure 5 , and can also include the following:
[0168] Step S501: inputting the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model to perform a deep feature extraction operation to determine the corresponding deep features;
[0169] Step S502: input the deep features into the output layer of a preset time series model, and determine the corresponding loss function according to the predicted features output by the output layer and the preset real features.
[0170] Optionally, in this embodiment, the preliminary features output by the first long short-term memory layer (LSTM layer) are input into the second long short-term memory layer of the preset time series model for deep feature extraction operation to determine the corresponding deep features, and the second LSTM layer receives the output from the first LSTM layer as input. If the first LSTM layer is set to return sequences (return_sequences=True), its output is a three-dimensional array that can be directly used as input for the second LSTM layer. The second LSTM layer processes the input data and executes its own forgetting, input, and output gating mechanisms to extract deeper features that may not be fully extracted in the first LSTM layer, such as longer-distance dependencies. The output of the second LSTM layer can be combined with the output of the first LSTM layer to provide a richer feature representation.
[0171] It is understandable that through this deep feature extraction process, the model is able to learn more abstract and advanced feature representations, which are very useful for solving complex sequence prediction and pattern recognition tasks.
[0172] Optionally, in this embodiment, the second LSTM layer may use a bidirectional LSTM to simultaneously process forward and reverse context information in the second LSTM layer, providing a more comprehensive sequence understanding.
[0173] Optionally, in this embodiment, the second LSTM layer can also be replaced by a GRU layer, which has fewer parameters but can provide similar functions.
[0174] Optionally, in this embodiment, a residual connection may be added between the first LSTM layer and the second LSTM layer to help gradient flow and alleviate the gradient vanishing or exploding problem in the deep network.
[0175] Optionally, in this embodiment, an attention mechanism can be introduced to enable the second LSTM layer to focus on the key parts of the input sequence to enhance the interpretability of the model.
[0176] It can be understood that the above steps, for adding or changing the second LSTM layer, are intended to enhance the expressiveness and adaptability of the model, and can be used alone or in combination without affecting the implementation of the embodiments of the present invention.
[0177] Through step S502, this embodiment successfully trains the second LSTM layer, which can be used to extract deep features and lay the foundation for subsequent calculation of model loss.
[0178] In one embodiment of the front-end form data maintenance method of the present application, see Figure 6 , and can also include the following:
[0179] Step S601: determining a corresponding regularization term according to the parameters of the loss function and a preset regularization coefficient;
[0180] Step S602: updating the loss function according to the regularization term, and determining the minimum loss function according to the updated loss function and a preset gradient descent algorithm.
[0181] Optionally, in this embodiment, in order to control the complexity of the model and prevent overfitting, a regularization term is added to the obtained loss function to improve the generalization ability of the model.
[0182] In the optional embodiment, the regularization term uses L2 regularization, and the steps are:
[0183] Define the regularization coefficient: Define a coefficient for the regularization term (usually expressed as λ or α), which controls the contribution of the regularization term to the total loss. (This coefficient is a hyperparameter and the optimal value needs to be determined through cross-validation)
[0184] Construct regularization term: According to the selected regularization type, construct the regularization term. For L2 regularization, it is the sum of the squares of the model parameters.
[0185] Modify the loss function: Add the regularization term to the original loss function. The new loss function is usually expressed as the sum of the original loss function and the regularization term.
[0186] Back propagation: During model training, the gradient of the new loss function with respect to the model parameters is calculated through the back propagation algorithm. (This process takes into account both the original loss and the regularization term to find the best model parameters that can not only minimize the prediction error, i.e. the minimum loss function, but also meet the requirements of regularization, thereby improving the generalization ability of the model)
[0187] Parameter update: Update the model parameters using gradient descent or its variants. The regularization term affects the direction and step size of the parameter update.
[0188] Adjust the regularization coefficient: According to the performance of the model on the validation set, adjust the regularization coefficient. Multiple attempts are required to find the optimal regularization strength.
