Resource allocation approval method, device, equipment, medium and program product

Abstract

The invention provides a resource allocation approval method which can be applied to the technical field of artificial intelligence or the field of financial science and technology. The method comprises the following steps: in response to a resource allocation approval request, calling an employee historical resource allocation value and employee attribute data; the resources represent computing resources of terminals used by the employees; performing resource allocation value prediction according to the employee historical resource allocation value and the employee attribute data through a pre-trained resource allocation value prediction model to output a predicted resource allocation value with resource allocation value details and a group resource allocation condition; and performing data comparison based on the predicted resource allocation value, the group resource allocation condition and a pre-acquired actual calculation resource allocation value to form a resource allocation approval report. The invention further provides a resource allocation approval device and equipment, a storage medium and a program product.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence, specifically to data approval methods, and more specifically to a resource allocation approval method, apparatus, equipment, medium, and program product. Background Technology

[0002] Currently, the allocation of resources by large enterprises requires processing, verification and approval by multiple relevant personnel from the human resources and finance departments. The extent of resource allocation for an employee is related to their specific educational background, work tasks and processes, and the data involved in their work.

[0003] However, existing methods for processing resource allocation values ​​either involve using objective computer programs to calculate an employee's resource allocation or using neural networks to determine the final terminal computing power. The former overlaps with the company's own resource allocation estimation methods and cannot confirm whether the employee's resource allocation plan is appropriate. Although the latter uses different technologies to verify whether the employee's resource allocation value is calculated correctly, it involves very little data and only outputs a final amount. It cannot intuitively obtain detailed information on the allocation of computing power for resources such as chips, databases, hard disk storage space, and cloud storage. As a result, some terminal computing power remains idle and wasted, while other terminal computing power is scarce and causes downtime. Summary of the Invention

[0004] In view of the above issues, this application provides resource allocation approval methods, apparatus, equipment, media and procedures.

[0005] According to a first aspect of this application, a resource allocation approval method is provided, comprising: responding to a resource allocation approval request by retrieving historical resource allocation values ​​and employee attribute data; the resource characterizing the computing resources possessed by the terminal used by the employee; predicting resource allocation values ​​based on historical resource allocation values ​​and employee attribute data using a pre-trained resource allocation value prediction model, to output a predicted resource allocation value with detailed resource allocation values ​​and a group resource allocation status; and comparing the predicted resource allocation value, the group resource allocation status, and the pre-acquired actual computing resource allocation value to form a resource allocation approval report.

[0006] According to embodiments of this application, in response to a resource allocation approval request, retrieving historical resource allocation values ​​and employee attribute data includes: parsing the received resource allocation approval request to obtain the associated employee; retrieving the associated employee's department, job type, and job level as a first employee attribute, and sending the first employee attribute to the requester who issued the resource allocation approval request so that the requester can verify the associated employee's identity; designating the verified associated employee as the target employee, receiving the identity verification receipt and identity key from the requester, and retrieving the target employee's historical resource allocation values ​​and employee attribute data based on the identity verification receipt and identity key; wherein...

[0007] The historical resource allocation value of employees includes the computing power data value of the historical terminals used by employees; the computing power data value is obtained by weighted summation and standardization of chip data, database data, hard disk storage space data, and cloud storage data.

[0008] According to an embodiment of this application, retrieving employee attribute data includes: obtaining department transfer data and job adjustment data of the target employee as dynamic attributes; obtaining the target employee's years of service data, educational background data, and work record data as second employee attributes, and using the first employee attribute and the second employee attribute as static attributes; and summarizing the data in the dynamic attributes and static attributes to form employee attribute data.

[0009] According to an embodiment of this application, before pre-training the resource allocation value prediction model, the method includes: acquiring employee attribute data of all employees within the target enterprise as a basic dataset; in response to the presence of non-continuous data in the basic dataset within a preset time interval, filling the gaps in the non-continuous data using the median within the time interval to form a continuous dataset; performing deduplication and standardization operations on the continuous dataset to form a standard dataset; and classifying and refining the standard dataset based on a preset overall classification to obtain sample data including the static attributes, dynamic attributes, and historical resource allocation value data of all employees.

[0010] According to an embodiment of this application, pre-training a resource allocation value prediction model includes: repeatedly training a preset recurrent neural network architecture using sample data based on preset constraint prompts, so that the recurrent neural network architecture outputs training prediction individual resource allocation value entries and training prediction group resource allocation data corresponding to each employee in the sample data; calculating a first weighted mean square error between the training prediction individual resource allocation value entries and the pre-acquired current resource allocation value data of employees, calculating a second weighted mean square error between the training prediction group resource allocation data and the pre-acquired current data of the employee group, and performing a weighted summation of the first weighted mean square error and the second weighted mean square error to obtain the loss function value of the recurrent neural network architecture; wherein, when the loss function value is lower than a preset loss threshold, training is stopped, and the optimal recurrent neural network architecture under the constraint prompts is used as the resource allocation value prediction model; wherein, the current data of the employee group is obtained by classifying and averaging the current resource allocation value data of employees.

[0011] According to embodiments of this application, the recurrent neural network architecture outputs training prediction individual resource allocation value entries and training prediction group resource allocation data corresponding to each employee in the sample data, including: extracting features from employee historical resource allocation value data, static attributes, and dynamic attributes respectively to obtain time series features, static features, and dynamic features; fusing the time series features, static features, and dynamic features to obtain fused features; using a preset individual task head algorithm to perform feature mapping on the fused features to obtain training prediction individual resource allocation value entries; and using a preset group task head algorithm to perform category feature mapping on the fused features according to the categories of the first employee attribute and the second employee attribute to obtain training prediction group resource allocation data corresponding to each category of the first employee attribute and the second employee attribute.

[0012] According to an embodiment of this application, the resource allocation value prediction model is a large model. This large model predicts resource allocation values ​​based on employees' historical resource allocation values ​​and employee attribute data, outputting predicted resource allocation values ​​and group resource allocation status with detailed resource allocation information. The model includes: extracting features based on the target employee's historical resource allocation values ​​and employee attribute data according to constraint prompts, to obtain historical resource allocation value features and attribute features; fusing the historical resource allocation value features and attribute features to obtain employee features; using an individual task head algorithm to perform feature mapping on the employee features based on preset resource allocation value detail categories, to obtain predicted resource allocation values ​​with detailed resource allocation information; and using a group task head algorithm to perform category feature mapping on the employee features according to the categories of a first employee attribute and a second employee attribute, to obtain the group resource allocation status corresponding to each category of the first and second employee attributes.

