WO2020019488A1 - Staff demission warning method and related devices - Google Patents

Abstract

A staff demission warning method and device. The staff demission warning method comprises: determining a demission attribute entry for staff demission warning (S1); obtaining leaving staff sample data corresponding to the demission attribute entry, and constructing a staff demission warning model according to the leaving staff sample data (S2); obtaining target staff attribute data corresponding to the demission attribute entry in staff data of a target staff, and inputting the target staff attribute data into the staff demission warning model (S3); and determining target demission dealing information corresponding to the target staff attribute data on the basis of the staff demission warning model (S4), the target demission dealing information being used for warning the demission of the target staff. The staff demission warning method and device can predict the staff demission dealing information, reduce human resource costs of staff management, improve the retention success rate of staff demission, reduce the damage caused by staff demission, and are high in applicability.

Description

Early warning method and related device for employee turnover

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on July 25, 2018, with application number 201810825916.9, and with the application name of "Employment Pre-Warning Method and Related Device", the entire contents of which are incorporated herein by reference. in.

Technical field

The present application relates to the field of communication technology, and in particular, to an early warning method and related device for employee turnover.

Background technique

In the development of enterprises in various fields and industries, the addition of new employees and the departure of old employees have long been commonplace. However, the role of core employees in the development of an enterprise is extremely important regardless of which company is in which field. For example, the performance of core employees can often add a strong stroke to the growth of the company's performance at all stages of the company's development. Can also bring major breakthroughs to the business development of the enterprise. In addition, core employees can also provide key directional guidance at key development stages of the enterprise, leading the enterprise to a new level of development. However, nowadays, the turnover of core employees of the company is serious. The turnover of core employees will not only cause the loss of human capital investment, increase the cost of human resources reconfiguration, but may also make the company's Development has stalled. Therefore, how to predict the dynamics of employee turnover or make timely response measures has become one of the issues that need to be resolved in the human resource management and even the development of enterprises.

Summary of the Invention

The embodiment of the present application provides an early warning method and related device for employee turnover, which can predict the response information of employee turnover as an early warning of employee turnover, can reduce the human resource cost of employee management, increase the retention success probability of employee turnover, and reduce employee turnover. The resulting loss is highly applicable.

In a first aspect, an embodiment of the present application provides an early warning method for employee turnover, which includes:

Determine the departure attribute entries for employee departure warnings;

Obtaining the sample data of the departing employee corresponding to the above-mentioned attributes of the leaving employee, and constructing the early warning model of the leaving employee based on the sample data of the leaving employee, wherein the sample data of the leaving employee includes at least the sample data of the first leaving employee and the sample data of the second leaving employee The sample data of the first departure employee includes the employee attribute data and the first departure response information of the first type of employees, and the above sample data of the second departure employee includes the employee attribute data and the second departure response information of the second type of employees;

Obtaining target employee attribute data corresponding to the above-mentioned departure attribute entry in the employee data of the target employee, and inputting the above target employee attribute data into the employee departure early warning model;

Based on the above-mentioned employee turnover warning model, target turnover information corresponding to the target employee attribute data is determined, and the target turnover response information is used for early warning of the turnover of the target employee.

In a second aspect, an embodiment of the present application provides an early warning device for employee turnover. The early warning device includes:

A determination unit for determining a departure attribute entry for employee departure warning;

A model building unit is configured to obtain the sample data of the turnover employee corresponding to the above-mentioned turnover attribute entry determined by the determination unit, and construct a warning model of the turnover employee based on the sample data of the turnover employee, wherein the sample data of the turnover employee includes at least the first turnover employee Sample data and sample data of the second employee. The sample data of the first employee includes the employee attribute data of the first type of employee and the first departure response information. The sample data of the second employee includes the information of the second type of employee. Employee attribute data and second departure response information;

A data processing unit, configured to obtain target employee attribute data corresponding to the above-mentioned departure attribute entry determined by the determining unit in the employee data of the target employee, and input the above target employee attribute data into the employee departure warning model;

The determination unit is further configured to determine target turnover response information corresponding to the target employee attribute data input by the data processing unit based on the employee turnover warning model constructed by the model building unit, and the target turnover response information is used to target the target. Early warning of employee departures.

In a third aspect, an embodiment of the present application provides a terminal device. The terminal device includes a processor and a memory, and the processor and the memory are connected to each other. The memory is configured to store a computer program that supports the terminal device to execute the method provided in the first aspect and / or any possible implementation manner of the first aspect. The computer program includes program instructions, and the processor is configured to call the foregoing. A program instruction executes the first aspect and / or the method provided in any possible implementation manner of the first aspect.

According to a fourth aspect, an embodiment of the present application provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, where the computer program includes program instructions, and when the program instructions are executed by a processor, the processor executes the instructions. The first aspect and / or the method provided by any possible implementation manner of the first aspect.

By adopting the embodiment of the present application, the human resource cost of employee management can be reduced, the retention success probability of employee turnover can be improved, the loss caused by employee turnover can be reduced, and the applicability is high.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to explain the technical solution of the embodiment of the present application more clearly, the drawings used in the description of the embodiment will be briefly introduced below.

FIG. 1 is a schematic flowchart of an early warning method for employee turnover provided by an embodiment of the present application; FIG.

2 is a schematic flowchart of a method for constructing an early warning model for employee turnover provided by an embodiment of the present application;

3 is a schematic structural diagram of an early warning device for employee turnover provided by an embodiment of the present application;

FIG. 4 is a schematic structural diagram of a terminal device according to an embodiment of the present application.

detailed description

In the following, the technical solutions in the embodiments of the present application will be clearly and completely described with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, but not all of them.

The early warning method for employee turnover provided in the embodiment of the present application (for convenience of description, the method provided in the embodiment of the present application may be referred to as short) can be applied to the human resource management system of various enterprises in various fields and industries, and the risk of human capital investment Control system, etc. are not limited here. The method provided in the embodiment of the present application may build an employee departure warning model based on the employee data in the employee database of the enterprise, or based on the employee data of the same and / or similar fields and / or industries obtained by big data analysis. Based on the employee departure warning model, the employee data of the target employees of the target enterprise collected in real time can be used to predict the target employee's departure response information, and the target employee's work status and / or departure dynamics can be grasped in time, and timely prevention and / or response can be achieved The turnover of the target employee can reduce the human resource management cost of the enterprise and reduce the risk of the target employee's significant loss to the enterprise due to the turnover of the target employee. The employee departure warning model provided in the embodiment of the present application can be constructed based on a large amount of employee data. Based on employee data in different fields and / or different industries, different models can be constructed to apply to employee departure warnings in various fields or industries, with high flexibility. ,Wide range of applications. For the convenience of description, the following can take any employee in a certain field and / or a target company in a certain industry (for the convenience of description, the target employee can be used as an example) to leave the early warning model as an example, the method provided in the embodiment of this application Describe.

The method and related devices provided in the embodiments of the present application will be described below with reference to FIGS. 1 to 4 respectively. The method provided in the embodiment of the present application may include determination of a departure attribute entry for employee departure warning, construction of an employee departure warning model, acquisition of employee attribute data of the target employee, and departure response of the target employee based on the employee departure warning model. Data processing stages such as information prediction. For implementation manners of the foregoing data processing stages, refer to the implementation manners shown in FIG. 1 to FIG. 2 below.

