Performance prediction method and device, electronic equipment and storage medium

Abstract

The application provides a performance prediction method and device, electronic equipment and a storage medium. The performance prediction method comprises the following steps: obtaining to-be-analyzed data, wherein the to-be-analyzed data comprises work records of an employee; determining a plurality of work information of the employee and a category corresponding to each work information according to the work records; encoding the corresponding work information according to the category to obtain an encoding vector of the to-be-analyzed data; and processing the encoding vector based on a pre-trained performance prediction model to obtain a predicted performance of the employee. The application relates to the technical field of performance evaluation, and can improve the accuracy of the predicted performance.

Description

Technical Field

[0001] This application relates to the field of performance evaluation, and more particularly to a performance prediction method, apparatus, electronic device, and storage medium. Background Technology

[0002] Currently, most companies and organizations tend to use performance appraisals to statistically analyze employee work results. To improve the efficiency of employee performance management, performance review standards and evaluation criteria are usually set in advance. By comparing employee performance with these criteria, it is determined whether the employee's performance meets the standards. However, this method relies on the identification of performance evaluation criteria and often depends on subjective evaluation and qualitative analysis, which may lead to inaccurate and unfair evaluation results, potentially resulting in significant errors in performance appraisals. Summary of the Invention

[0003] In view of the above, it is necessary to propose a performance prediction method, device, electronic device and storage medium to solve the technical problem of low accuracy in performance prediction.

[0004] This application provides a performance prediction method applied to an electronic device. The method includes: acquiring data to be analyzed, the data to be analyzed including employee work records; determining multiple work information of the employee and a category corresponding to each work information based on the work records; encoding the corresponding work information according to the category to obtain an encoding vector of the data to be analyzed; and processing the encoding vector based on a pre-trained performance prediction model to obtain the predicted performance of the employee.

[0005] In some embodiments, the method further includes training the prediction model, wherein training the prediction model includes: acquiring historical work records and corresponding historical performance; constructing graph data based on the historical work records; obtaining a first predicted performance based on the graph data using a pre-constructed initial prediction model; determining the loss value of the initial prediction model based on the first predicted performance and the historical performance; updating the initial prediction model using a backpropagation algorithm until the loss value meets a preset condition, thereby obtaining a trained performance prediction model.

[0006] In some embodiments, constructing graph data based on the historical work records includes: preprocessing the historical work records to obtain multiple corresponding historical work information; classifying the multiple historical work information to obtain a category for each historical work information; sorting the historical work information according to the category to obtain an information sequence; and constructing graph data based on the information sequence.

[0007] In some embodiments, sorting the historical work information according to the category to obtain an information sequence includes: determining the arrangement order of the historical work information according to a preset category sequence and the category; determining the association relationship between every two adjacent historical work information according to the arrangement order; and determining the information sequence according to the arrangement order and the association relationship between every two adjacent historical work information.

[0008] In some embodiments, determining the loss value of the initial prediction model based on the first predicted performance and the historical performance includes: determining the historical performance corresponding to the information sequence based on the plurality of historical work information; determining the first predicted performance corresponding to the information sequence based on the historical work records; and determining the loss value of the initial prediction model based on the difference between the first predicted performance and the historical performance.

[0009] In some embodiments, determining multiple job information of the employee and the category corresponding to each job information based on the job record includes: determining prompt data based on the job record and a pre-stored prompt template; and determining multiple job information of the employee and the category corresponding to each job information based on the prompt data using a generative pre-trained model.

[0010] In some embodiments, encoding the corresponding work information according to the category to obtain the encoded vector of the data to be analyzed includes: determining the arrangement order of the work information according to a preset category sequence and the category; determining the association relationship between every two adjacent work information according to the arrangement order; encoding the work information and the association relationship to obtain a first vector corresponding to the work information and a second vector corresponding to the association relationship; and combining the first vector and the second vector according to the arrangement order to obtain the encoded vector.

[0011] This application embodiment also provides a performance prediction device, the device comprising: an acquisition module for acquiring data to be analyzed, the data to be analyzed including employee work records; a determination module for determining multiple work information of the employee and a category corresponding to each work information based on the work records; the determination module is further configured to encode the corresponding work information according to the category to obtain an encoding vector of the data to be analyzed; and a prediction module for processing the encoding vector based on a pre-trained performance prediction model to obtain the predicted performance of the employee.

[0012] This application also provides an electronic device, comprising: a memory storing at least one instruction; and a processor executing the instructions stored in the memory to implement the performance prediction method.

[0013] This application also provides a computer-readable storage medium storing at least one instruction, which is executed by a processor in an electronic device to implement the performance prediction method described above.