[0013] According to embodiments of this application, constraint prompts represent constraints on historical resource allocation value features and attribute features, enabling large models to enhance the extraction and fusion of corresponding predicted resource allocation values ​​and group resource allocation status based on historical resource allocation value features and attribute features.

[0014] According to an embodiment of this application, a resource allocation approval report is generated by comparing data based on predicted resource allocation values, group resource allocation status, and pre-obtained actual calculated resource allocation values. This includes: processing the group resource allocation status to obtain a group resource allocation statistics table corresponding to the target employee; wherein the group resource allocation statistics table uses each category of the first and second employee attributes as its header, and is obtained by matching the group resource allocation status with the header data; comparing the actual calculated resource allocation value with individual data in the group resource allocation statistics table to obtain an external comparison ratio; comparing the actual calculated resource allocation value with the predicted resource allocation value to obtain an internal ratio; performing a four-quadrant analysis based on the external and internal ratios to obtain resource allocation value results; and performing a horizontal approval analysis based on the resource allocation value results to generate a resource allocation approval report for the target employee.

[0015] A second aspect of this application provides a resource allocation approval device, comprising: an approval response module, used to respond to a resource allocation approval request by calling historical resource allocation values ​​and employee attribute data; a resource representation of the computing resources possessed by the terminal used by the employee; a resource allocation value prediction module, used to predict resource allocation values ​​based on historical resource allocation values ​​and employee attribute data using a pre-trained resource allocation value prediction model, to output a predicted resource allocation value with detailed resource allocation values ​​and a group resource allocation status; and a resource allocation value analysis module, used to compare data based on the predicted resource allocation value, the group resource allocation status, and the pre-acquired actual computing resource allocation value, to generate a resource allocation approval report.

[0016] A third aspect of this application provides an electronic device comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method described above.

[0017] A fourth aspect of this application also provides a computer-readable storage medium having a computer program or instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method.

[0018] The fifth aspect of this application also provides a computer program product, including a computer program or instructions that, when executed by a processor, implement the steps of the above-described method.

[0019] One or more of the above embodiments have the following beneficial effects: A pre-trained resource allocation value prediction model predicts resource allocation values ​​based on employees' historical resource allocation values ​​and employee attribute data, outputting predicted resource allocation values ​​with detailed resource allocation information and group resource allocation status. Then, based on this predicted resource allocation value, the group resource allocation status, and the pre-acquired actual calculated resource allocation value, a data comparison is performed to generate a resource allocation approval report. Thus, the resource allocation value prediction model generates predicted resource allocation values ​​adapted to the employee's situation and the group to which the employee belongs, based on the employee's historical resource allocation values ​​and employee attributes. The appropriate group resource allocation not only predicts the individual resource allocation value of an employee, but also the group resource allocation that is appropriate for that employee. This improves the generalization ability and output coverage of the resource allocation value prediction model. Based on this, comparison and approval can improve the comprehensiveness of resource allocation comparison and the objectivity of approval reference. In this way, the human resources department can accurately confirm whether the employee's resource allocation value is calculated accurately and objectively grasp whether the employee's resource allocation is consistent with his or her own position. This avoids the phenomenon of some terminal computing power being idle and wasted, while other terminal computing power is scarce and causes downtime, thereby improving the overall utilization rate of enterprise resources. Attached Figure Description

[0020] The above-mentioned contents, other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0021] Figure 1 The illustrations depict application scenarios of the resource allocation approval method, apparatus, device, medium, and program products according to embodiments of this application.

[0022] Figure 2 A flowchart illustrating a resource allocation approval method according to an embodiment of this application is shown schematically.

[0023] Figure 3 This illustration schematically shows a functional architecture diagram involved in the resource allocation approval method according to an embodiment of this application;

[0024] Figure 4 This diagram illustrates the key operational steps involved in the resource allocation approval method according to an embodiment of this application.

[0025] Figure 5 This schematically illustrates a structural block diagram of a resource allocation approval device according to an embodiment of this application; and

[0026] Figure 6 A block diagram schematically illustrates an electronic device suitable for implementing a resource allocation approval method according to an embodiment of this application. Detailed Implementation

[0027] The embodiments of this application will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of this application. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of this application for ease of explanation. However, it will be apparent that one or more embodiments may be implemented without these specific details. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concepts of this application.

[0028] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0029] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0030] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0031] In the technical solution of this application, the data related to users / employees is collected, stored, used, processed, transmitted, provided, disclosed and applied with the authorization of users / employees, and all such processing complies with the provisions of relevant laws and regulations, necessary confidentiality measures have been taken, and does not violate public order and good morals.

[0032] In existing resource allocation value processing methods, either the resource allocation for an employee is calculated in summary by an objective computer program, or the final terminal computing power is determined by a neural network. The former overlaps with the enterprise's own resource allocation estimation method and cannot confirm whether the employee's resource allocation plan is appropriate. Although the latter uses different technologies to verify whether the employee's resource allocation value is calculated correctly, it involves very little data and only outputs a final amount. It cannot intuitively obtain specific details of the allocated computing power of resources such as chips, databases, hard disk storage space, and cloud storage resources. As a result, some terminal computing power is still idle and wasted, while other terminal computing power is scarce and causes downtime.

[0033] This application provides a resource allocation approval method, apparatus, device, medium, and program product. The resource allocation approval method includes: responding to a resource allocation approval request by retrieving historical resource allocation values ​​and employee attribute data; a resource characterization representing the computing resources possessed by the terminal used by the employee; predicting resource allocation values ​​based on historical resource allocation values ​​and employee attribute data using a pre-trained resource allocation value prediction model, to output a predicted resource allocation value with detailed resource allocation information and a group resource allocation status; and comparing the predicted resource allocation value, the group resource allocation status, and the pre-acquired actual computing resource allocation value to generate a resource allocation approval report.

[0034] Figure 1 The diagram illustrates an application scenario of the resource allocation approval method according to an embodiment of this application.

[0035] like Figure 1 As shown, the application scenario 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105.

[0036] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 via the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as web browser applications, search applications, instant messaging tools, email clients, social media platform software, etc.

[0037] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0038] Network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables. Network 104 serves as the medium for providing communication links between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. Network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables.

[0039] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0040] It should be noted that the resource allocation approval method provided in this application embodiment can generally be executed by server 105. Correspondingly, the resource allocation approval device provided in this application embodiment can generally be located in server 105. The resource allocation approval method provided in this application embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Correspondingly, the resource allocation approval device provided in this application embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105.