Referring to FIG. 1, FIG. 1 is a schematic flowchart of an early warning method for employee turnover provided by an embodiment of the present application. The method provided in the embodiment of the present application may include the following steps S1 to S4:

S1. Determine the departure attribute entry for employee departure warning.

In some feasible implementation methods, in order to reduce the amount of data that needs to be processed in the data processing process such as employee departure warning model construction or departure response information prediction, improve the data processing efficiency of employee departure warning, and improve employee departure based on the employee departure warning model. The accuracy of the early warning, in the implementation of the early warning of employee turnover based on the early warning model of employee turnover, the attributes of the departure attributes used for early warning of employee turnover can be determined first. Based on the determined departure attribute entries, more effective information can be selected from the employee data with a larger amount of data as the data basis for employee departure warnings, including the data basis for employee departure warning model training and the prediction of departure response information based on the employee departure warning model. The data foundation can reduce the amount of data processing for employee turnover response information prediction, improve the accuracy of employee turnover response information prediction, and have stronger applicability.

In some feasible implementation manners, the above-mentioned departure attribute entries include, but are not limited to, basic employee information, employee positions, employee tenure, employee income level, employee performance, employee's latest promotion interval, employee reward and punishment items, employee home address, One or more combinations of employee attendance, employee leave frequency, and frequency of login job application URL may be determined according to actual application scenarios, and there is no limitation here.

S2. Obtaining the sample data of the departing employee corresponding to the above-mentioned attributes of the leaving employee, and constructing an early warning model of the leaving employee according to the above sample data of the leaving employee.

In some feasible implementation manners, the construction of an employee departure warning model may include modeling data acquisition of the employee departure warning model, training of the employee departure warning model, and testing of the employee departure warning model during data processing stages. Please refer to FIG. 2 together. FIG. 2 is a schematic flowchart of a method for constructing an early warning model for employee turnover provided by an embodiment of the present application. The construction of the employee departure warning model provided in the embodiment of the present application can be described by the implementation methods provided in steps S21 to S23 as follows.

S21. The modeling data collection of the employee departure warning model.

In some feasible implementation manners, the modeling data of the above-mentioned employee departure warning model may be sample data of the employee departure warning model. For the convenience of description, the modeling data is used as an example for description below. Optionally, the modeling data of the above-mentioned employee departure warning model may be derived from the employee data in the employee data system of the target enterprise. The target company's employee data system can be a data management system that stores employee data, and can also monitor employee data transmission and reception records (including Internet access records, instant communication records, and / or file transmission and reception records, etc.) on the corporate LAN. Restricted) data monitoring system (employees' data monitored and recorded by the target company's data monitoring system), based on which the data monitoring system can collect and record the employee's data of the target company, etc., without any restrictions here. For the convenience of description, the employee data system will be taken as an example for description below. Among them, the modeling data of the above-mentioned employee departure warning model can also be derived from employee data of other companies in the same and / or similar fields and / or industries as the target enterprise obtained by big data analysis, or other more data acquisition paths The obtained employee data can be determined according to the actual application scenario, and there is no limitation here. In the embodiment of the present application, the modeling data of the employee departure warning model can come from multiple paths, which improves the source diversity of the modeling data of the employee departure warning model, thereby improving the prediction accuracy of the employee departure warning model during training. Enhance the applicability of the employee departure warning model.

In some feasible implementation manners, after at least two types of employee turnover data corresponding to at least two types of termination response information collected from the above-mentioned multiple data acquisition paths, the corresponding employee attributes can be selected based on the acquired employee data. The sample data of the resignation employees is used to obtain the modeling data of the employee departure warning model. For the convenience of description, the sample data of the attributive employees corresponding to the above at least two types of termination response information will be described below by using the sample data of the first and second termination responses as examples. Among them, at least two types of turnover response sample data including the first departure response information and the second departure response information mentioned above include any type of turnover employee sample data corresponding to any type of turnover response information. Employee attribute data and turnover information. The above-mentioned sample data of the departing employees may include the sample data of the departing employees and / or the employees who have submitted the application for resignation but have not yet left, and at least the sample data of the first leaving employee and the second leaving response corresponding to the first departure response information. Sample data of the second departure employee corresponding to the information. For example, the first departure employee sample data corresponding to the above first departure response information may be employee sample data corresponding to a type of employee who has the departure reason as the first departure reason (for example, salary). The first departure response information may be Including retention schemes or stop-loss schemes for the resignation of retired employees such as salaries, or the deadline for dealing with the resignation of such employees. Among them, the above-mentioned transaction processing period may be a brewing period for employees who leave for this type of reason for resignation (that is, salary). During the brewing period, a retention plan or a stop-loss plan for employee resignation is adopted to handle retention or stop-loss transactions. duration. In other words, during the brewing period of employee turnover, a corresponding retention plan or stop loss plan can be taken in time to handle retention or stop loss matters in response to employee turnover.

Optionally, in some feasible implementation manners, the brewing period for employee turnover can be understood as the period from when the employee's intention to leave is revealed to when the employee leaves. Understandably, employee turnover is often not an immediate action made by the brain, but an idea that has occurred within a certain period of time, and then in the process of daily work, there have been frequent leave of absence, slow work and online job applications, such as job posting intentions. The performance of these performances until the employee leaves the post is also the period during which the employee is planning to leave. This period of time can be referred to as the period during which the employee is leaving. During the training of the employee departure warning model, the modeling data of different data types and / or data contents can be correspondingly trained to obtain an employee departure warning model suitable for outputting different lengths of the brewing period. In other words, the modeling data includes the departure period of a certain type of employees, and a set of network parameters can be obtained through training in this part of the modeling data, so that the employee departure warning model with such a set of network parameters can be based on The input employee attribute data corresponds to the ability to output the brewing period of employee turnover. At this time, the brewing period of employee turnover based on the employee departure warning model can be used as one of the employee's departure response information. Among them, the brewing period included in the above modeling data may exist through a label or a threshold. Based on the differences in the known brewing periods of different departing employees in the modeled data, the brewing periods in the modeled data can be divided into multiple categories according to one month, two months, three months, or half a year. The gestation period can be marked with labels of different values. For example, assuming that the form of the brewing period in the modeling data is a label, you can mark the brewing period of one month, two months, three months, or half a year by the value of the label 0, 1, 2, or 3, Therefore, multiple sets of network parameters can be obtained through training of modeling data containing multiple categories of gestation deadlines, so that the employee departure warning model can predict the employee attribute data of the input target employees to obtain the gestation deadlines of the target employees' departure. Coping with information.