[0014] As can be seen from the above technical solutions, the embodiments of this application, by parsing employee work records, determine the different types of work information within the records, enabling fine-grained classification of work records and improving the accuracy of employee performance prediction. After obtaining different types of work information, the relationships between these different types of work information are also determined, and the corresponding encoding vectors for employee work records are determined based on the work information and relationships. This allows for a quantitative representation of the semantic relationships between employee work records and different types of work information within them, thereby improving the accuracy of performance prediction. Attached Figure Description

[0015] Figure 1 This is an application scenario diagram of a performance prediction method provided in an embodiment of this application.

[0016] Figure 2 This is a flowchart of a performance prediction method provided in an embodiment of this application.

[0017] Figure 3 This is a flowchart of a method for training a performance prediction model according to an embodiment of this application.

[0018] Figure 4 This is a flowchart of a method for constructing graph data according to an embodiment of this application.

[0019] Figure 5 This is a flowchart of a method for determining an information sequence provided in an embodiment of this application.

[0020] Figure 6 This is a functional block diagram of a performance prediction device provided in an embodiment of this application.

[0021] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0022] To better understand the purpose, features, and advantages of this application, a detailed description of the application is provided below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of this application can be combined with each other. Numerous specific details are set forth in the following description to provide a thorough understanding of this application; the described embodiments are only a part of the embodiments of this application, and not all of them.

[0023] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0025] This application provides a performance prediction method that can be applied to one or more electronic devices. An electronic device is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions. Its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0026] Electronic devices can be any electronic product that allows human-computer interaction with a customer, such as personal computers, tablets, smartphones, personal digital assistants (PDAs), game consoles, interactive network television (IPTV), smart wearable devices, etc.

[0027] Electronic devices may also include network devices and / or client devices. The network devices include, but are not limited to, a single network server, a server group consisting of multiple network servers, or a cloud based on cloud computing consisting of a large number of hosts or network servers.

[0028] The networks in which electronic devices are located include, but are not limited to, the Internet, wide area networks, metropolitan area networks, local area networks, and virtual private networks (VPNs).

[0029] like Figure 1As shown, the performance prediction method provided in this application can be applied to an electronic device 100, which is communicatively connected to a server 200. The electronic device 100 is used to obtain data to be analyzed from the server 200, including employee work records. Based on the work records, it determines multiple work information items for the employee and the category corresponding to each work information item. Then, it encodes the corresponding work information according to the category to obtain an encoding vector for the data to be analyzed. Finally, it processes the encoding vector based on a pre-trained performance prediction model to obtain the employee's predicted performance.

[0030] like Figure 2 The diagram shown is a flowchart of a performance prediction method according to an embodiment of this application. The order of the steps in the flowchart can be changed, and some steps can be omitted, depending on different needs. The performance prediction method provided in this embodiment includes the following steps.

[0031] S20, Obtain the data to be analyzed, which includes employee work records.

[0032] In one embodiment of this application, the data to be analyzed can be a text file input by the user into the electronic device, or a data table file. This application does not limit the specific format of the data to be analyzed. The employee's work record includes at least the employee's department, employee ID, work tasks assigned to the employee, and the performance record obtained by the employee after completing a specific work task.

[0033] S21, determine multiple job information of the employee and the category corresponding to each job information based on the job record.

[0034] In one embodiment of this application, in order to represent the relationships between different work information of different employees based on a graph structure, multiple work information of an employee and the category corresponding to each work information can be determined according to the work records. Specifically, the data to be analyzed can be parsed to obtain the text data used to record work records in the data to be analyzed, and multiple work information can be obtained by parsing the text data. The multiple work information used to represent the employee's work record includes at least the employee's department, employee information, work items that the employee is responsible for, and the performance obtained by the employee after completing a certain work item.

[0035] In one embodiment of this application, to improve the data quality of work information, preprocessing operations are performed on the parsed text data after acquisition to remove redundant and missing data. These preprocessing operations include data deduplication and missing information filtering, which are not limited in this application. This improves the quality of data related to employee work records before predicting employee performance, thereby enhancing the accuracy of performance prediction. For example, text cleaning can remove irrelevant characters, noise, and redundant information from the text data; word segmentation and part-of-speech tagging can also be performed on the text data in the work records to facilitate subsequent extraction of work information and improve the efficiency of information extraction. Furthermore, when identifying multiple pieces of work information, generative pre-trained models can be used to identify named entities in the work records, such as names of people, places, and organizations. This provides data support for subsequently determining the relationships between different pieces of work information. Based on named entity recognition, relationships between entities can also be determined, such as determining relationships like "department-item" or "item-profit" from the text data. Event extraction can also be performed on work records to identify events and related elements in the text, such as event type, time, location, and participants. Furthermore, generative pre-trained models can be used to determine the semantics of text data, extracting key information and validating and standardizing the extracted information to ensure accuracy and consistency. For example, regular expressions can be used for validation, and data formats can be standardized based on entity properties.