[0189] Monitor the training process: Monitor the loss and validation performance during the training process to ensure that the model is not overfitting while maintaining a good fit.
[0190] It can be understood that by determining the loss function corresponding to the optimal regularization strength as the minimum loss function and using this loss function to reversely update the model parameters, the best time series model can be obtained, which can solve complex sequence prediction and pattern recognition tasks.
[0191] Through step S602, this embodiment successfully obtains a time series model for solving complex sequence prediction and pattern recognition tasks, laying a foundation for real-time performance adjustment of subsequent prediction form patterns.
[0192] In one embodiment of the front-end form data maintenance method of the present application, see Figure 7 , and can also include the following:
[0193] Step S701: judging whether the form performance has decreased according to the form performance prediction result and the preset form performance baseline, and if decreased, performing a scroll event monitoring operation to determine the corresponding optimal resource loading time;
[0194] Step S702: performing resource optimization and adjustment operations according to the optimal resource loading time;
[0195] Optionally, in this embodiment, the preset performance baseline is a performance baseline established using historical data, which is used to determine normal and abnormal performance indicators. The prediction results output by the time series model are compared with the performance baseline indicators to determine whether they are normal. If it is predicted that the form performance will deteriorate due to the user's addition operation, the lazy loading technology is applied in advance for this change to improve the performance of the front-end form.
[0196] Optionally, in this embodiment, the lazy loading technology determines when to load resources by monitoring scrolling events.
[0197] Optionally, in this embodiment, lazy loading is an optimization technology that is mainly used to improve the loading speed and performance of web pages or applications. Its core idea is to delay the loading of resources that are not immediately needed until the user needs them. In this way, the amount of data at the initial loading time can be reduced, the response speed of the page can be accelerated, and the user experience can be improved.
[0198] In the context of front-end forms, lazy loading technology can be applied to the following aspects:
[0199] Lazy loading of fields: For forms containing a large number of fields, you can only load the fields that the user currently needs to fill in, and load other fields when the user scrolls to the corresponding position.
[0200] Asynchronous data loading: If the form needs to obtain data from the server (such as drop-down list options), you can use lazy loading technology to request data only when the user needs to fill in the relevant fields.
[0201] Lazy loading of images and multimedia resources: For multimedia resources such as images or videos contained in the form, they can be loaded only when the user scrolls to their location to reduce the burden of initial page loading.
[0202] Lazy loading of scripts and style sheets: For non-critical JavaScript scripts or CSS style sheets, you can use lazy loading to load them only when needed.
[0203] Component lazy loading: In a single-page application (SPA), you can use route-level lazy loading, which loads components of a page only when a user visits that specific page.
[0204] Through step S702, this embodiment successfully predicts performance bottlenecks in advance based on the interaction between the user and the front-end form and implements real-time resource allocation and scheduling based on lazy loading technology.
[0205] In order to improve the efficiency and flexibility of the front-end form based on predictive performance optimization, the present application provides an embodiment of a front-end form data maintenance device for implementing all or part of the content of the front-end form data maintenance method, see Figure 8 The front-end form data maintenance device specifically includes the following contents:
[0206] The form feature extraction module 10 is used to collect historical form usage data, perform feature evaluation operations on the historical form usage data according to preset feature correlation analysis rules, and determine corresponding form features;
[0207] The model training module 20 is used to input the form features into the first long short-term memory layer of the preset time series model for preliminary feature extraction, input the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model for deep feature extraction, determine the corresponding deep features, determine the corresponding loss function according to the deep features, adjust the weight of the preset time series model according to the minimum loss function and the back propagation algorithm, and determine the corresponding time series model, wherein the first long short-term memory layer and the second long short-term memory layer are connected by a residual connection;
[0208] The prediction and maintenance optimization module 30 is used to receive the user's modification operation on the front-end form, determine the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, and judge whether the form performance has declined according to the form performance prediction result and the preset form performance baseline. If it has declined, use lazy loading technology to perform resource optimization and adjustment operations on the front-end form.