[0041] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0042] The following will be based on Figure 1 The described scene, through Figures 2-4 The resource allocation approval method according to the embodiments of this application will be described in detail.

[0043] Figure 2 A flowchart illustrating a resource allocation approval method according to an embodiment of this application is shown schematically.

[0044] like Figure 2 As shown, the resource allocation approval method of this embodiment includes S210~S230, as detailed below:

[0045] Operation S210 responds to the resource allocation approval request by calling the employee's historical resource allocation values ​​and employee attribute data; resources represent the computing resources available on the terminal used by the employee; operation S220 uses a pre-trained resource allocation value prediction model to predict resource allocation values ​​based on the employee's historical resource allocation values ​​and employee attribute data, and outputs predicted resource allocation values ​​with detailed resource allocation values ​​and group resource allocation status; operation S230 compares the predicted resource allocation values, group resource allocation status, and pre-acquired actual computing resource allocation values ​​to form a resource allocation approval report.

[0046] As an example, the resource allocation approval request can be initiated by a pre-set approver terminal, which is a personal computer. Upon receiving the request, the approver first identifies the target employee, then retrieves the employee's historical resource allocation values, such as those over the past five years, and also retrieves the employee's attribute data, such as their current job level, years of service, department (business / functional), educational background, and work history. Finally, the employee's historical resource allocation values ​​and attribute data are input into a pre-trained... In the resource allocation value prediction model, the predicted resource allocation value for the target employee and the group's resource allocation status are output. If the employee's historical resource allocation value is the same, it is obtained by weighted summation of chip data, database data, hard disk storage space data, and cloud storage data. The specific weighted summation method is not limited; for example, different weights can be assigned to chip data, database data, hard disk storage space data, and cloud storage data respectively. Then, the chip parameters are multiplied by the chip weight, the database parameters are multiplied by the database weight, and the hard disk storage space size is multiplied by the hard disk weight. Adding the cloud storage space size multiplied by the cloud storage weight yields the final summation. This summation is then standardized to keep the resource allocation value below 100 points for easier comparison later. For example, if the predicted resource allocation value is 90 points, the overall resource allocation status shows the average resource allocation value for the target employee is 91 points, and the actual calculated resource allocation value for the current year / month is 72 points, then a data comparison is performed based on the predicted resource allocation value, the group resource allocation status, and the pre-obtained actual calculated resource allocation value to generate a resource allocation approval report. This report displays the actual calculated resource allocation. The allocated value is close to the predicted resource allocation value, so its calculation is likely correct. However, since the actual calculated resource allocation value (predicted resource allocation value) differs significantly from the average resource allocation value of the target employee in the group, the resource allocation approval report can show that the target employee's personal resource allocation value is lower than the average level of his / her group. Based on this resource allocation approval report, the human resources department can understand whether the employee's resource allocation is consistent with his / her own positioning, thereby avoiding the phenomenon of some terminal computing power being idle and wasted, while other terminal computing power is scarce and causes downtime, thus improving the overall utilization rate of enterprise resources.

[0047] Therefore, the resource allocation value prediction model generates a predicted resource allocation value that matches the employee's situation based on the employee's historical resource allocation value and employee attribute data, as well as a group resource allocation status that matches the employee's group. Based on this, a resource allocation approval report for the employee is generated. This makes it easier for human resources departments to understand whether the employee's resource allocation is accurate and whether the employee's resource allocation is consistent with their own positioning. It also makes it easier to control the matching degree between the computing power of the terminal used by the employee and their workload. This avoids the phenomenon of some terminal computing power being idle and wasted, while other terminal computing power is scarce and causes downtime. It improves the overall utilization rate of enterprise resources and enhances the comfort of employees in the office.

[0048] Figure 3 A schematic diagram of the functional architecture involved in the resource allocation approval method of this embodiment is shown.

[0049] like Figure 2 , Figure 3 As shown in the figure, the resource allocation approval method in this embodiment is based on a preset resource allocation value management system server, which includes a data storage device for storing employees' historical resource allocation values ​​and employee attribute data, a main processing unit that executes operations S210-S230, a maintenance terminal for maintaining the resource allocation value management system server, and an approver terminal for initiating resource allocation approval requests. The maintenance terminal communicates with the main processing unit via a preset resource allocation value maintenance interface and a model maintenance interface through the network. The approver terminal initiates resource allocation approval requests to the main processing unit via the resource allocation approval interface through the network.

[0050] In this embodiment, after the main processing unit in the resource allocation value management system server receives the resource allocation approval request initiated by the approver's terminal, it executes operation S210 in response to the resource allocation approval request, calling the employee's historical resource allocation value and employee attribute data. In this embodiment, this includes parsing the received resource allocation approval request to obtain the associated employee of the resource allocation approval request; calling the associated employee's department, job type, and job level as the first employee attribute, and sending the first employee attribute to the requester who issued the resource allocation approval request so that the requester can verify the identity of the associated employee; taking the associated employee whose identity has been verified as the target employee, receiving the identity verification receipt and identity key issued by the requester, and retrieving the target employee's historical resource allocation value and employee attribute data based on the identity verification receipt and identity key; the employee's historical resource allocation value includes the computing power data value of the historical terminal used by the employee; the computing power data value is obtained by weighted summation and standardization of chip data, database data, hard disk storage space data, and cloud storage data.

[0051] As an example, when responding to an approval request, it's necessary to first identify the specific target employee. Therefore, the received resource allocation approval request needs to be parsed to obtain the associated employees involved. However, in large enterprises, employee names are often duplicated. If multiple associated employees are involved, it affects the accuracy of subsequent predictions and approvals. Furthermore, employee resource allocation values ​​are highly privacy-sensitive data, so a verification mechanism is required. This ensures that the entire resource allocation value management service system can only obtain the information / data of the individual employee whose resource allocation approval request is actually being approved. Therefore, in this example, after obtaining the associated employees for the resource allocation approval request, it's necessary to first retrieve the associated employee's department, job type, and position. The system uses the employee level as the primary employee attribute and sends this attribute to the requester who issued the resource allocation approval request. This allows the requester to verify the identity of the associated employee. After receiving the primary employee attribute, the requester can easily determine which employee is the specific one they want to approve. Therefore, the requester will select the actual target employee from the associated employees and then send the target employee's confirmation information (confirmation receipt) and identity key back to the main processing unit. The main processing unit then uses the associated employee corresponding to the confirmation information as the target employee, receives the identity confirmation receipt and identity key from the requester, and retrieves the target employee's historical resource allocation value and employee attribute data based on the identity confirmation receipt and identity key.