[0036] In one embodiment of this application, determining multiple work information items of the employee and the category corresponding to each work information item based on the work records includes: determining prompt data based on the work records and a pre-stored prompt template; and determining multiple work information items of the employee and the category corresponding to each work information item based on the prompt data using a generative pre-trained model. Specifically, historical work records are filled into a pre-edited prompt template to obtain corresponding prompt data. The generative pre-trained model processes the prompt data to generate text data containing multiple work information items (e.g., project name, executor, execution time, work summary) and the type of each work information item (e.g., management tasks, requirements analysis, development coding, testing tasks, maintenance tasks, etc.). The classification result output by the generative pre-trained model based on the prompt data also includes the probability that historical work information belongs to a certain category, and the category with the highest probability can be determined as the category corresponding to the historical work information. The generative pre-trained model can be a model with text classification capabilities, such as the BERT model, XLNet model, or GPT model, and this application does not limit it to this type. To improve the performance of generative pre-trained models and thus the accuracy of data classification, the models can be trained to adjust their parameters using labeled datasets.

[0037] S22, the corresponding work information is encoded according to the category to obtain the encoded vector of the data to be analyzed.

[0038] In one embodiment of this application, in order to quantify and characterize work information, thereby improving the accuracy of performance prediction in subsequent performance forecasting, the corresponding work information can be encoded according to categories. Specifically, encoding work information can improve the storage and management efficiency of work information, converting text information into a form that electronic devices can recognize and process, thus effectively storing it in the memory of electronic devices, and also improving the efficiency of subsequent information retrieval and the accuracy of information processing. In addition, encoding work information can also achieve information standardization and normalization. A unified encoding standard helps to determine the relationship between different types of work information, and can also protect the security and confidentiality of work information. Some encoding methods can also encrypt text information, increasing the confidentiality of information. Through encryption encoding, it can be ensured that work information is not accessed or tampered with by unauthorized personnel during transmission and processing.

[0039] In one embodiment of this application, the step of encoding the corresponding work information according to the category to obtain the encoding vector of the data to be analyzed includes: determining the arrangement order of the work information according to a preset category sequence and the category; determining the association relationship between every two adjacent work information according to the arrangement order; encoding the work information and the association relationship to obtain a first vector corresponding to the work information and a second vector corresponding to the association relationship; and combining the first vector and the second vector according to the arrangement order to obtain the encoding vector.

[0040] Specifically, the predefined category sequence represents the order in which work information is concatenated during the encoding of multiple pieces of work information. Within the category sequence, the relationship between two adjacent pieces of work information can be either a "subordinate relationship" or a "correspondence relationship." For example, the category sequence could be "Organization, Employee ID, Performance, Item." The association between each pair of adjacent pieces of work information represents the connection between work information of different categories. For example, the association between work information of category "Organization" and work information of category "Employee ID" is a subordinate relationship, as is the association between work information of category "Employee ID" and work information of category "Performance." The methods for encoding work information include, but are not limited to, one-hot encoding, bag-of-words model, inverse document frequency model, and word embedding model.

[0041] For example, after encoding the work information, the code corresponding to the work information of category "organization" can be

[1000] , the code corresponding to the work information of category "employee number" can be

[0100] , the code corresponding to the work information of category "performance" can be

[0010] , and the code corresponding to the work information of category "item" can be

[0001] . The second code corresponding to the association between the work information of category "organization" and the work information of category "employee number" can be

[1000] , the second code corresponding to the association between the work information of category "employee number" and the work information of category "performance" can be

[2000] , and the second code corresponding to the association between the work information of category "performance" and the work information of category "item" can be

[3000] . If the category sequence is "organization, employee number, performance, item", then the first vector and the second vector are combined according to this arrangement order to obtain the encoding vector [1000100001002000001030000001].

[0042] In one embodiment of this application, after determining the encoding vectors corresponding to multiple job information, each dimension of the encoding vector can be normalized to avoid the dimensional differences in the values ​​of different dimensions that occur during the encoding process, thereby improving the accuracy of subsequent prediction of employee performance based on the encoding vectors.

[0043] S23, The encoded vector is processed based on a pre-trained performance prediction model to obtain the predicted performance of the employee.