Optionally, in some feasible implementation manners, the employee's departure is often due to more different reasons, different objective reasons, and / or different subjective reasons. Different reasons for leaving may determine whether employees' thoughts of leaving can be eliminated, whether employees can be retained, or whether the losses caused by employees leaving can be eliminated in a timely manner. During the training of the employee departure warning model, the modeling data of different data types and / or data contents can be correspondingly trained to obtain an employee departure warning model suitable for the output retention plan and / or the stop loss plan. In other words, the modeling data includes retention plans and / or stop-loss plans for different types of employees who are leaving based on different reasons for leaving, and a set of network parameters can be obtained through this part of the modeling data training, so that A set of network parameter employee departure warning models have the ability to output employee departure response information such as retention plans and / or stop-loss plans based on the input employee attribute data. Therefore, based on the above-mentioned employee departure warning model, multiple sets of network parameters can be obtained through training of modeling data of multiple retention plans and / or stop-loss plans, so that the employee departure warning model can predict any target employee data input. Get a retention plan and / or a stop loss plan for the target employee. Optionally, the modeling data of the above-mentioned employee departure warning model may further include employee data and reason for departure in historical data of employee departure within a certain period of time, including a retention plan and / or termination plan for employee departure. Loss scheme. In the same way, based on the same data processing method of the above-mentioned employee's departure method, the retention plan and / or stop loss plan included in the modeled data can also exist in the form of labels or thresholds. For details, refer to the foregoing implementation manners. More details. For the convenience of description, the information such as the brewing deadline, retention plan, and / or stop loss plan for the above employees will be collectively referred to as the termination information below.

In some feasible implementation manners, data other than the departure response information in the modeling data of the above-mentioned employee departure warning model may be referred to as employee attribute data. For example, for the first departure employee sample data corresponding to this type of departure employee whose reason for departure is the first departure reason, in addition to the first departure response information, there is also the attribute data of the first departure employee corresponding to the first departure employee ( For the convenience of description, it can be set as the attribute data of the first employee to leave). In the same way, for the second departure employee sample data for the second departure reason, the second departure employee sample data includes the second departure response information, as well as the second departure employee attribute data. (For the convenience of description, it can be set as the attribute data of the second departing employee), which will not be described below. The modeled data of the above employee departure warning model includes the sample data of the first employee and the sample data of the second employee. The sample data of the employee who responded to the departure information includes, but is not limited to, basic employee information and employee positions. One or more of one or more of the leaving attribute entries in the employee's working tenure, employee income level, employee performance, employee's latest promotion interval, employee reward and punishment program, employee home address, employee attendance, employee leave frequency, login job application website frequency Combined employee data. Wherein, the employee data obtained by combining one or more of the above-mentioned departure attribute entries is the employee attribute data of the departing employee. The above basic employee information may include, but is not limited to, personal identification information of employees (including but not limited to identity, ID, such as an ID card number, etc.), education and contact information, etc., and employee enterprise identification information (including but not limited to Title, position, or position level, etc.), employee employment history, employee personality, and employee communication mode, and other data types can be determined based on actual application scenarios, and there are no restrictions here. Among them, the above-mentioned employee attributes such as job title, employee tenure, employee income level, employee performance, employee attendance, employee leave frequency, and login job website frequency can be used as employee departure risk prediction parameters for employee departure response information prediction. The above employee's latest promotion interval (which can reflect the employee's promotion space), employee reward and punishment items, and the employee's home address (which can reflect the employee's time to commute) can be used as employee departure induction parameters for employee departure response information. No restrictions are predicted here.

In some feasible implementation manners, after obtaining the at least two types of resignation response information (including the first resignation employee sample data and the second resignation employee sample data), the corresponding resignation response information may be based on the respective resignation response information. The employee data is obtained from the sample data of the employee who has left the corresponding attributes of the employee, so as to obtain the sample data of the employee who corresponds to each of the information on the response to the departure, which is used as the modeling data of the employee departure warning model. In other words, the above sample data of the employees who left the corresponding response information is part of the data of the employees who left the corresponding response information. This part of the data can be filtered based on the above-mentioned attributes of the departure attributes, and can be obtained from the larger amount of data. More effective information is selected from the employee data as the data basis for the prediction of employee turnover information, which can reduce the amount of data processing for the prediction of employee turnover information, improve the accuracy of the prediction of employee turnover information, and have stronger applicability.

In some feasible implementation manners, based on the above-mentioned turnover attribute entries, the sample data of the turnover employees corresponding to each of the turnover response information is selected from the data of the turnover employees corresponding to each of the turnover response information as modeling data, and then the modeling data may be constructed based on the modeling data. The attribute characteristic pairs of departing employees. In the process of constructing the attributes of the resignation employee, the resignation response information corresponding to the sample data of each resignation employee can be labeled in the form of a label, and the characteristics corresponding to the resignation response information are generated based on the label. For example, the first departure response information may be based on the tag 1 tag, the second departure response information may be based on the tag 2 tag, and further, a feature corresponding to the first departure response information (such as a feature represented by a character “0”) may be generated based on the tag 1. The feature corresponding to the second departure response information based on the tag 2 (for example, a feature represented by a character “1”) is not limited herein. For another example, an attribute characteristic pair of a departing employee includes the attribute characteristic of a first employee corresponding to the first departure response information and the attribute characteristic of a second employee corresponding to the second departure response information. The attribute characteristics of the first departure employee corresponding to the first departure response information may be constructed from sample data of the first departure employee corresponding to the first departure response information. Correspondingly, the attribute characteristics of the second departure employee corresponding to the second departure response information may be constructed from sample data of the second departure employee corresponding to the second departure response information. In other words, it can be understood that the attribute characteristic of the first departure employee corresponding to the above first departure response information can be used as a positive sample feature in the attribute characteristic pair of the departure employee, and the attribute characteristic of the second departure employee corresponding to the above second departure response information can be used as The negative sample features in the attribute characteristic pairs of departing employees can be used to train the initial network model for the prediction of termination response information based on one positive and one negative sample characteristics to obtain the first departure response information or the second departure response information. Competitive employee departure warning model. Among them, the above-mentioned predicted departure turnover information based on the employee departure warning model is the first departure response information or the second departure response information is only an example, including but not limited to the first departure response information and the second departure response information, which may be specifically based on actual applications. The scene is determined, and there are no restrictions here.

In some feasible implementation manners, based on the attribute data of the resignation employee corresponding to any resignation response information, according to a preset abstraction rule of the attribute data of the resignation employee, an abstract feature representation of the attribute data of the resignation employee corresponding to each resignation attribute entry can be obtained, and further, According to the abstract features of the attribute data of the resignation employee corresponding to each resignation attribute entry, the attribute characteristics of the resignation employee corresponding to the resignation response information are obtained. For example, the attribute characteristics of the first departing employee may be represented by a multi-character feature vector, and the feature vector may be composed of three partial features. Among them, the above-mentioned three partial characteristics may include the basic information of the employee in the attribute data of the first employee, the employee turnover risk prediction parameter, and the employee's departure induction parameter. Among them, any one of the above three partial features may be composed of one or more characters, one or more sets of characters, and / or one or more dimensions of characters, and is not limited herein. For example, assuming that the above basic employee information can include 5 dimensions of employee personal identity information, employee corporate identity information, employee employment history, employee personality, and employee communication mode, 5 characters (or 5 groups of characters or 5 Dimension characters, etc. are not limited here) are used to represent the features abstracted from the basic information of employees. Wherein, each of the above 5 characters can represent information of one dimension. Among them, the information of each dimension can be classified separately, and different types of information (such as 0 or 1) are used to mark the information of different categories, and then the corresponding identifier of the information of each dimension can be obtained, so that the information of each dimension can be correspondingly The combination of the identifiers obtains the basic information characteristics of the employees containing 5 characters. For example, if the employee's personal identification number is used to represent the employee's personal identification information, it can be classified according to the year of the year in the ID number, and each age can be marked with a logo, which can be used in the above basic employee information The characteristics of the characters corresponding to the employee's personal identification information in the characteristics. By analogy, the characteristics of each character in the above-mentioned five-character employee basic information characteristics can be determined, so that the five-character employee basic characteristic information characteristics can be obtained. Here, for the dimensions of the default information in the above five dimensions of information, the characters corresponding to the dimensions may be filled in with blanks, etc. to construct the features corresponding to the dimensions, which is not limited here.