[0052] In this way, the associated employees are identified based on the resource allocation approval request, and then the requester confirms the identity of the associated employees to determine the target employee, thereby avoiding the risk of misoperation due to duplicate names. Moreover, information such as resource allocation value attributes can only be retrieved after obtaining the identity key, thus ensuring the privacy of employee information.

[0053] In this embodiment, retrieving employee attribute data includes: obtaining the target employee's department transfer data and job adjustment data as dynamic attributes; obtaining the target employee's years of service data, educational background data, and work record data as second employee attributes; and using the first and second employee attributes as static attributes; and summarizing the data in the dynamic and static attributes to form employee attribute data.

[0054] As an example, static attributes are those that don't change easily, or whose changes are orderly, such as years of service. These increase year by year, while department, job type, and job level generally remain constant. Like years of service, they are adjusted in an orderly manner. Often, the longer the years of service, the more memory space is needed; simply increasing the terminal's storage space periodically is sufficient. Dynamic attributes, on the other hand, reflect the employee's dynamic processes, such as departmental transfers and job adjustments. This type of employee attribute data can almost cover all the key data of the target employee's entire career. If dynamic attributes change, resource allocation is often significantly affected. For example, the original terminal data may need to be erased, and a new resource allocation strategy needs to be planned for the new job, such as allocating computer with which type of chip, database type, and the amount of hard drive and cloud storage space. Using such comprehensive employee attribute data as input to a subsequent resource allocation prediction model can improve the model's generalization ability, thereby improving the accuracy and relevance of the predicted resource allocation values ​​and group resource allocation status.

[0055] Figure 4 This diagram illustrates the key operational steps of the resource allocation value prediction model involved in the resource allocation approval method of this embodiment.

[0056] like Figure 4 As shown, in this embodiment, before operating S220, it is necessary to set the relevant attributes of the resource allocation value prediction model, collect data samples and clean data, etc., and then establish the model (resource allocation value prediction model), and train and optimize the reverse algorithm. Then, the trained resource allocation value prediction model is used to predict the resource allocation value of employees, etc. Then, the prediction results (predicted resource allocation value, group resource allocation status) are compared with the actual resource allocation value (actual calculated resource allocation value) to form a resource allocation approval report.

[0057] That is, before pre-training the resource allocation value prediction model, the process includes: acquiring employee attribute data of all employees in the target enterprise as a base dataset; in response to the presence of non-continuous data in the base dataset within a preset time interval, filling the gaps in the non-continuous data using the median within the time interval to form a continuous dataset; performing deduplication and standardization operations on the continuous dataset to form a standard dataset; and classifying and refining the standard dataset based on a preset overall classification to obtain sample data including the static attributes, dynamic attributes, and historical resource allocation value data of all employees.

[0058] In this embodiment, historical resource allocation data, job level, years of service, department attributes (business / functionality), educational background, annual performance evaluation, etc., of employees can be obtained from the data storage device and processed to form a standard dataset. The specific steps are as follows: Set relevant static attributes for the model, including job level, years of service, department attributes (business / functionality), optional educational background, work records, etc.; record dynamic characteristics of employees, including department transfers and job adjustments. Obtain employee resource allocation value distribution records for the past several years from the system, including the employee's current job level, date of employment, department attributes, educational background, annual performance evaluation, etc., and call pandas functions for data cleaning. The processing steps include: handling missing values ​​by filling discontinuous parts of resource allocation value distribution with the column median; handling duplicate values ​​by deleting all rows with identical columns; handling outliers by setting boundary values ​​for resource allocation values, such as [100, 1000000], and replacing outlier data with boundary values, thereby forming a standard dataset. In addition, group data can be obtained by dividing the entire population into groups. When training the resource allocation value prediction model, this data can be used as the current data of the employee group, which facilitates the calculation of the weighted mean square error when the resource allocation value prediction model outputs the training prediction data of the group's resource allocation. That is, the standard dataset is classified and refined based on the preset group division categories to obtain sample data including the static attributes, dynamic attributes and historical resource allocation value data of all employees. The specific classification process is as follows: for example, work experience is divided into one year as a unit; job level is divided into job level codes (1, P1; 2, P2; 3, P3); the average resource allocation value of the group is calculated by dividing the resource allocation of the group with the same work experience and job level at a specific time by the number of employees in that group at that specific time.

[0059] Therefore, by performing gap filling, deduplication, and standardization operations through continuity judgment before pre-training the resource allocation value prediction model, the comprehensiveness and standardization of the training sample set are improved, thereby enhancing the continuity and standardization of the sample set and ultimately improving the generalization ability and prediction accuracy of the trained resource allocation value prediction model.

[0060] In this embodiment, the pre-training of the resource allocation value prediction model includes: based on preset constraint prompts, repeatedly training a preset recurrent neural network architecture using sample data, so that the recurrent neural network architecture outputs training prediction individual resource allocation value entries and training prediction group resource allocation data corresponding to each employee in the sample data; calculating the first weighted mean square error between the training prediction individual resource allocation value entries and the pre-acquired current resource allocation value data of employees, calculating the second weighted mean square error between the training prediction group resource allocation data and the pre-acquired current data of the employee group, and performing a weighted summation of the first weighted mean square error and the second weighted mean square error to obtain the loss function value of the recurrent neural network architecture; wherein, when the loss function value is lower than a preset loss threshold, training is stopped, and the optimal recurrent neural network architecture under the constraint prompts is used as the resource allocation value prediction model; wherein, the current data of the employee group is obtained by classifying and averaging the current resource allocation value data of employees.

[0061] As an example, the process of pre-training the resource allocation value prediction model is the process of building the model (resource allocation value prediction model), as well as the process of training and optimizing the inverse algorithm. Based on the preset constraint prompts, the preset recurrent neural network architecture is repeatedly trained using sample data so that the output of the recurrent neural network architecture corresponds to the training prediction individual resource allocation value entries and training prediction group resource allocation data for each employee in the sample data. Then, the inverse optimization is performed based on the loss function value of the recurrent neural network architecture. When the loss function value is lower than the preset loss threshold, the training stops, and the optimal recurrent neural network architecture under the constraint prompts is used as the resource allocation value prediction model. In this example, the recurrent neural network architecture uses an LSTM model. The data from the first three years of the nearly five-year data set can be divided into a training set, a validation set (year four), and a test set (year five). Time-series cross-validation is then performed, allowing the LSTM model to perform individual and group predictions during repeated training. This yields training predictions of individual resource allocation values ​​and group resource allocation data. Furthermore, in this example, the data generated from individual predictions can be corrected based on the group prediction data using the group standard deviation to form the final training predictions of individual resource allocation values. To ensure continuous optimization of the model, the LSTM model is continuously trained and tuned each year when the latest resource allocation data is obtained, ensuring that the generated resource allocation prediction model remains in optimal condition.