[0044] In one embodiment of this application, the encoded vector can represent an employee's work information in a quantitative form. The encoded vector can be input into a pre-trained performance prediction model to obtain the employee's predicted performance. The performance prediction model can be a model with graph structure data processing capabilities, used to predict employee performance based on the encoded vector. For example, the performance prediction model can be a Metapath Aggregated Graph Neural Network (MAGNN).

[0045] In one embodiment of this application, the output of the performance prediction model can be data in vector form. For example, when the performance prediction model outputs data in vector form, the prediction result can be [(0-60, 0.1), (60-70, 0.2), (70-80, 0.25), (80-90, 0.35), (90-100, 0.1)]. This prediction result represents the probability that an employee's predicted performance is between 0 and 60 (0.1), between 60 and 70 (0.2), between 70 and 80 (0.25), between 80 and 90 (0.35), and between 90 and 100 (0.1). Based on this prediction result, the employee's predicted performance can be determined to be in the range of 80 to 90 points.

[0046] In one embodiment of this application, to improve the accuracy of employee performance prediction based on the encoded vector, a performance prediction model can be trained using pre-stored historical work records of employees. This improves the performance of the performance prediction model and thus enhances the accuracy of performance prediction. For specific methods of training the performance prediction model, please refer to [link to relevant documentation]. Figure 3 The corresponding detailed explanation.

[0047] As can be seen from the above technical solutions, the embodiments of this application, by parsing employee work records, determine the different types of work information within the records, enabling fine-grained classification of work records and improving the accuracy of employee performance prediction. After obtaining different types of work information, the relationships between these different types of work information are also determined, and the corresponding encoding vectors for employee work records are determined based on the work information and relationships. This allows for a quantitative representation of the semantic relationships between employee work records and different types of work information within them, thereby improving the accuracy of performance prediction.

[0048] like Figure 3 The diagram shown is a flowchart of a method for training a performance prediction model according to an embodiment of this application. The order of steps in this flowchart can be changed, and some steps can be omitted, depending on different needs. The method for training a performance prediction model according to an embodiment of this application includes the following steps.

[0049] S30: Obtain historical work records and corresponding historical performance data.

[0050] In one embodiment of this application, a pre-stored historical work log can be parsed to obtain historical work records and corresponding historical performance data. The historical work log can be a text file pre-stored in an electronic device or a data table file; this application does not limit the specific format of the historical work log. The employee's historical work records at least include the employee's department, employee ID, work tasks assigned to the employee, and historical performance data obtained after completing a specific work task.

[0051] In one embodiment of this application, historical performance data can be in vector form. For example, when the historical performance score is 85, the historical performance score can be [(0-60, 0), (60-70, 0), (70-80, 0), (80-90, 1), (90-100, 0)], where the historical performance score represents the probability that the employee's historical performance score is between 0 and 60, between 60 and 70, between 70 and 80, between 80 and 90, and between 90 and 100.

[0052] S31, Construct graph data based on the historical work records.

[0053] In one embodiment of this application, multiple pieces of historical work information in historical work records can be encoded to construct a node corresponding to each piece of historical work information based on the encoding vector of the historical work information. Multiple information sequences are then constructed based on the nodes, and graph data is built based on the multiple information sequences. The information sequences are used to connect two nodes, representing the relationship between the two nodes and revealing the potential connections and semantic information between different nodes in the graph data.

[0054] Specifically, historical work records can be cleaned and preprocessed. Preprocessing includes removing duplicate data, handling missing values, and standardizing data formats. When constructing nodes and relationships based on historical work information in the historical work records, different types of nodes in the graph can be defined first according to the information sequence, and each node can be assigned a unique identifier for differentiation and referencing within the graph data. Then, based on the information sequence and node types, different types of relationships in the graph can be defined, each relationship can be assigned a unique identifier, and the types of nodes connected by each relationship can be determined. According to the order of nodes and relationships in the information sequence, subgraph data containing relevant nodes and relationships can be extracted from the historical work information. Multiple subgraph data can be merged to obtain the graph data corresponding to the historical work records.

[0055] In one embodiment of this application, the method for constructing graph data based on historical work records can be found in [link to relevant documentation]. Figure 4 The corresponding detailed explanation.

[0056] S32, Based on the graph data, the first prediction performance is obtained using a pre-constructed initial prediction model.

[0057] In one embodiment of this application, graph data can be input into a pre-built initial prediction model to obtain the corresponding first predicted performance. The graph data is used to represent an employee's historical work records and the relationships between multiple pieces of historical work information within those records.

[0058] The first predicted performance can be data in vector form. For example, the first predicted performance can be [(0-60, 0.1), (60-70, 0.2), (70-80, 0.25), (80-90, 0.35), (90-100, 0.1)]. This first predicted performance is used to characterize the probability that an employee's predicted performance is between 0 and 60 (0.1), between 60 and 70 (0.2), between 70 and 80 (0.25), between 80 and 90 (0.35), and between 90 and 100 (0.1).