For the same reason, it is assumed that the employee turnover risk prediction parameters include the position of the employee (can be classified and identified according to the position, etc.), the term of the employee's work (can be classified and classified according to the length of the term, etc.), and the employee's income level (can be classified according to the level of income) Classification and identification, etc.), employee performance (classification and identification according to excellent, good or poor grades, etc.), employee attendance (classification and identification according to the number of days of absence, etc.), employee leave frequency (can be performed in accordance with frequency and segmentation) Category ID, etc.) and the frequency of login job search URLs (such as classification and identification based on frequency segmentation), and 7 dimensions of information, you can use 7 characters (or 7 groups of characters or 7 dimensions of characters, etc.), not here Restriction) represents the characteristics abstracted by the employee turnover risk prediction parameters. For the convenience of description, it may be referred to as the employee turnover risk prediction parameter characteristics, etc., and is not limited here. The implementation process of the above-mentioned employee turnover risk prediction parameter characteristics obtained by abstracting the above-mentioned employee turnover risk prediction parameter characteristics can refer to the corresponding implementation manner of the above-mentioned employee basic information characteristics, which is not limited here. Assume that the employee departure induction parameters include the employee's latest promotion interval (can be classified according to the promotion interval of different lengths, etc.), employee reward and punishment items (according to the category of employees' reward and punishment item classification, etc.), and the employee's home address (according to the employee The time-consuming time of commuting by segmentation, classification, identification, etc.), 3 dimensions of information, etc., can use 3 characters (or 3 groups of characters or 3 dimensions of characters, etc., without restrictions here) to indicate employee induction of departure The features abstracted by the parameters are simply referred to as the features of employee induction parameters for the convenience of description, and are not limited here.

S22. Training of employee departure warning model.

In some feasible implementation manners, based on the acquired attribute data of the employee departure warning model obtained based on the above-mentioned step S21, the attribute attribute pairs of the employee turnover may be input into the initial network model of the employee departure alert model. The above-mentioned initial network model is used to learn the data characteristics of the employee turnover included in the input employee attribute feature pair and the corresponding labeling characteristics of the termination response information to obtain employees having the ability to output the termination response information corresponding to any of the employee data characteristics. Departure warning model. Wherein, the feature data of the turnover employee included in the above modeling data includes the attributes of the first turnover employee attribute and the attributes of the second departure employee attribute, and the tag characteristics corresponding to each of the termination response information include at least the tags corresponding to the first termination response information. Characteristics, and the label characteristics corresponding to the above second departure response information. Here, the initial network model of the above employee departure warning model may be a backpropagation (BP) neural network model, or other types of neural network models, which are not limited herein. The activation function of the above employee departure warning model may be a sigmoid function, etc., which may be specifically determined according to an actual application scenario, and is not limited herein. Optionally, in some feasible implementation manners, the output of the above-mentioned employee departure warning model is a label or threshold corresponding to each departure response information, and further specific departure response information may be determined based on the label or threshold corresponding to each departure response information. It can be determined according to the actual application scenario, and there is no limitation here.

S23. Testing of the employee departure warning model.

In some feasible implementation manners, after the construction of the employee departure warning model is completed, after the selected time period of the modeling data of the above employee departure warning model is selected, employees who have left the employee in the period closest to the current time may be selected. Data is used as test data for departing employees. Here, the test data of the resignation employee includes at least one kind of test information of the resignation response information, and further, at least one test feature of the resignation employee may be constructed based on the test data of the at least one kind of resignation response information. Based on the test data of the above-mentioned employee turnover, the characteristics of the employee turnover test can be constructed, and then the accuracy of the departure response information prediction of the employee departure warning model can be tested based on the constructed employee test features. Among them, the departure attribute entries of the employee data included in the above-mentioned employee turnover test data, and the data types (or data dimensions) included in each of the departure attribute entries can be compared with the employees included in the modeling data of the employee departure warning model. The departure attribute entries of the data and the data types (or data dimensions) included in each departure attribute entry are the same. There is no restriction here, which can ensure the test effectiveness of the employee departure warning model and improve the test results of the employee departure warning model. Accuracy, enhancing the applicability of the employee departure warning model.

In some feasible implementation manners, based on the above-mentioned employee departure warning model, the characteristics of the departure employee test data constructed by the above-mentioned employee departure test data can be learned and the termination response information corresponding to the departure employee test data can be output, and then the employee departure alert model can be based on the foregoing employee departure warning model. Predict the termination response information for the test characteristics of the above employee departures, and combine the test departure data with the actual employee departure response information corresponding to the known employee departures to calculate the loss value of the output of the employee departure warning model (such as the label corresponding to the departure termination information and / Or the difference between the thresholds, etc., which is not limited here). The output loss value of the above-mentioned employee departure warning model can be fed back to the employee departure warning model, and the optimization of network parameters of the above employee departure warning model based on the above-mentioned loss value can improve the prediction accuracy and applicability of the employee departure warning model. Stronger.

In the embodiment of the present application, during the training of the employee departure warning model, the training samples are obtained from the sample data of various employees who left the company. Based on the attribute data of the employees who correspond to the various departure response information, a departure for model training can be constructed. Feature pairs of employee attributes, used to train the employee departure warning model, so that the model has the ability to output the corresponding response information for the employee attribute characteristics corresponding to any employee data, so that the employee departure corresponding to any employee data can be predicted based on the employee departure warning model. Response information, realizing effective prediction of employee turnover response information, which can enhance the response feasibility of employee turnover, reduce the human resource cost of employee management, increase the retention success probability of employee turnover, reduce the loss caused by employee turnover, and applicability high.

S3. Obtain the target employee attribute data corresponding to the above-mentioned departure attribute entry in the employee data of the target employee, and input the above target employee attribute data into the employee departure warning model.

In some feasible implementation methods, when the target company needs to monitor the status of employees in a certain part including the target employees (such as monitoring the departure status of employees who have been employed for half a year), they must grasp the employee's resignation dynamics and possible resignation responses. When information is obtained, the employee data of some employees (for example, target employees) of the target enterprise to be tested can be obtained, and based on the obtained employee data of some employees to be tested, the employee attribute characteristics corresponding to each employee can be constructed and the employee departure warning model can be input. Each employee attribute characteristic of the part to be tested is learned through the employee departure warning model (and / or processed, used, etc., for convenience of description, the following will take learning as an example), and the predicted departure response information of each employee is correspondingly output. For the convenience of description, the prediction of the turnover information of the target employee will be taken as an example.