[0062] Therefore, this resource allocation value prediction model essentially generates weighted mean square errors based on the predicted individual resource allocation value entries and the predicted group resource allocation data, and then sums the two weighted mean square errors in a weighted manner. This ensures that the resource allocation value prediction model can perform dual-thread prediction in the optimal state, predicting not only the individual resource allocation value of the employee, but also the resource allocation value that the group to which the employee belongs should have. This lays the foundation for comprehensively locating the employee's resource allocation value through the resource allocation value prediction model in the subsequent use stage.

[0063] In this embodiment, the recurrent neural network architecture outputs training prediction individual resource allocation value entries and training prediction group resource allocation data corresponding to each employee in the sample data. This includes: extracting features from the employee's historical resource allocation value data, static attributes, and dynamic attributes to obtain time-series features, static features, and dynamic features; fusing the time-series features, static features, and dynamic features to obtain fused features; using a preset individual task head algorithm to perform feature mapping on the fused features to obtain training prediction individual resource allocation value entries; and using a preset group task head algorithm to perform category feature mapping on the fused features according to the categories of the first employee attribute and the second employee attribute to obtain training prediction group resource allocation data corresponding to each category of the first employee attribute and the second employee attribute.

[0064] As an example, the resource allocation value sequence of the past 12 months can be input into the input layer of this recurrent neural network architecture to obtain a time series. Information such as job level, department attribute, years of service, department transfer data, and job adjustment data can be input into this input layer to obtain static and dynamic information. The shared layer of this recurrent neural network architecture then uses LSTM to extract features from the time series, static, and dynamic information to obtain temporal features, static features, and dynamic features. These features are then fused through the fusion layer of the recurrent neural network architecture to obtain fused features. Subsequently, a dual-task head is used for individual prediction and group prediction, including an individual task head using an individual task head algorithm and a group task head using a group task head algorithm. This yields training prediction entries for individual resource allocation values ​​corresponding to the employee's individual status, and training prediction group resource allocation data corresponding to the categories of the employee's first and second employee attributes. Based on this, a loss function composed of weighted mean squared error is obtained. When calculating this loss function, the weight corresponding to the individual task head can be set to 0.7 to prioritize individual prediction accuracy.

[0065] Based on this, the two independent task head algorithms enable the trained resource allocation value prediction model to automatically output predicted resource allocation values ​​with detailed resource allocation information based on the employee's resource allocation value attribute information, and the data flow of outputting individual resource allocation and group resource allocation values ​​does not affect each other, thereby improving the rationality and accuracy of subsequent employee resource allocation value positioning and approval.

[0066] In this embodiment, the resource allocation value prediction model is a large-scale model. This model predicts resource allocation values ​​based on employees' historical resource allocation values ​​and employee attribute data, outputting predicted resource allocation values ​​and group resource allocation status with detailed resource allocation information. The process includes: extracting features based on the target employee's historical resource allocation values ​​and employee attribute data using constraint prompts to obtain historical resource allocation value features and attribute features; fusing these features to obtain employee features; using an individual task head algorithm to perform feature mapping on the employee features based on preset resource allocation value detail categories to obtain predicted resource allocation values ​​with detailed resource allocation information; and using a group task head algorithm to perform category feature mapping on the employee features according to the categories of the first and second employee attributes to obtain the group resource allocation status corresponding to each category of the first and second employee attributes.

[0067] As an example, the trained large model can be directly applied, and its application process is similar to the training process. First, the input layer and shared layer of the trained large model extract features based on the target employee's historical resource allocation values ​​and employee attribute data to obtain historical resource allocation value features and attribute features. Then, based on the fusion layer of the large model, feature fusion is performed on the historical resource allocation value features and attribute features to obtain employee features. Then, based on the dual task head, the individual task head algorithm in the individual task head performs feature mapping on the employee features based on the preset resource allocation value detail categories to obtain the predicted resource allocation value with resource allocation value details. The group task head algorithm in the group task head performs category feature mapping on the employee features according to the categories of the first employee attribute and the second employee attribute to obtain the group resource allocation status corresponding to each category of the first employee attribute and the second employee attribute.

[0068] Therefore, this large model outputs predicted resource allocation values ​​with detailed resource allocation information and group resource allocation status corresponding to each category of employee attributes. Based on this, it can not only output predicted resource allocation values ​​that are suitable for the target employee's situation, but also group resource allocation status that is suitable for the employee's group. Based on this, a resource allocation approval report for the employee can be generated, allowing the human resources department to fully understand the employee's own resource allocation value structure and the reasonable resource allocation for the entire group to which the employee belongs. This makes it easier to confirm whether the employee's resource allocation is calculated accurately and whether the employee's resource allocation is consistent with their own positioning. This makes it easier to control the matching degree between the computing power of the terminal used by the employee and their own work situation, thereby improving the company's resource utilization rate and improving the employee's office comfort.

[0069] In this embodiment, the constraint prompt information represents the constraints on historical resource allocation value features and attribute features. This allows the large model to enhance the extraction and fusion of corresponding predicted resource allocation values ​​and group resource allocation status based on these historical resource allocation value features and attribute features. This enables the large model to better extract predicted resource allocation values ​​and group resource allocation status for target employees based on historical resource allocation value features and attribute features, thereby improving the prediction accuracy and precision of the large model.

[0070] In this embodiment, a resource allocation approval report is generated by comparing the predicted resource allocation value, the group resource allocation status, and the pre-acquired actual calculated resource allocation value. This includes: processing the group resource allocation status to obtain a group resource allocation statistics table corresponding to the target employee; wherein the group resource allocation statistics table uses each category of the first and second employee attributes as its header, and is obtained by matching the group resource allocation status with the header data; comparing the actual calculated resource allocation value with individual data in the group resource allocation statistics table to obtain an external comparison ratio; comparing the actual calculated resource allocation value with the predicted resource allocation value to obtain an internal ratio; performing a four-quadrant analysis based on the external and internal ratios to obtain the resource allocation value result; and performing a horizontal approval analysis based on the resource allocation value result to generate a resource allocation approval report for the target employee.