[0059] S33, Based on the first predicted performance and the historical performance, determine the loss value of the initial prediction model, update the initial prediction model using the backpropagation algorithm until the loss value meets the preset conditions, and obtain the trained performance prediction model.

[0060] In one embodiment of this application, determining the loss value of the initial prediction model based on the first predicted performance and the historical performance includes: determining the historical performance corresponding to the information sequence based on the plurality of historical work information; determining the first predicted performance corresponding to the information sequence based on the historical work records; and determining the loss value of the initial prediction model based on the difference between the first predicted performance and the historical performance. Specifically, the difference between the first predicted performance and the historical performance can be determined based on the cross-entropy function, and this difference can be determined as the loss value of the initial prediction model.

[0061] like Figure 4 The diagram shown is a flowchart of a method for constructing graph data according to an embodiment of this application. The order of steps in this flowchart can be changed, and some steps can be omitted, depending on different requirements. The method for constructing graph data provided in this embodiment includes the following steps.

[0062] S40, preprocess the historical work records to obtain multiple corresponding historical work information.

[0063] In one embodiment of this application, in order to represent the relationships between different work information of different employees based on a graph structure, multiple work information of an employee and the category corresponding to each work information can be determined according to work records. Specifically, historical work records can be parsed to obtain text data in the data to be analyzed that records historical work information, and multiple historical work information can be obtained by parsing the text data. The multiple historical work information used to represent an employee's work record includes at least the employee's department, employee information, work items the employee is responsible for, and the performance obtained by the employee after completing a certain work item.

[0064] In one embodiment of this application, to improve the data quality of work information, preprocessing operations are performed on the parsed text data after acquisition to remove redundant and missing data. These preprocessing operations include data deduplication and missing information filtering, which are not limited in this application. This improves the quality of data related to employee work records before predicting employee performance, thereby enhancing the accuracy of performance prediction. For example, text cleaning can remove irrelevant characters, noise, and redundant information from the text data; word segmentation and part-of-speech tagging can also be performed on the text data in the work records to facilitate subsequent extraction of work information and improve the efficiency of information extraction. Furthermore, when identifying multiple pieces of work information, generative pre-trained models can be used to identify named entities in the work records, such as names of people, places, and organizations. This provides data support for subsequently determining the relationships between different pieces of work information. Based on named entity recognition, relationships between entities can also be determined, such as determining relationships like "department-item" or "item-profit" from the text data. Event extraction can also be performed on work records to identify events and related elements in the text, such as event type, time, location, and participants. Furthermore, generative pre-trained models can be used to determine the semantics of text data, extracting key information and validating and standardizing the extracted information to ensure accuracy and consistency. For example, regular expressions can be used for validation, and data formats can be standardized based on entity properties.

[0065] S41, classify the multiple historical work information to obtain the category of each historical work information.

[0066] In one embodiment of this application, historical work records can be filled into a pre-edited prompt template to obtain corresponding prompt data. This prompt data is processed using a generative pre-trained model to generate text data containing multiple pieces of work information (e.g., project name, executor, execution time, work summary), as well as the type of each type of work information (e.g., management tasks, requirements analysis, development coding, testing tasks, maintenance tasks, etc.). The classification results output by the generative pre-trained model based on the prompt data also include the probability that the historical work information belongs to a certain category; the category with the highest probability can be determined as the category corresponding to the historical work information.

[0067] S42, Sort the historical work information according to the category to obtain an information sequence.

[0068] In one embodiment of this application, the information sequence is used to connect paths between nodes corresponding to at least two different types of working information, and can characterize the association relationships between different nodes in graph data. Before determining the information sequence, the types of different nodes can be determined, that is, the node types of the start and end points of the information sequence. Specifically, the selected node types and defined relationship types are connected in logical order to form a complete information sequence. For example, in a knowledge graph in the e-commerce field, the information sequence could be "user-purchased product-review-user". This information sequence represents that in the e-commerce field's indicator graph, a query can start from any user's information, find the product through the user's purchase relationship, and then find another user through the review relationship. In the process of evaluating employee performance, if the category sequence is "organization, employee ID, performance, task", then the order of historical work information in the information sequence can be set to "organization-employee ID-performance-task". This information sequence is used to represent that in the graph data representing user performance, a query can start from the organization information, find the employee's ID through the subordinate relationship, then find the corresponding performance through the employee's ID, and finally find the corresponding work task through the performance. This can represent the performance data achieved by an employee belonging to a certain organization after completing a certain work task.