[0209] From the above description, it can be seen that the front-end form data maintenance device provided in the embodiment of the present application can determine the corresponding form features by collecting historical form usage data, input the form features into the first long short-term memory layer of the preset time series model for preliminary feature extraction operations, input the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model for deep feature extraction operations, determine the corresponding deep features, determine the corresponding time series model according to the deep features, receive the user's modification operation on the front-end form, determine the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, judge whether the form performance has declined according to the form performance prediction result and the preset form performance baseline, and if it has declined, use lazy loading technology to perform resource optimization and adjustment operations on the front-end form, thereby improving the efficiency and flexibility of the front-end form based on predictive performance optimization.
[0210] From the hardware level, in order to improve the efficiency and flexibility of the front-end form based on predictive performance optimization, the present application provides an embodiment of an electronic device for implementing all or part of the content of the front-end form data maintenance method, and the electronic device specifically includes the following content:
[0211] 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 maintenance 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 maintenance method and the embodiment of the front-end form data maintenance method in the embodiment, and the contents thereof are incorporated herein, and the repeated parts are not repeated.
[0212] 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.
[0213] In practical applications, part of the front-end form data maintenance 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.
[0214] 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.
[0215] 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.
[0216] In one embodiment, the front-end form data maintenance method function may be integrated into the central processing unit 9100. The central processing unit 9100 may be configured to perform the following control:
[0217] Step S101: collecting historical form usage data, performing feature evaluation operations on the historical form usage data according to preset feature correlation analysis rules, and determining corresponding form features;
[0218] Step S102: input the form features into the first long short-term memory layer of the preset time series model to perform a preliminary feature extraction operation, input the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model to perform a deep feature extraction operation, determine the corresponding deep features, determine the corresponding loss function according to the deep features, adjust the weight of the preset time series model according to the minimum loss function and the back propagation algorithm, and determine the corresponding time series model, wherein the first long short-term memory layer and the second long short-term memory layer are connected by a residual connection;
[0219] Step S103: Receive the user's modification operation on the front-end form, determine the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, and judge whether the form performance has declined according to the form performance prediction result and the preset form performance baseline. If declined, use lazy loading technology to perform resource optimization and adjustment operations on the front-end form.
[0220] From the above description, it can be seen that the electronic device provided in the embodiment of the present application collects historical form usage data, determines the corresponding form features, inputs the form features into the first long short-term memory layer of the preset time series model for preliminary feature extraction, inputs the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model for deep feature extraction, determines the corresponding deep features, determines the corresponding time series model according to the deep features, receives the user's modification operation on the front-end form, determines the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, determines whether the form performance has decreased according to the form performance prediction result and the preset form performance baseline, and if it has decreased, uses lazy loading technology to perform resource optimization and adjustment operations on the front-end form, thereby improving the efficiency and flexibility of the front-end form based on predictive performance optimization.
[0221] In another embodiment, the front-end form data maintenance method can be configured separately from the central processing unit 9100. For example, the front-end form data maintenance method can be configured as a chip connected to the central processing unit 9100, and the functions of the front-end form data maintenance method are realized through the control of the central processing unit.
[0222] like Fig. 9As 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.
[0223] like Fig. 9 As 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. The central processing unit 9100 receives input and controls the operation of various components of the electronic device 9600.
[0224] 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.
[0225] 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.
[0226] 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.
[0227] 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.).
[0228] 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.
[0229] 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.