[0071] As an example, an external comparison ratio is obtained by comparing the actual computing resource allocation value with individual data in the group resource allocation statistics table. This can be achieved by dividing the actual computing resource allocation value by the individual data in the group resource allocation statistics table to obtain the individual external ratio. If the individual external ratio is equal to 1, it means that the actual computing resource allocation value of the target employee is completely consistent with the computing power resource level of its group. If the individual external ratio is greater than 1 (e.g., 1.1-1.2), it means that the actual computing resource allocation value of the target employee is higher than the computing power resource level of the employee's group, indicating that the target employee's terminal has surplus computing power. If the individual external ratio is less than 1 (e.g., 0.8-0.9), it means that the actual computing resource allocation value of the target employee is lower than the computing power resource level of the employee's group, thus indicating that the employee's terminal has limited computing power, which may affect its office comfort. The internal ratio is calculated by comparing the actual and predicted resource allocation values. This can be achieved by dividing the actual resource allocation value by the predicted value. An internal ratio of 1 indicates that the resource allocation aligns with the company's internal positioning. A ratio greater than 1 indicates that the resource allocation is higher than the company's expected computing power. A ratio less than 1 indicates that the resource allocation is lower than the company's expected computing power. When performing a four-quadrant analysis based on these external and internal ratios, Quadrant 1 can be defined as "high internally and high externally," meaning the resource allocation is higher than both the internal standard and the group level. Quadrant 2 can be defined as "low internally and high externally," meaning the resource allocation is lower than the internal standard but higher than the group level. Quadrant 3 can be defined as "low internally and low externally," meaning the resource allocation is lower than both the internal standard and the group level. Quadrant 4 can be defined as "high internally and low externally," meaning the resource allocation is higher than the internal standard but lower than the group level. Then, based on the quadrant in which the target employee is located and the corresponding level of their current resource allocation value, the situation of the target employee is analyzed based on this level, and a resource allocation approval report for the target employee is generated.

[0072] In this way, the internal ratio is obtained by comparing the predicted resource allocation value with the actual calculated resource allocation value. Then, a four-quadrant analysis is performed based on the external ratio and the internal ratio to generate a resource allocation approval report. This allows for strict approval of the accuracy of the employee's resource allocation value calculation and a comprehensive assessment of whether the employee's resource allocation is consistent with that of other groups with similar attributes. This facilitates control over the overall computing power allocation of the enterprise, improves the overall resource utilization rate of the enterprise, and enhances the overall office comfort of employees.

[0073] As described above, the resource allocation approval method provided in this embodiment uses the resource allocation value prediction model to generate a predicted resource allocation value that matches the employee's situation based on the employee's historical resource allocation value and employee attributes, as well as a group resource allocation status that matches the employee's group. Based on this, it not only predicts the employee's individual resource allocation value but also predicts the group resource allocation status, thereby improving the generalization ability and output coverage of the resource allocation value prediction model. This, in turn, improves the comprehensiveness of resource allocation comparison and the objectivity of approval reference. As a result, it is easier for the human resources department to accurately confirm whether the employee's resource allocation is calculated accurately and to objectively grasp whether the employee's resource allocation is consistent with their own positioning. This also makes it easier to control the matching degree between the computing power of the terminal used by the employee and the corresponding workload, improve the overall resource utilization rate and resource balance of the enterprise, and avoid the phenomenon that some terminals have idle computing power while others have strained computing power.

[0074] Based on the above-mentioned resource allocation approval method, this application also provides a resource allocation approval device. The following will be combined with... Figure 5 The device is described in detail.

[0075] Figure 5 A schematic diagram of the structure of a resource allocation approval device according to an embodiment of this application is shown.

[0076] like Figure 5 As shown, the resource allocation approval device 500 in this embodiment includes a request response module 510, a resource allocation value prediction module 520, and a report generation module 530.

[0077] The request response module 510 can perform operation S210 to respond to a resource allocation approval request by calling the employee's historical resource allocation value and employee attribute data; the resource represents the computing resources possessed by the terminal used by the employee.

[0078] The resource allocation value prediction module 520 can perform operation S220 to predict resource allocation values ​​based on employees' historical resource allocation values ​​and employee attribute data using a pre-trained resource allocation value prediction model, so as to output predicted resource allocation values ​​and group resource allocation status with resource allocation value details.

[0079] The report generation module 530 can perform operation S230 to compare data based on the predicted resource allocation value, the group resource allocation status, and the pre-acquired actual calculated resource allocation value to form a resource allocation approval report.

[0080] In this embodiment, the request response module 510 includes: a data parsing unit, used to parse the received resource allocation approval request to obtain the associated employee of the resource allocation approval request; an identity verification unit, used to retrieve the associated employee's department, job type, and job level as the first employee attribute, and send the first employee attribute to the requester who issued the resource allocation approval request so that the requester can verify the identity of the associated employee; and an information retrieval unit, used to select the associated employee whose identity has been verified as the target employee, receive the identity verification receipt and identity key issued by the requester, and retrieve the target employee's historical resource allocation based on the identity verification receipt and identity key. The data includes resource allocation values ​​and employee attribute data. Historical resource allocation values ​​for employees include the computing power data of the historical terminals used by the employees. Computing power data is obtained by weighted summation and standardization of chip data, database data, hard disk storage space data, and cloud storage data. Retrieving employee attribute data includes: obtaining the target employee's departmental transfer data and job adjustment data as dynamic attributes; obtaining the target employee's years of service data, educational background data, and work record data as secondary employee attributes, and using the primary and secondary employee attributes as static attributes; and summarizing the data from the dynamic and static attributes to form the employee attribute data.

[0081] In this embodiment, the resource allocation value prediction module 520 performs operations through a pre-trained resource allocation value prediction model, which is pre-trained using a preset model training unit. Before pre-training the resource allocation value prediction model, the process includes: acquiring employee attribute data for all employees within the target enterprise as a base dataset; filling in the gaps in the base dataset by using the median within the time interval to form a continuous dataset, given that the base dataset contains non-continuous data within a preset time interval; performing deduplication and standardization operations on the continuous dataset to form a standard dataset; and classifying and refining the standard dataset based on a preset overall classification to obtain sample data including static attributes, dynamic attributes, and historical resource allocation value data of all employees.