In some feasible implementation manners, before obtaining the target employee attribute data of the target employee, a departure attribute entry for the employee departure warning may be determined first, and the target employee attribute data collected by the employee departure warning and the establishment of the employee departure warning model may be maintained. The data types of the model data are consistent, which can ensure the prediction accuracy of the above-mentioned employee turnover response information and enhance the reliability of the employee departure warning. The types of data included in the above-mentioned departure attribute entry may be specifically referred to the implementation manner provided in step S21 in the foregoing embodiment, and details are not described herein again.

In some feasible implementation manners, the employee data of the target employee may be obtained from the employee data system of the company to which the target employee belongs, and then the target employee attribute corresponding to the departure attribute entry may be obtained from the employee data of the target employee obtained above. data. Optionally, the above-mentioned employee data of the target employee may also be based on the collected instant messaging records of the target employee (for example, historical chat records based on instant messaging devices, or mobile phone SMS records, etc.) and / or the online data records of the target employee , Obtaining target employee attribute data corresponding to the above-mentioned departure attribute entry from the above instant messaging record and / or online data record. In specific implementation, the data type included in the target employee attribute data may be consistent with the data type of the attribute data of the terminated employee included in the modeling data of the employee departure warning model. For details, refer to the implementation manner provided in step S21 above. , Will not repeat them here. In the embodiment of the present application, the collection of employee data of the target employee is various, which can improve the richness of the employee data of the target employee, and further enhance the effectiveness of the target employee attribute data for the employee departure warning, thereby improving the employee departure. The accuracy of prediction of response information is more applicable.

S4. Determine target turnover response information corresponding to the target employee attribute data based on the above-mentioned employee turnover warning model.

In some feasible implementation manners, the implementation method of constructing the target employee attribute characteristics based on the employee attribute data of the target employee may be the same as the construction method of the attribute characteristics of the departing employee in the modeling data of the above-mentioned employee departure warning model. For details, refer to the above steps. The implementation manner provided by S21 is not repeated here. After the target employee attribute characteristics are constructed based on the target employee attribute data, the target employee attribute characteristics can be input into the employee departure warning model, the target employee attribute characteristics are learned based on the employee departure warning model, and the target employee correspondence is determined. Termination Response Information. Optionally, here, after the employee departure warning model learns the attribute characteristics of the target employee, it can output the target employee's departure response information label or the departure response information threshold value correspondingly, and then the target can be determined based on the above departure response information label or the departure response information threshold value. Staff turnover information. For example, when the employee departure warning model learns the attributes of the target employee and the corresponding output departure information label is label 1 or the departure response information threshold is threshold 1, it can be determined that the result of the prediction of the departure response information for the target employee is First departure response information. Or, after the employee departure warning model learns the attributes of the target employee, the output response information label corresponding to the output can be label 2 or the threshold value of the departure response information is threshold 2, then it can be determined that the prediction result of the target employee's departure response information is the first Second departure response information. By analogy, the possible termination response information of the target employee can be determined based on the termination response information label output by the employee departure warning model. In the embodiment of the present application, the data type and / or data content corresponding to the target employee attribute characteristics of the employee departure warning model is obtained by filtering and inputting based on the employee data of the target employee, and may be related to the training and / or testing stage of the employee departure warning model. The input modeling data and / or test data have the same data type and / or data content. The data types and / or data contents collected and filtered during the training, testing, and use phases of the employee departure warning model are the same, so that the employee departure warning model can better utilize the input employee's target employee attribute characteristics. Learning and outputting the corresponding departure response information can increase the accuracy of the prediction of the departure response information of the employee departure warning model.

In some feasible implementation manners, after outputting the departure response information of the target employee based on the above-mentioned employee departure warning model, the above-mentioned departure response information may be used to generate early-warning information of the employee departure and send it to the target employee for human resource management and / or Human resources risk management management of employee managers. The above-mentioned early-warning information may be an early-warning email of the employee, or an early-warning report file composed of the information about the employee's leaving response and its corresponding employee attribute data, etc., which can be determined according to the actual application scenario, and will not be done here. limit. The above employee management personnel may be the boss of the target employee, the department head of the department to which the target employee belongs, or the human resource supervisor and / or human resource risk control supervisor of the target enterprise, etc., which can be determined according to the actual application scenario, and there is no limitation here. The early-warning information generated through the above-mentioned turnover response information of the target employee can be used to prompt the employee management personnel to determine the source of the turnover risk of the target employee based on the predicted turnover response information in combination with the target employee's employee data. Based on the above sources of turnover risk, adopt the corresponding above-mentioned resignation countermeasures to eliminate the source of turnover risk and retain the target employees in time to avoid the loss of human resources of the company, or for the sources of turnover risk that cannot be eliminated, remind the staff management to do the employee turnover in a timely manner Remedial measures such as personnel replenishment or job adjustment to stop losses in a timely manner to avoid the operation status of the enterprise brought about by the departure of target employees, and the operation is flexible.

In the embodiment of the present application, the employee departure warning model may be constructed based on the employee data obtained from the enterprise employee database or the employee data obtained from multiple data paths such as enterprise employee data obtained from big data analysis. Based on the employee departure warning model, the employee data of the target employees of the target company monitored in real time can be used to predict the departure response information of the target employees. According to the target employee's departure response information, the target employee's turnover trends can be grasped in time to prevent and / or respond. The turnover of the target employee can reduce the human resource management cost of the enterprise, increase the retention success probability of the employee turnover, and then reduce the risk of the target employee's departure from bringing significant losses to the enterprise. The employee departure warning model provided in the embodiment of this application is constructed based on a large amount of employee data. Based on employee data in different fields and / or different industries, different models can be constructed to apply to employee departure warnings of enterprises in various fields or industries, with high flexibility. ,Wide range of applications.

Referring to FIG. 3, FIG. 3 is a schematic structural diagram of an early warning device for employee turnover provided by an embodiment of the present application. The early warning device for employee turnover provided in the embodiments of the present application includes:

A determining unit 31 is configured to determine a departure attribute entry for employee turnover warning.

A model constructing unit 32 is configured to obtain sample data of the terminated employee corresponding to the above-mentioned termination attribute entry determined by the determining unit 31, and construct a early warning model of the terminated employee according to the above-mentioned sampled employee data, where the sampled employee data includes at least the first Sample data of former employees and sample data of second employees. The above sample data of first employees includes the employee attribute data of the first type of employees and first departure response information, and the above sample data of the second employees includes the second type of employees. Employee attribute data and second departure response information for employees.

The data processing unit 33 is configured to obtain target employee attribute data corresponding to the above-mentioned departure attribute entry determined by the determining unit 31 among the employee data of the target employee, and input the above-mentioned target employee attribute data into the employee departure warning model.

The determining unit 31 is further configured to determine target turnover response information corresponding to the target employee attribute data input by the data processing unit based on the employee turnover warning model constructed by the model building unit, and the target turnover response information is used for the Alert the departure of target employees.

In some feasible implementation manners, the above-mentioned departure attribute entries include: employee basic information, employee position, employee working term, employee income level, employee performance, employee's latest promotion interval, employee reward and punishment program, employee home address, employee attendance , One or more combinations of employee leave frequency and login job frequency.