[0230] 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 maintenance 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 maintenance method in the above-mentioned embodiment are implemented. For example, when the processor executes the computer program, the following steps are implemented:
[0231] Step S101: collecting historical form usage data, performing feature evaluation operations on the historical form usage data according to preset feature correlation analysis rules, and determining corresponding form features;
[0232] Step S102: input the form features into the first long short-term memory layer of the preset time series model for preliminary feature extraction, input the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model for deep feature extraction, determine the corresponding deep features, determine the corresponding loss function according to the deep features, adjust the weight of the preset time series model according to the minimum loss function and the back propagation algorithm, and determine the corresponding time series model, wherein the first long short-term memory layer and the second long short-term memory layer are connected by a residual connection message analysis model;
[0233] Step S103: Receive the user's modification operation on the front-end form, determine the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, and judge whether the form performance has declined according to the form performance prediction result and the preset form performance baseline. If declined, use lazy loading technology to perform resource optimization and adjustment operations on the front-end form.
[0234] From the above description, it can be seen that the computer-readable storage medium provided in the embodiment of the present application collects historical form usage data, determines the corresponding form features, inputs the form features into the first long short-term memory layer of the preset time series model for preliminary feature extraction, inputs the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model for deep feature extraction, determines the corresponding deep features, determines the corresponding time series model according to the deep features, receives the user's modification operation on the front-end form, determines the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, determines whether the form performance has decreased according to the form performance prediction result and the preset form performance baseline, and if it has decreased, uses lazy loading technology to perform resource optimization and adjustment operations on the front-end form, thereby improving the efficiency and flexibility of the front-end form based on predictive performance optimization.
[0235] The embodiments of the present application also provide a computer program product capable of implementing all steps of the front-end form data maintenance 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 maintenance method are implemented. For example, the computer program / instruction implements the following steps:
[0236] Step S101: collecting historical form usage data, performing feature evaluation operations on the historical form usage data according to preset feature correlation analysis rules, and determining corresponding form features;
[0237] Step S102: input the form features into the first long short-term memory layer of the preset time series model to perform a preliminary feature extraction operation, input the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model to perform a deep feature extraction operation, determine the corresponding deep features, determine the corresponding loss function according to the deep features, adjust the weight of the preset time series model according to the minimum loss function and the back propagation algorithm, and determine the corresponding time series model, wherein the first long short-term memory layer and the second long short-term memory layer are connected by a residual connection;
[0238] Step S103: Receive the user's modification operation on the front-end form, determine the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, and judge whether the form performance has declined according to the form performance prediction result and the preset form performance baseline. If declined, use lazy loading technology to perform resource optimization and adjustment operations on the front-end form.
[0239] From the above description, it can be seen that the computer program product provided by the embodiment of the present application collects historical form usage data, determines the corresponding form features, inputs the form features into the first long short-term memory layer of the preset time series model for preliminary feature extraction operations, inputs the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model for deep feature extraction operations, determines the corresponding deep features, determines the corresponding time series model according to the deep features, receives the user's modification operation on the front-end form, determines the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, determines whether the form performance has decreased according to the form performance prediction result and the preset form performance baseline, and if it has decreased, uses lazy loading technology to perform resource optimization and adjustment operations on the front-end form, thereby improving the efficiency and flexibility of the front-end form based on predictive performance optimization.
[0240] 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.
[0241] 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.
[0242] 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.
[0243] 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.
[0244] 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 maintenance method, characterized in that: The method comprises: Collect historical form usage data, perform feature evaluation operations on the historical form usage data according to preset feature correlation analysis rules, and determine corresponding form features; Input the form features into the first long short-term memory layer of the preset time series model for preliminary feature extraction, input the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model for deep feature extraction, determine the corresponding deep features, determine the corresponding loss function according to the deep features, adjust the weight of the preset time series model according to the minimum loss function and the back propagation algorithm, and determine the corresponding time series model, wherein the first long short-term memory layer and the second long short-term memory layer are connected by a residual connection; Receive the user's modification operation on the front-end form, determine the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, and judge whether the form performance has declined according to the form performance prediction result and a preset form performance baseline. If declined, use lazy loading technology to perform resource optimization and adjustment operations on the front-end form.