[0069] In one embodiment of this application, for a specific method of sorting historical work information according to its category to determine the information sequence, please refer to [link to relevant documentation]. Figure 5 The corresponding detailed explanation.

[0070] S43, construct graph data based on the information sequence.

[0071] In one embodiment of this application, when constructing nodes and relationships based on historical work information in historical work records, different types of nodes existing in the graph can be defined first according to the information sequence, and each node can be assigned a unique identifier for differentiation and reference in the graph data. Then, based on the information sequence and node types, different types of relationships existing in the graph are defined, each relationship is assigned a unique identifier, and the type of nodes connected by each relationship is determined. According to the order of nodes and relationships in the information sequence, subgraph data containing relevant nodes and relationships is extracted from the historical work information, and multiple subgraph data are merged to obtain the graph data corresponding to the historical work records.

[0072] In one embodiment of this application, if conflicts arise in nodes or relationships (e.g., duplicate nodes, inconsistent node relationships, etc.) during the merging of multiple subgraph data, any one node and any one relationship can be retained from among the conflicting nodes. Furthermore, the constructed graph data can be verified based on arbitrary graph data processing instructions to ensure that the nodes and relationships in the graph data conform to the definition and logical order of the information sequence.

[0073] like Figure 5 The diagram shown is a flowchart of a method for determining an information sequence according to an embodiment of this application. The order of the steps in this flowchart can be changed, and some steps can be omitted, depending on different requirements. The method for determining an information sequence provided in this embodiment includes the following steps.

[0074] S50, determine the arrangement order of the historical work information according to the preset category sequence and the category.

[0075] In one embodiment of this application, a preset category sequence is used to characterize the order of multiple historical information items during the training of the performance prediction model. For example, the category sequence could be "organization, employee number, performance, task". The association between each pair of adjacent work information items is used to characterize the relationship between different categories of work information.

[0076] In one embodiment of this application, a preset category sequence is used to characterize the splicing order of historical work information during the encoding of multiple historical work information. For example, when the preset category sequence is "organization, employee number, performance, matter", the arrangement order of historical work information can be determined as "historical work information of category organization - historical work information of category employee number - historical work information of category performance - historical work information of category matter".

[0077] S51, Based on the arrangement order, determine the association between every two adjacent historical work information.

[0078] In one embodiment of this application, the association between every two adjacent historical work information is used to characterize the connection between historical work information of different categories. For example, the association between historical work information of category "organization" and historical work information of category "employee number" is a subordinate relationship, and the association between historical work information of category "employee number" and historical work information of category "performance" is also a subordinate relationship.

[0079] S52, determine the information sequence based on the arrangement order and the correlation between every two adjacent historical work information.

[0080] In one embodiment of this application, the information sequence is used to characterize the arrangement order of different categories of historical work information and the relationships between every two adjacent historical work information. For example, the information sequence may be "historical work information categorized as organization - subordinate relationship - historical work information categorized as employee number - corresponding relationship - historical work information categorized as performance - corresponding relationship - historical work information categorized as matter".

[0081] Please see Figure 6 , Figure 6 This is a functional block diagram of a performance prediction device according to an embodiment of this application. A performance prediction device 61 includes an acquisition module 610, a determination module 611, and a prediction module 612. The module / unit referred to in this application refers to a series of computer-readable instruction segments that can be executed by the processor 13 and perform a fixed function, and are stored in the memory 12. In this embodiment, the functions of each module / unit will be described in detail in subsequent embodiments.

[0082] The acquisition module 610 is used to acquire data to be analyzed, including employee work records.

[0083] The determining module 611 is used to determine multiple work information of the employee and the category corresponding to each work information based on the work record.

[0084] The determining module 611 is further configured to encode the corresponding work information according to the category to obtain the encoding vector of the data to be analyzed.

[0085] The prediction module 612 is used to process the encoded vector based on a pre-trained performance prediction model to obtain the predicted performance of the employee.

[0086] In some embodiments, the prediction module 612 is further configured to acquire historical work records and corresponding historical performance; construct graph data based on the historical work records; obtain a first predicted performance based on the graph data using a pre-constructed initial prediction model; determine the loss value of the initial prediction model based on the first predicted performance and the historical performance; update the initial prediction model using a backpropagation algorithm until the loss value meets a preset condition, thereby obtaining a trained performance prediction model.

[0087] In some embodiments, the prediction module 612 is further configured to preprocess the historical work records to obtain a plurality of corresponding historical work information; classify the plurality of historical work information to obtain a category for each historical work information; sort the historical work information according to the category to obtain an information sequence; and construct graph data according to the information sequence.