In some feasible implementation manners, the foregoing data processing unit 33 is configured to:

Obtaining the employee data of the target employee from the employee data system of the company to which the target employee belongs, and obtaining the target employee attribute data corresponding to the departure attribute entry from the employee data; and / or

Collecting an instant messaging record of the target employee and / or an online data record of the target employee, and acquiring target employee attribute data corresponding to the departure attribute entry from the instant communication record and / or the online data record.

In some feasible implementation manners, the model building unit 32 is configured to:

Acquiring data of at least two types of departure response information for employee departure warning training, wherein the at least two types of departure response information include at least the first departure response information and the second departure response information;

From the above-mentioned employee data, select the sample of employee data corresponding to the above-mentioned attributes of the employee, including any sample of employee response information including the above-mentioned first employee sample data and / or the second employee sample data The data includes employee attribute data and departure response information for a class of departing employees;

At least one attrition employee attribute characteristic pair is constructed based on the sample data of the at least two attrition employees corresponding to the above-mentioned resignation response information, and an early warning model of employee turnover is constructed according to the at least one attrition employee attribute characteristic pair.

In some feasible implementation manners, the model building unit 32 is configured to:

Obtain employee data of the employees who have left and / or employees who have submitted their application for resignation but have not yet resigned from the employee data system of the company to which the target employee belongs, and select from the obtained employee data the above first departure response information and the above second departure The employee data corresponding to at least two types of termination response information including the response information is used as the termination employee data for the employee departure warning training.

In some feasible implementation manners, the model building unit 32 is configured to:

Based on big data analysis, obtain employee data of employees who have left the company and / or employees who have submitted their application for resignation but have not yet resigned from the employee data of the same industry as the target employee's company;

From the obtained employee data, employee data corresponding to at least two types of resignation response information including the first resignation response information and the second resignation response information are selected as the resignation employee data used for employee resignation early warning training.

In some feasible implementation manners, the early warning device further includes:

An early warning information output unit 34, configured to generate early warning information for employee turnover based on the target departure response information determined by the determining unit 31, and output the early warning information to an employee manager who manages the target employee;

The above-mentioned early warning information includes early warning emails and / or early warning report files;

The above-mentioned early warning emails and / or early warning report files include one or more types of information on the deadlines for dealing with employee departures, retention plans for employee departures, and stop loss programs for employee departures.

In specific implementation, the above-mentioned early warning device for employee turnover can implement the implementation manners provided by the steps in FIG. 1 to FIG. 2 described above through each of its built-in functional modules. For example, the above-mentioned determining unit 31 may be used to perform the implementation methods such as determining the attribute of the departure attribute in each of the above steps, and the prediction of the employee departure response information. For details, refer to the implementation manners provided in the foregoing steps, and details are not described herein again. The above model building unit 32 may be used to execute the implementation manners described in the relevant steps in the construction of the employee departure warning model in the above steps. For details, refer to the implementation manners provided in the above steps, and details are not described herein again. The above-mentioned data processing unit 33 may be used to implement implementation methods such as collecting employee data, filtering employee attribute data, and constructing employee attribute characteristics in the foregoing steps. For details, refer to the implementation methods provided in the foregoing steps, and details are not described herein again. The above-mentioned early warning information output unit 34 may be used to implement the implementation methods of generating and outputting the early warning information of the employee departure warning in the foregoing embodiments. For details, refer to the implementation methods provided in the foregoing embodiments, and details are not described herein again.

In the embodiment of the present application, the early warning device for employee departure may build an early warning model for employee departure based on employee data obtained from multiple data paths, such as enterprise employee data obtained from an enterprise employee database or enterprise data obtained by big data analysis. Based on the employee departure warning model, the employee data of the target employees of the target company monitored in real time can be used to predict the departure response information of the target employees. According to the target employee's departure response information, the target employee's turnover trends can be grasped in time to prevent and / or respond. The turnover of the target employee can reduce the human resource management cost of the enterprise, increase the retention success probability of the employee turnover, and then reduce the risk of the target employee's departure from bringing significant losses to the enterprise. The employee departure warning model provided in the embodiment of this application is constructed based on a large amount of employee data. Based on employee data in different fields and / or different industries, different models can be constructed to apply to employee departure warnings of enterprises in various fields or industries, with high flexibility. ,Wide range of applications.

Referring to FIG. 4, FIG. 4 is a schematic structural diagram of a terminal device according to an embodiment of the present application. As shown in FIG. 4, the terminal device in this embodiment may include one or more processors 401 and a memory 402. The processor 401 and the memory 402 are connected via a bus 403. The memory 402 is configured to store a computer program. The computer program includes program instructions. The processor 401 is configured to execute the program instructions stored in the memory 402, and perform the following operations:

Determine the departure attribute entries for employee departure warnings;

Obtaining the sample data of the departing employee corresponding to the above-mentioned attributes of the leaving employee, and constructing an early warning model of the leaving employee based on the sample data of the leaving employee, wherein the sample data of the leaving employee includes at least the sample data of the first leaving employee and the sample data of the second leaving employee. The sample data of the first departure employee includes the employee attribute data and the first departure response information of the first type of employees, and the above sample data of the second departure employee includes the employee attribute data and the second departure response information of the second type of employees;

Obtaining target employee attribute data corresponding to the above-mentioned departure attribute entry in the employee data of the target employee, and inputting the above target employee attribute data into the employee departure early warning model;

Based on the above-mentioned employee turnover warning model, target turnover information corresponding to the target employee attribute data is determined, and the target turnover response information is used for early warning of the turnover of the target employee.

In some feasible implementation manners, the above-mentioned departure attribute entries include: employee basic information, employee position, employee working term, employee income level, employee performance, employee's latest promotion interval, employee reward and punishment program, employee home address, employee attendance , One or more combinations of employee leave frequency and login job frequency. Based on the departure attribute entry, more effective information can be selected from the employee data with a larger amount of data as the data basis for the employee departure response information prediction, thereby reducing the data processing amount of the employee departure response information prediction and improving the accuracy of the employee departure response information prediction. Rate, more applicable.

In some feasible implementation manners, the foregoing processor 401 is configured to:

Obtain the employee data of the target employee from the employee data system of the company to which the target employee belongs, and obtain the target employee attribute data corresponding to the above-mentioned departure attribute entry from the employee data; and / or

Collect the instant messaging record of the target employee and / or the online data record of the target employee, and obtain the target employee attribute data corresponding to the departure attribute entry from the instant communication record and / or the online data record.

In some feasible implementation manners, the foregoing processor 401 is configured to:

Acquiring data of at least two types of departure response information for employee departure warning training, wherein the at least two types of departure response information include at least the first departure response information and the second departure response information;

From the above-mentioned employee data, select the sample of employee data corresponding to the above-mentioned attributes of the employee, including any sample of employee response information including the above-mentioned first employee sample data and / or the second employee sample data The data includes employee attribute data and departure response information for a class of departing employees;

Construct at least one attribute attribute pair of the terminated employee based on the sample data of the at least two employee termination response information, wherein the attribute attribute pair of the at least one terminated employee attribute includes the attribute attribute of the employee of the first type of termination and the characteristic attribute of the second type of termination Employee attribute characteristics of employees,

An early warning model for employee turnover is constructed based on at least one of the attributes of the employee turnover.