2. The front-end form data maintenance method according to claim 1, characterized in that: The collecting of historical form usage data, performing feature evaluation operations on the historical form usage data according to preset feature correlation analysis rules, and determining corresponding form features, includes: Collect historical form data, perform data cleaning operations on the historical form data, and determine corresponding clean form data; A feature evaluation operation is performed on the clean form data according to a preset correlation coefficient calculation technology and a feature selection algorithm to determine corresponding form features.
3. The front-end form data maintenance method according to claim 2 is characterized in that: The step of performing a feature evaluation operation on the clean form data according to a preset correlation coefficient calculation technique and a feature selection algorithm to determine corresponding form features includes: Performing a correlation coefficient calculation operation on the clean form data according to a preset Spearman rank correlation coefficient rule to determine corresponding correlation features; The corresponding form features are determined according to the relevant features and the recursive feature elimination algorithm.
4. The front-end form data maintenance method according to claim 1, characterized in that: The step of inputting the form features into the first long short-term memory layer of the preset time series model for preliminary feature extraction includes: Inputting the form feature into the input gate of the first long short-term memory layer of the preset time series model to determine the corresponding input feature; Inputting the input feature into the forget gate of the first long short-term memory layer to determine the corresponding retained feature; The retained features are input into the output gate of the first long short-term memory layer to determine corresponding preliminary features.
5. The front-end form data maintenance method according to claim 1, characterized in that: The step of inputting the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model to perform a deep feature extraction operation, determining the corresponding deep features, and determining the corresponding loss function according to the deep features includes: Inputting the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model to perform a deep feature extraction operation to determine the corresponding deep features; The deep features are input into the output layer of a preset time series model, and the corresponding loss function is determined according to the predicted features output by the output layer and the preset real features.
6. The front-end form data maintenance method according to claim 1, characterized in that: After inputting the preliminary features output by the first long short-term memory layer into the second long short-term memory layer of the preset time series model to perform a deep feature extraction operation, determining the corresponding deep features, and determining the corresponding loss function according to the deep features, the method further includes: Determining a corresponding regularization term according to the parameters of the loss function and a preset regularization coefficient; The loss function is updated according to the regularization term, and the minimum loss function is determined according to the updated loss function and a preset gradient descent algorithm.
7. The front-end form data maintenance method according to claim 1, characterized in that: The determining whether the form performance has decreased according to the form performance prediction result and the preset form performance baseline, and if decreased, performing resource optimization and adjustment operations on the front-end form using lazy loading technology, includes: Determine whether the form performance has decreased according to the form performance prediction result and the preset form performance baseline, and if so, perform a scroll event monitoring operation to determine the corresponding optimal resource loading time; A resource optimization adjustment operation is performed according to the optimal resource loading time.
8. A front-end form data maintenance device, characterized in that: The device comprises: A form feature extraction module is used to collect historical form usage data, perform feature evaluation operations on the historical form usage data according to preset feature correlation analysis rules, and determine corresponding form features; A model training module, used for inputting the form features into a first long short-term memory layer of a preset time series model for preliminary feature extraction, inputting the preliminary features output by the first long short-term memory layer into a second long short-term memory layer of the preset time series model for deep feature extraction, determining corresponding deep features, determining corresponding loss functions according to the deep features, adjusting the weights of the preset time series model according to the minimum loss function and a back propagation algorithm, and determining a corresponding time series model, wherein the first long short-term memory layer and the second long short-term memory layer are connected via a residual connection; The prediction and maintenance optimization module is used to receive the user's modification operation on the front-end form, determine the form performance prediction result corresponding to the modification operation of the front-end form according to the time series model, and judge whether the form performance has declined according to the form performance prediction result and the preset form performance baseline. If it has declined, use lazy loading technology to perform resource optimization and adjustment operations on the front-end form.
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 maintenance method according to 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 maintenance method described in any one of claims 1 to 7 are implemented.