[0082] The model training unit pre-trains a resource allocation value prediction model, including: repeatedly training a pre-defined recurrent neural network architecture using sample data based on preset constraint prompts, so that the recurrent neural network architecture outputs training prediction individual resource allocation value entries and training prediction group resource allocation data corresponding to each employee in the sample data; calculating the first weighted mean square error between the training prediction individual resource allocation value entries and the pre-acquired current resource allocation value data of employees, and calculating the second weighted mean square error between the training prediction group resource allocation data and the pre-acquired current data of the employee group; weighted summing of the first weighted mean square error and the second weighted mean square error to obtain the loss function value of the recurrent neural network architecture; wherein, training stops when the loss function value is lower than a preset loss threshold, and the optimal recurrent neural network architecture under the constraint prompts is used as the resource allocation value prediction model; wherein, the employee The current data of the workforce is obtained by classifying and averaging the current resource allocation values ​​of employees. The recurrent neural network architecture outputs training prediction individual resource allocation value entries and training prediction group resource allocation data corresponding to each employee in the sample data. This includes: extracting features from historical resource allocation value data, static attributes, and dynamic attributes of employees to obtain time-series features, static features, and dynamic features; fusing the time-series features, static features, and dynamic features to obtain fused features; using a preset individual task head algorithm to perform feature mapping on the fused features to obtain training prediction individual resource allocation value entries; and using a preset group task head algorithm to perform category feature mapping on the fused features according to the categories of the first employee attribute and the second employee attribute to obtain training prediction group resource allocation data corresponding to each category of the first employee attribute and the second employee attribute.

[0083] In this embodiment, the resource allocation value prediction model is a large-scale model. This model predicts resource allocation values ​​based on employees' historical resource allocation values ​​and employee attribute data, outputting predicted resource allocation values ​​and group resource allocation status with detailed resource allocation information. This includes: extracting features based on the target employee's historical resource allocation values ​​and employee attribute data using constraint prompts to obtain historical resource allocation value features and attribute features; fusing these features to obtain employee features; using an individual task head algorithm to perform feature mapping on the employee features based on preset resource allocation value detail categories to obtain predicted resource allocation values ​​with detailed resource allocation information; and using a group task head algorithm to perform category feature mapping on the employee features according to the categories of a first employee attribute and a second employee attribute to obtain the group resource allocation status corresponding to each category of the first and second employee attributes. The constraint prompts represent the constraints on the historical resource allocation value features and attribute features, enabling the large-scale model to strengthen the extraction and fusion of predicted resource allocation values ​​and group resource allocation status based on these features.

[0084] The report generation module 530 includes: a statistical table generation unit, used to process data on the group resource allocation status to obtain a group resource allocation statistical table corresponding to the target employee; wherein, the group resource allocation statistical table uses the categories of the first employee attribute and the second employee attribute as table headers, and is obtained by matching the group resource allocation status with the table headers; a ratio calculation unit, used to compare the actual calculated resource allocation value with the individual data in the group resource allocation statistical table to obtain the external ratio, and to compare the actual calculated resource allocation value with the predicted resource allocation value to obtain the internal ratio; and an approval analysis unit, used to perform four-quadrant analysis based on the external ratio and the internal ratio to obtain the resource allocation value result, and to perform horizontal approval analysis based on the resource allocation value result to form a resource allocation approval report for the target employee.

[0085] Furthermore, according to embodiments of this application, any multiple modules among the request response module 510, resource allocation value prediction module 520, and report generation module 530 can be merged into one module, or any one of these modules can be split into multiple modules. Alternatively, at least some of the functions of one or more of these modules can be combined with at least some of the functions of other modules and implemented in one module. According to embodiments of this application, at least one of the request response module 510, resource allocation value prediction module 520, and report generation module 530 can be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or any other reasonable means of integrating or packaging circuitry, or implemented in software, hardware, or firmware, or in any suitable combination of any of these three implementation methods. Alternatively, at least one of the request response module 510, resource allocation value prediction module 520, and report generation module 530 can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is run.

[0086] It should be noted that the implementation methods, technical problems solved, functions achieved, and technical effects of each module in the device embodiment are the same as or similar to the implementation methods, technical problems solved, functions achieved, and technical effects of each corresponding step in the method embodiment, and will not be repeated here.

[0087] Figure 6 A block diagram schematically illustrates an electronic device suitable for implementing a resource allocation approval method according to an embodiment of this application.

[0088] like Figure 6 As shown, an electronic device 600 according to an embodiment of this application includes a processor 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage portion 608 into a random access memory (RAM) 603. The processor 601 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 601 may also include onboard memory for caching purposes. The processor 601 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of this application.

[0089] RAM 603 stores various programs and data required for the operation of electronic device 600. Processor 601, ROM 602, and RAM 603 are interconnected via bus 604. Processor 601 executes various operations of the method flow according to embodiments of this application by executing programs in ROM 602 and / or RAM 603. It should be noted that the programs may also be stored in one or more memories other than ROM 602 and RAM 603. Processor 601 may also execute various operations of the method flow according to embodiments of this application by executing programs stored in said one or more memories.

[0090] According to embodiments of this application, the electronic device 600 may further include an input / output (I / O) interface 605, which is also connected to a bus 604. The electronic device 600 may also include one or more of the following components connected to the input / output (I / O) interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the input / output (I / O) interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 610 as needed so that computer programs read from it can be installed into the storage section 608 as needed.

[0091] This application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of this application.

[0092] According to embodiments of this application, the computer-readable storage medium can be a non-volatile computer-readable storage medium, such as including but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this application, the computer-readable storage medium may include ROM 602 and / or RAM 603 and / or one or more memories other than ROM 602 and RAM 603 described above.

[0093] Embodiments of this application also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code enables the computer system to implement the resource allocation approval method provided in the embodiments of this application.

[0094] When the computer program is executed by the processor 601, it performs the functions defined in the system / apparatus of this application embodiment. According to the embodiments of this application, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0095] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and downloaded and installed via the communication section 609, and / or installed from the removable medium 611. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0096] In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 609, and / or installed from the removable medium 611. When the computer program is executed by the processor 601, it performs the functions defined in the system of this application embodiment. According to the embodiments of this application, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0097] According to embodiments of this application, program code for executing the computer programs provided in the embodiments of this application can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0098] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0099] Those skilled in the art will understand that the features described in the various embodiments of this application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this application. In particular, the features described in the various embodiments of this application can be combined and / or combined in various ways without departing from the spirit and teachings of this application. All such combinations and / or combinations fall within the scope of this application.

Claims

1. A resource allocation approval method, characterized in that, The method includes: In response to a resource allocation approval request, the system retrieves the employee's historical resource allocation values ​​and employee attribute data; the resources represent the computing resources available on the terminal used by the employee. The resource allocation value prediction model, which is pre-trained, predicts resource allocation values ​​based on the employees' historical resource allocation values ​​and employee attribute data, and outputs predicted resource allocation values ​​and group resource allocation status with detailed resource allocation values. A resource allocation approval report is generated by comparing the predicted resource allocation value, the group resource allocation status, and the pre-acquired actual calculated resource allocation value.