In some feasible implementation manners, the foregoing processor 401 is configured to:

Obtain employee data of the employees who have left and / or employees who have submitted their application for resignation but have not yet resigned from the employee data system of the company to which the target employee belongs, and select from the obtained employee data the above first departure response information and the above second departure The employee data corresponding to at least two types of termination response information including the response information is used as the termination employee data for the employee departure warning training.

In some feasible implementation manners, the foregoing processor 401 is configured to:

Based on big data analysis, obtain employee data of employees who have left the company and / or employees who have submitted their application for resignation but have not yet resigned from the employee data of the same industry as the target employee's company;

From the obtained employee data, employee data including at least two types of departure response information including the above-mentioned first departure response information and the above-mentioned second departure response information is selected as the employee data used for employee departure warning training.

In some feasible implementation manners, the foregoing processor 401 is further configured to:

Generate early warning information for employee turnover based on the above target departure response information, and output the above warning information to the employee management personnel who manage the target employee;

The above warning information includes a warning email and / or a warning report file. The above-mentioned early warning emails and / or early warning report files include one or more types of information on the deadlines for dealing with employee departures, retention plans for employee departures, and stop loss programs for employee departures.

It should be understood that, in some feasible implementation manners, the processor 401 may be a central processing unit (CPU), and the processor may also be another general-purpose processor or a digital signal processor (DSP). , Application specific integrated circuit (ASIC), ready-made programmable gate array (field programmable gate array), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The memory 402 may include a read-only memory and a random access memory, and provide instructions and data to the processor 401. A part of the memory 402 may further include a non-volatile random access memory. For example, the memory 402 may also store information of a device type.

In specific implementation, the terminal device can implement the implementation manners provided by the steps in FIG. 1 to FIG. 2 through the built-in functional modules. For details, refer to the implementation manners provided in the foregoing steps, and details are not described herein again.

In the embodiment of the present application, the terminal device may construct an employee departure warning model based on employee data obtained from multiple data paths such as enterprise employee data obtained from big data analysis or enterprise employee data obtained by big data analysis. Based on the employee departure warning model, the employee data of the target employees of the target company monitored in real time can be used to predict the departure response information of the target employees. According to the target employee's departure response information, the target employee's turnover trends can be grasped in time to prevent and / or respond The turnover of the target employee can reduce the human resource management cost of the enterprise, increase the retention success probability of the employee turnover, and then reduce the risk of the target employee's departure from bringing significant losses to the enterprise. The employee departure warning model provided in the embodiment of this application is constructed based on a large amount of employee data. Based on employee data in different fields and / or different industries, different models can be constructed to apply to employee departure warnings of enterprises in various fields or industries, with high flexibility. ,Wide range of applications.

An embodiment of the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program. The computer program includes program instructions. When the program instructions are executed by a processor, each step in FIG. 1 to FIG. 2 is implemented. For the early warning method for employee turnover, please refer to the implementation methods provided in the above steps for details.

The computer-readable storage medium may be an early warning device for employee turnover provided by any of the foregoing embodiments or an internal storage unit of the terminal device, such as a hard disk or a memory of an electronic device. The computer-readable storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, Flash card, etc. Further, the computer-readable storage medium may include both an internal storage unit and an external storage device of the electronic device. The computer-readable storage medium is used to store the computer program and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been or will be output.

The terms "first", "second", "third", "fourth", etc. in the claims and the description of the present application and the drawings are used to distinguish different objects, and are not used to describe a specific order. Furthermore, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. Reference to "an embodiment" herein means that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the present application. The term "and / or" used in this specification and the appended claims refers to and includes any combination of one or more of the associated listed items and all possible combinations. The method and related devices provided in the embodiments of the present application are described with reference to the method flowchart and / or the structural schematic diagram provided in the embodiments of the present application, and each process of the method flowchart and / or the structural schematic diagram can be implemented by computer program instructions and / or Or a combination of blocks and processes and / or blocks in the flowcharts and / or block diagrams.

Claims (20)

1. An early warning method for employee turnover, characterized in that the method includes:

   Determine the departure attribute entries for employee departure warnings;

   Obtaining sample data of the departing employee corresponding to the departure attribute entry, and constructing a warning model of the leaving employee based on the sample data of the leaving employee, wherein the sample data of the leaving employee includes at least the sample data of the first leaving employee and the sample of the second leaving employee Data, the sample data of the first departure employee includes the employee attribute data of the first type of employees and the first departure response information, and the sample data of the second departure employee includes the employee attribute data of the second type of employees and the second Termination Response Information;

   Obtaining target employee attribute data corresponding to the departure attribute entry in the employee data of the target employee, and inputting the target employee attribute data into an employee departure warning model;

   Based on the employee departure warning model, target departure response information corresponding to the target employee attribute data is determined, and the target departure response information is used to provide early warning of the departure of the target employee.
2. The method according to claim 1, wherein the termination attribute entries include: employee basic information, employee position, employee working term, employee income level, employee performance, employee's latest promotion interval, employee reward and punishment items, One or more combinations of the employee's home address, employee attendance, employee leave frequency, and login job website frequency.
3. The method according to claim 1 or 2, wherein the obtaining the target employee attribute data corresponding to the departure attribute entry in the employee data of the target employee comprises:

   Obtaining the employee data of the target employee from the employee data system of the company to which the target employee belongs, and obtaining the target employee attribute data corresponding to the departure attribute entry from the employee data; and / or

   Collecting an instant messaging record of the target employee and / or an online data record of the target employee, and acquiring target employee attribute data corresponding to the departure attribute entry from the instant communication record and / or the online data record.
4. The method according to any one of claims 1 to 3, wherein the step of obtaining sample data of the terminated employee corresponding to the departure attribute entry, and constructing an early warning model of the terminated employee based on the sampled employee data include:

   Acquiring data on at least two types of departure response information for employee departure warning training, wherein the at least two types of departure response information include at least the first departure response information and the second departure response information;

   Select the sample data of the former employee corresponding to the last attribute attribute data from the former employee data, including any sample of the former employee and / or the second former employee sample data The data includes employee attribute data and departure response information for a class of departing employees;

   Construct at least one attribute attribute pair of the terminated employee based on the sample data of the at least two employees corresponding to the termination response information, and construct an early warning model of the employee termination based on the at least one attribute characteristic pair of the terminated employee.
5. The method according to claim 4, wherein the acquiring the data of the at least two types of termination response information used for the early departure training of the employee includes:

   Obtain the employee data of the employee who has left the company and / or the employee who has submitted a job application but has not left the company from the employee data system of the company to which the target employee belongs, and select from the obtained employee data to include the first departure response information and the The employee data corresponding to at least two types of termination response information including the second termination response information is used as the termination employee data for employee termination warning training.
6. The method according to claim 4, wherein the acquiring the data of the at least two types of termination response information used for the early departure training of the employee includes:

   Based on big data analysis, obtain employee data of employees who have left the company and / or employees who have submitted their application for resignation but have not yet resigned from the employee data of the same industry as the company to which the target employee belongs;