[0088] In some embodiments, the prediction module 612 is further configured to determine the arrangement order of the historical work information according to a preset category sequence and the category; determine the association relationship between every two adjacent historical work information according to the arrangement order; and determine the information sequence according to the arrangement order and the association relationship between every two adjacent historical work information.

[0089] In some embodiments, the prediction module 612 is further configured to: determine the historical performance corresponding to the information sequence based on the plurality of historical work information; determine the first predicted performance corresponding to the information sequence based on the historical work records; and determine the loss value of the initial prediction model based on the difference between the first predicted performance and the historical performance.

[0090] In some embodiments, the determining module 611 is further configured to determine prompt data based on the work record and the pre-stored prompt template; and determine multiple work information of the employee and the category corresponding to each work information based on the prompt data and a generative pre-trained model.

[0091] In some embodiments, the determining module 611 is further configured to determine the arrangement order of the work information according to a preset category sequence and the category; determine the association relationship between every two adjacent work information according to the arrangement order; encode the work information and the association relationship to obtain a first vector corresponding to the work information and a second vector corresponding to the association relationship; and combine the first vector and the second vector according to the arrangement order to obtain an encoded vector.

[0092] Please see Figure 7This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device 100 includes a memory 12 and a processor 13. The memory 12 is used to store computer-readable instructions, and the processor 13 executes the computer-readable instructions stored in the memory to implement a performance prediction method as described in any of the above embodiments.

[0093] In one embodiment of this application, the electronic device 100 further includes a bus and a computer program stored in the memory 12 and executable on the processor 13, such as a performance prediction program.

[0094] Figure 7 Only an electronic device 100 with memory 12 and processor 13 is shown; those skilled in the art will understand that... Figure 7 The structure shown does not constitute a limitation on the electronic device 100, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0095] Combination Figure 2 The memory 12 in the electronic device 100 stores a plurality of computer-readable instructions to implement a performance prediction method. The processor 13 can execute the plurality of instructions to: acquire data to be analyzed, the data to be analyzed including employee work records; determine multiple work information of the employee and the category corresponding to each work information based on the work records; encode the corresponding work information according to the category to obtain an encoding vector of the data to be analyzed; and process the encoding vector based on a pre-trained performance prediction model to obtain the predicted performance of the employee.

[0096] Specifically, the processor 13's implementation method for the above instructions can be found in [reference needed]. Figure 2 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0097] Those skilled in the art will understand that the schematic diagram is merely an example of the electronic device 100 and does not constitute a limitation on the electronic device 100. The electronic device 100 may be a bus-type structure or a star-type structure. The electronic device 100 may also include more or fewer other hardware or software than shown in the diagram, or different component arrangements. For example, the electronic device 100 may also include input / output devices, network access devices, etc.

[0098] It should be noted that electronic device 100 is only an example. Other existing or future electronic products that are suitable for this application should also be included within the scope of protection of this application and are incorporated herein by reference.

[0099] The memory 12 includes at least one type of readable storage medium, which can be non-volatile or volatile. The readable storage medium includes flash memory, portable hard drives, multimedia cards, card-type memory (e.g., SD or DX memory), magnetic storage, magnetic disks, optical disks, etc. In some embodiments, the memory 12 can be an internal storage unit of the electronic device 100, such as the portable hard drive of the electronic device 100. In other embodiments, the memory 12 can also be an external storage device of the electronic device 100, such as a plug-in portable hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the electronic device 100. The memory 12 can be used not only to store application software and various types of data installed on the electronic device 100, such as the code of a performance prediction program, but also to temporarily store data that has been output or will be output.

[0100] In some embodiments, the processor 13 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits packaged with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 13 is the control unit of the electronic device 100, connecting to various components of the electronic device 100 via various interfaces and lines. It executes programs or modules stored in the memory 12 (e.g., executing a performance prediction program) and calls data stored in the memory 12 to perform various functions of the electronic device 100 and process data.

[0101] The processor 13 executes the operating system of the electronic device 100 and various installed applications. The processor 13 executes the applications to implement the steps in each of the above-described performance prediction method embodiments, for example... Figure 2 The steps are shown.

[0102] For example, the computer program may be divided into one or more modules / units, which are stored in the memory 12 and executed by the processor 13 to complete this application. The one or more modules / units may be a series of computer-readable instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device 100. For example, the computer program may be divided into an acquisition module 610, a determination module 611, and a prediction module 612.

[0103] The integrated unit implemented as a software functional module described above can be stored in a computer-readable storage medium. This software functional module, stored in a storage medium, includes several instructions to cause a computer device (which may be a personal computer, computer equipment, or network device, etc.) or processor to execute a portion of the performance prediction method described in the various embodiments of this application.