2. The resource allocation approval method according to claim 1, characterized in that, In response to a resource allocation approval request, retrieve the employee's historical resource allocation values ​​and employee attribute data, including: The received resource allocation approval requests are parsed to obtain the associated employees of the resource allocation approval requests; The associated employee's department, job type, and job level are retrieved as the first employee attribute, and the first employee attribute is sent to the requesting party that issued the resource allocation approval request so that the requesting party can verify the identity of the associated employee. The associated employee whose identity has been verified is designated as the target employee. The system receives the identity verification receipt and identity key issued in the request, and retrieves the target employee's historical resource allocation values ​​and employee attribute data based on the identity verification receipt and identity key. The employee's historical resource allocation value includes the computing power data value of the historical terminals used by the employee; the computing power data value is obtained by weighted summation and standardization of chip data, database data, hard disk storage space data, and cloud storage data.

3. The resource allocation approval method according to claim 2, characterized in that, Retrieving the employee attribute data includes: Obtain the department transfer data and job adjustment data of the target employee as dynamic attributes; Obtain the target employee's years of service, educational background, and work record data as a second employee attribute, and use the first employee attribute and the second employee attribute as static attributes; The data in the dynamic attributes and the static attributes are aggregated to form employee attribute data.

4. The resource allocation approval method according to claim 3, characterized in that, Before pre-training the resource allocation value prediction model, the following steps are included: Obtain employee attribute data for all employees within the target company as the base dataset; In response to the presence of non-continuous data in the basic dataset within a preset time interval, the median within the time interval is used to fill the gaps in the non-continuous data to form a continuous dataset. The continuous dataset is deduplicated and standardized to form a standard dataset; The standard dataset is classified and refined based on a preset overall classification system to obtain sample data including the static attributes, dynamic attributes, and historical resource allocation values ​​of all employees.

5. The resource allocation approval method according to claim 4, characterized in that, Pre-training the resource allocation value prediction model includes: Based on the preset constraint prompts, the preset recurrent neural network architecture is repeatedly trained using the sample data, so that the recurrent neural network architecture outputs training prediction individual resource allocation value entries and training prediction group resource allocation data corresponding to each employee in the sample data. Calculate the first weighted mean square error between the training predicted individual resource allocation value entries and the pre-acquired current employee resource allocation value data, and calculate the second weighted mean square error between the training predicted group resource allocation data and the pre-acquired current employee group data; The first weighted mean square error and the second weighted mean square error are summed in a weighted manner to obtain the loss function value of the recurrent neural network architecture; wherein, Training stops when the loss function value falls below a preset loss threshold, and the optimal recurrent neural network architecture under the constraint prompts is used as the resource allocation value prediction model; wherein, The current data for the employee group is obtained by categorizing and averaging the current resource allocation values ​​of the employees.

6. The resource allocation approval method according to claim 5, characterized in that, The recurrent neural network architecture outputs training prediction individual resource allocation value entries and training prediction group resource allocation data corresponding to each employee in the sample data, including: Feature extraction is performed on the employee's historical resource allocation value data, the static attribute, and the dynamic attribute to obtain time series features, static features, and dynamic features; The time series features, static features, and dynamic features are fused to obtain fused features; A preset individual task head algorithm is used to perform feature mapping on the fused features to obtain training prediction individual resource allocation value entries. A preset group task head algorithm is used to perform category feature mapping on the fused features according to the categories of the first employee attribute and the second employee attribute to obtain training prediction group resource allocation data corresponding to each category of the first employee attribute and the second employee attribute.

7. The resource allocation approval method according to claim 6, characterized in that, The resource allocation value prediction model is a large-scale model. Based on the employee's historical resource allocation values ​​and employee attribute data, the model predicts resource allocation values ​​and outputs predicted resource allocation values ​​and group resource allocation status with detailed resource allocation information, including: Based on the constraint prompt information, feature extraction is performed on the target employee's historical resource allocation value and employee attribute data to obtain historical resource allocation value features and attribute features. The historical resource allocation value features and the attribute features are fused to obtain employee features; The individual task head algorithm is used to perform feature mapping on the employee characteristics based on a preset resource allocation value detail category to obtain a predicted resource allocation value with resource allocation value details; and the group task head algorithm is used to perform category feature mapping on the employee characteristics according to the categories of the first employee attribute and the second employee attribute to obtain the group resource allocation status corresponding to each category of the first employee attribute and the second employee attribute.

8. The resource allocation approval method according to claim 7, characterized in that, The constraint prompt information represents the constraint conditions for the historical resource allocation value features and the attribute features, so that the large model can strengthen the extraction of the predicted resource allocation value and group resource allocation status corresponding to the fusion features based on the historical resource allocation value features and the attribute features.

9. The resource allocation approval method according to claim 3, characterized in that, Based on the predicted resource allocation value, the group's resource allocation status, and the pre-acquired actual calculated resource allocation value, a data comparison is performed to generate a resource allocation approval report, including: The group resource allocation status is processed to obtain a group resource allocation statistics table corresponding to the target employee; wherein, the group resource allocation statistics table uses the categories of the first employee attribute and the second employee attribute as table headers, and is obtained by data mapping between the group resource allocation status and the table headers; The actual calculated resource allocation value is compared with the individual data in the group resource allocation statistics table to obtain the external comparison ratio, and the actual calculated resource allocation value is compared with the predicted resource allocation value to obtain the internal ratio. A four-quadrant analysis is performed based on the external ratio and the internal ratio to obtain resource allocation value results. A horizontal approval analysis is then performed based on the resource allocation value results to generate a resource allocation approval report for the target employee.

10. A resource allocation approval device, characterized in that, The device includes: The approval response module is used to respond to resource allocation approval requests by retrieving employees' historical resource allocation values ​​and employee attribute data; the resources represent the computing resources available on the terminal used by the employee. The resource allocation value prediction module is used to predict resource allocation values ​​based on the employee's historical resource allocation values ​​and employee attribute data using a pre-trained resource allocation value prediction model, so as to output the predicted resource allocation value and the group's resource allocation status with resource allocation value details. The resource allocation value analysis module is used to compare data based on the predicted resource allocation value, the group resource allocation status, and the pre-acquired actual calculated resource allocation value to generate a resource allocation approval report.

11. An electronic device, comprising: One or more processors; Memory, used to store one or more computer programs. The characteristic feature is that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 9.

12. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 9.

13. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 9.

Images