   From the acquired employee data, employee data corresponding to at least two types of termination response information including the first departure response information and the second departure response information is selected as the departure employee data used for employee departure warning training.
7. The method according to any one of claims 1-6, wherein after determining the target departure response information corresponding to the target employee attribute data, the method further comprises:

   Generating early warning information for employee turnover based on the target departure response information, and outputting the early warning information to an employee manager who manages the target employee;

   Wherein, the warning information includes a warning email and / or a warning report file;

   The early warning email and / or early warning report file includes one or more kinds of information in a transaction processing period for responding to employee turnover, a retention plan for responding to employee turnover, and a stop loss plan for responding to employee turnover.
8. An early warning device for employee turnover, characterized in that the early warning device includes:

   A determination unit for determining a departure attribute entry for employee departure warning;

   A model building unit is configured to obtain sample data of the terminated employee corresponding to the departure attribute entry determined by the determination unit, and construct a early warning model of the terminated employee based on the sampled employee data, wherein the sampled data of the terminated employee includes at least Sample data of the first departure employee and sample data of the second departure employee. The sample data of the first departure employee includes employee attribute data and first departure response information of the first type of termination employee, and the sample data of the second departure employee includes Employee attribute data and second departure response information for the second type of departing employees;

   A data processing unit, configured to obtain target employee attribute data corresponding to the departure attribute entry determined by the determining unit among the employee data of the target employee, and input the target employee attribute data into an employee departure warning model;

   The determining unit is further configured to determine target turnover response information corresponding to the target employee attribute data input by the data processing unit based on the employee turnover warning model constructed by the model building unit, and the target turnover response information It is used to warn the departure of the target employee.
9. The early warning device according to claim 8, wherein the data processing unit is configured to:

   Obtaining the employee data of the target employee from the employee data system of the company to which the target employee belongs, and obtaining the target employee attribute data corresponding to the departure attribute entry from the employee data; and / or

   Collecting an instant messaging record of the target employee and / or an online data record of the target employee, and acquiring target employee attribute data corresponding to the departure attribute entry from the instant communication record and / or the online data record.
10. The early warning device according to claim 8 or 9, wherein the model construction unit is configured to:

    Acquiring data on at least two types of departure response information for employee departure warning training, wherein the at least two types of departure response information include at least the first departure response information and the second departure response information;

    Select the sample data of the former employee corresponding to the last attribute attribute data from the former employee data, including any sample of the former employee and / or the second former employee sample data The data includes employee attribute data and departure response information for a class of departing employees;

    Construct at least one attribute attribute pair of the terminated employee based on the sample data of the at least two employees corresponding to the termination response information, and construct an early warning model of the employee termination based on the at least one attribute characteristic pair of the terminated employee.
11. The early warning device according to claim 10, wherein the model construction unit is configured to:

    Obtain the employee data of the employee who has left the company and / or the employee who has submitted a job application but has not left the company from the employee data system of the company to which the target employee belongs, and select from the obtained employee data to include the first departure response information and the The employee data corresponding to at least two types of termination response information including the second termination response information is used as the termination employee data for employee termination warning training.
12. The early warning device according to claim 10, wherein the model construction unit is configured to:

    Based on big data analysis, obtain employee data of employees who have left the company and / or employees who have submitted their application for resignation but have not yet resigned from the employee data of the same industry as the company to which the target employee belongs;

    From the acquired employee data, employee data corresponding to at least two types of termination response information including the first departure response information and the second departure response information is selected as the departure employee data used for employee departure warning training.
13. The early warning device according to any one of claims 8 to 12, wherein the early warning device further comprises:

    An early warning information output unit, configured to generate early warning information for employee departure based on the target departure response information determined by the determining unit, and output the early warning information to an employee manager who manages the target employee;

    Wherein, the warning information includes a warning email and / or a warning report file;

    The early warning email and / or early warning report file includes one or more kinds of information in a transaction processing period for responding to employee turnover, a retention plan for responding to employee turnover, and a stop loss plan for responding to employee turnover.
14. A terminal device, comprising a processor and a memory, and the processor and the memory are connected to each other;

    The memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to perform the following operations:

    Determine the departure attribute entries for employee departure warnings;

    Obtaining sample data of the departing employee corresponding to the departure attribute entry, and constructing a warning model of the leaving employee based on the sample data of the leaving employee, wherein the sample data of the leaving employee includes at least the sample data of the first leaving employee and the sample of the second leaving employee Data, the sample data of the first departure employee includes the employee attribute data of the first type of employees and the first departure response information, and the sample data of the second departure employee includes the employee attribute data of the second type of employees and the second Termination Response Information;

    Obtaining target employee attribute data corresponding to the departure attribute entry in the employee data of the target employee, and inputting the target employee attribute data into an employee departure warning model;

    Based on the employee departure warning model, target departure response information corresponding to the target employee attribute data is determined, and the target departure response information is used to provide early warning of the departure of the target employee.
15. The terminal device according to claim 14, wherein the processor is configured to:

    Obtaining the employee data of the target employee from the employee data system of the company to which the target employee belongs, and obtaining the target employee attribute data corresponding to the departure attribute entry from the employee data; and / or

    Collecting an instant messaging record of the target employee and / or an online data record of the target employee, and acquiring target employee attribute data corresponding to the departure attribute entry from the instant communication record and / or the online data record.
16. The terminal device according to claim 14 or 15, wherein the processor is configured to:

    Acquiring data on at least two types of departure response information for employee departure warning training, wherein the at least two types of departure response information include at least the first departure response information and the second departure response information;

    Select the sample data of the former employee corresponding to the last attribute attribute data from the former employee data, including any sample of the former employee and / or the second former employee sample data The data includes employee attribute data and departure response information for a class of departing employees;

    Construct at least one attribute attribute pair of the terminated employee based on the sample data of the at least two employees corresponding to the termination response information, and construct an early warning model of the employee termination based on the at least one attribute characteristic pair of the terminated employee.
17. The terminal device according to claim 16, wherein the processor is configured to:

    Obtain the employee data of the employee who has left the company and / or the employee who has submitted a job application but has not left the company from the employee data system of the company to which the target employee belongs, and select from the obtained employee data to include the first departure response information and the The employee data corresponding to at least two types of termination response information including the second termination response information is used as the termination employee data for employee termination warning training.
18. The terminal device according to claim 16, wherein the processor is configured to:

    Based on big data analysis, obtain employee data of employees who have left the company and / or employees who have submitted their application for resignation but have not yet resigned from the employee data of the same industry as the company to which the target employee belongs;

    From the acquired employee data, employee data corresponding to at least two types of termination response information including the first departure response information and the second departure response information is selected as the departure employee data used for employee departure warning training.
19. The terminal device according to any one of claims 14 to 18, wherein the processor is further configured to:

    Generating early warning information for employee turnover based on the target departure response information, and outputting the early warning information to an employee manager who manages the target employee;

    Wherein, the warning information includes a warning email and / or a warning report file;

    The early warning email and / or early warning report file includes one or more kinds of information in a transaction processing period for responding to employee turnover, a retention plan for responding to employee turnover, and a stop loss plan for responding to employee turnover.
20. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, the computer program includes program instructions, and when the program instructions are executed by a processor, the processor executes The method according to any one of 1-7 is required.

Images