[0104] If the modules / units integrated in the electronic device 100 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware devices. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above.

[0105] The computer program includes computer program code, which may be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory, and other memory.

[0106] Furthermore, the computer-readable storage medium may primarily include a stored program area and a stored data area, wherein the stored program area may store the operating system, an application program required for at least one function, etc.; and the stored data area may store data created based on the use of blockchain nodes, etc.

[0107] The bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, in... Figure 7 The symbol is represented by only one arrow, but this does not indicate that there is only one bus or one type of bus. The bus is configured to enable communication between the memory 12 and at least one processor 13, etc.

[0108] This application also provides a computer-readable storage medium (not shown) storing computer-readable instructions, which are executed by a processor in an electronic device to implement a performance prediction method described in any of the above embodiments.

[0109] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0110] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0111] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0112] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices described in the specification may also be implemented by a single unit or device through software or hardware. Terms such as "first," "second," etc., are used to indicate names and do not indicate any specific order.

[0113] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of this application without departing from the spirit and scope of the technical solutions of this application.

Claims

1. A performance prediction method applied to electronic devices, characterized in that, The method includes: Acquire the data to be analyzed, which includes employee work records; Based on the work records, determine multiple work information of the employee and the category corresponding to each work information; The corresponding work information is encoded according to the category to obtain the encoding vector of the data to be analyzed; The encoded vector is processed based on a pre-trained performance prediction model to obtain the predicted performance of the employee.

2. The performance prediction method as described in claim 1, characterized in that, The method further includes training the prediction model, wherein training the prediction model includes: Obtain historical work records and corresponding historical performance; Construct graph data based on the historical work records; Based on the graph data, the first prediction performance is obtained using a pre-built initial prediction model; Based on the first predicted performance and the historical performance, the loss value of the initial prediction model is determined, and the initial prediction model is updated using the backpropagation algorithm until the loss value meets the preset conditions, thereby obtaining the trained performance prediction model.

3. The performance prediction method as described in claim 2, characterized in that, The graph data constructed based on the historical work records includes: The historical work records are preprocessed to obtain multiple corresponding historical work information; The multiple historical work information items are classified to obtain the category of each historical work information item; The historical work information is sorted according to the category to obtain an information sequence; The graph data is constructed based on the information sequence.

4. The performance prediction method as described in claim 3, characterized in that, The process of sorting the historical work information according to the category to obtain the information sequence includes: The order of the historical work information is determined according to a preset category sequence and the categories; Based on the arrangement order, determine the association between every two adjacent historical work information entries; The information sequence is determined based on the arrangement order and the correlation between each pair of adjacent historical work information.

5. The performance prediction method as described in claim 3, characterized in that, The step of determining the loss value of the initial prediction model based on the first prediction performance and the historical performance includes: Based on the multiple historical work information, determine the historical performance corresponding to the information sequence; Based on the historical work records, determine the first predicted performance corresponding to the information sequence; The loss value of the initial prediction model is determined based on the difference between the first predicted performance and the historical performance.

6. The performance prediction method as described in claim 1, characterized in that, The step of determining multiple work information items of the employee and the category corresponding to each work information item based on the work records includes: The prompt data is determined based on the work records and the pre-stored prompt templates; Based on the provided prompt data, a generative pre-trained model is used to determine multiple job information items for the employee and the category corresponding to each job information item.

7. The performance prediction method as described in claim 1, characterized in that, The step of encoding the corresponding work information according to the category to obtain the encoding vector of the data to be analyzed includes: The arrangement order of the work information is determined according to a preset category sequence and the categories; Based on the arrangement order, determine the association between every two adjacent pieces of work information; The work information and the association relationship are encoded to obtain a first vector corresponding to the work information and a second vector corresponding to the association relationship; The first vector and the second vector are combined according to the given arrangement order to obtain the encoded vector.

8. A performance prediction device, characterized in that, The device includes: The acquisition module is used to acquire data to be analyzed, including employee work records. The determination module is used to determine multiple work information of the employee and the category corresponding to each work information based on the work records; The determining module is further configured to encode the corresponding work information according to the category to obtain the encoding vector of the data to be analyzed; The prediction module is used to process the encoded vector based on a pre-trained performance prediction model to obtain the predicted performance of the employee.

9. An electronic device, characterized in that, The electronic device includes a processor and a memory, the processor being configured to implement the performance prediction method as described in any one of claims 1 to 7 when executing a computer program stored in the memory.

10. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is executed by a processor, it implements the performance prediction method as described in any one of claims 1 to 7.

Images