Training method of course label determination model, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a course label determination model training method, electronic equipment and a storage medium. The method comprises the steps of obtaining a first to-be-trained set and a second to-be-trained set; wherein the first to-be-trained set comprises attribute information of a first course and label information of the first course, the second to-be-trained set comprises attribute information of a second course, and the label information represents categories of the courses; performing screening processing on attribute information of a first course in the first to-be-trained set to obtain a first target set, and determining label information of a second course in the second to-be-trained set to obtain a second target set; according to the first target set and the second target set, training an initial model to obtain a trained course label determination model; wherein the course label determination model is used for determining the label information of the course. According to the method, the effect of quickly and accurately determining the label information of the course according to the trained course label determination model is achieved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and in particular to a training method, electronic device and storage medium for a course label determination model. Background Art

[0002] With the development of Internet technology and digital technology, online learning platforms based on Internet technology have gradually been widely used in various enterprises. Enterprise online learning platforms have become an important way for internal training and internal knowledge sharing. Enterprise online learning platforms provide learning services for the company's employees. Employees can learn the courses that they should learn in their positions or at the stage of their tenure on the platform.

[0003] As the course development and system construction work progresses, more and more courses are accumulated on the company's online learning platform. When employees are selecting courses, it is difficult to quickly and accurately select the courses they need. Therefore, it is necessary to add course labels to the courses to distinguish the target audiences of the courses. Summary of the invention

[0004] The embodiments of the present application provide a training method, electronic device and storage medium for a course label determination model, which are used to obtain a trained course label model and quickly and accurately generate course labels for courses based on the course label model.

[0005] In a first aspect, an embodiment of the present application provides a training method for a course label determination model, comprising:

[0006] Obtain a first set to be trained and a second set to be trained; wherein the first set to be trained includes attribute information of the first course and label information of the first course, and the second set to be trained includes attribute information of the second course, and the label information represents the category of the course;

[0007] The attribute information of the first course in the first set to be trained is screened to obtain a first target set, and the label information of the second course in the second set to be trained is determined to obtain a second target set; wherein the first target set represents the first set to be trained after the attribute information is screened, and the second target set includes the attribute information of the second course and the label information of the second course;

[0008] The initial model is trained according to the first target set and the second target set to obtain a trained course label determination model; wherein the course label determination model is used to determine the label information of the course.

[0009] In a possible implementation, the attribute information includes the total number of employees, employee majors, and the number of majors, the total number of employees represents the total number of employees who have browsed the second course, the number of majors represents the number of employees belonging to the employee major among the employees who have browsed the second course, the label information of the second course represents the professional category label, and the professional category label represents the professional category to which the second course belongs; determining the label information of the second course in the second to-be-trained set includes:

[0010] According to the number of professionals corresponding to each employee's major, determine the professional category variance; wherein the professional category variance represents the degree of dispersion of the employee's major in the second course;

[0011] If the variance of the major category is greater than the preset variance threshold, the label information of the second course is determined according to the major of each employee, the number of majors corresponding to each employee's major, and the total number of employees.

[0012] In a possible implementation, the label information of the second course is determined according to the major of each employee, the number of professionals corresponding to each employee major, and the total number of employees, including:

[0013] According to the number of professionals corresponding to each employee's major, sort the majors of each employee to obtain the sorting results;

[0014] Determine an average value based on the total number of employees and the number of employee specialties; wherein the average value represents the average value of the number of employees under each employee specialty;

[0015] Determine the first control indicator corresponding to each employee's major according to the ranking result, the total number of employees, the average value, and the number of professionals corresponding to each employee's major; wherein the first control indicator represents whether the professional category label of the second course is an employee's major;

[0016] If the first control indicator is a preset first threshold, the employee's major is determined as the professional category label of the second course.

[0017] In a possible implementation, the first control indicator corresponding to each employee's major is determined according to the ranking result, the total number of employees, the average value, and the number of professionals corresponding to each employee's major, including:

[0018] If it is determined according to the sorting result that the employee's major is located in a preset sorting position, the first control indicator corresponding to each employee's major is determined according to the total number of employees, the average value, and the number of professionals corresponding to each employee's major.

[0019] In a possible implementation, the attribute information includes the total number of employees, the time when the employees join the company, and the length of time the employees study. The total number of employees represents the total number of employees who have browsed the second course, the length of time the employees study represents the length of time the employees browse the course, and the label information of the second course represents a new employee category label, and the new employee category label represents that the second course is a course provided to new employees. Determining the label information of the second course in the second set to be trained includes:

[0020] Get the current time;

[0021] Determine a second control indicator for the second course based on the current time, the total number of employees, the employee's entry time, and the employee's learning time; wherein the second control indicator indicates whether the second course is a course provided to new employees;

[0022] If the second control indicator is greater than a preset second threshold, it is determined that the new employee category label is label information of the second course.

[0023] In a possible implementation, the second control indicator of the second course is determined according to the current time, the total number of employees, the employee entry time, and the employee learning time, including:

[0024] Determine the time difference between the current time and the employee's employee entry time, and if the time difference is less than a preset difference threshold, determine that the employee is a new employee;

[0025] A second control indicator for the second course is determined based on the total number of employees, the employee learning hours of each employee, the number of new employees, and the employee learning hours of each new employee.

[0026] In a possible implementation, the attribute information includes the total number of employees, employee positions, and employee learning time. The total number of employees represents the total number of employees who have browsed the second course. The employee learning time represents the time employees have browsed the course. The label information of the second course represents the management position category label. The management position category label represents that the second course is a course provided to employees in management positions. Determining the label information of the second course in the second set to be trained includes:

[0027] If an employee is determined to be in a management position based on his / her position, the sum of the learning hours of all employees in management positions is determined as the time;

[0028] Determine a third control indicator of the second course according to the number of employees in management positions, the total number of employees, the learning time of each employee, and the time; wherein the third control indicator indicates whether the second course is a course provided to employees in management positions;

[0029] If the third control indicator is greater than a preset third threshold, it is determined that the management position category label is the label information of the second course.

[0030] In a possible implementation, the attribute information of the first course in the first to-be-trained set is screened to obtain a first target set, including:

[0031] Determine the confidence between each attribute information and each label information according to the attribute information and label information of all the first courses in the first to-be-trained set; wherein the confidence represents the degree of association between the attribute information and the label information;

[0032] If the confidence is less than a preset confidence threshold, the attribute information is deleted to obtain a first target set.

[0033] In a possible implementation, determining the confidence between each attribute information and each label information according to the attribute information and label information of all first courses in the first to-be-trained set includes:

[0034] Arrange and combine the attribute information and label information of all the first courses in the first set to be trained to obtain at least one information group; wherein the information group includes one attribute information and one label information;

[0035] Determine each piece of label information of the first to-be-trained set as a target label, and determine each piece of attribute information of the first to-be-trained set as a target attribute;

[0036] Determine the number of information groups containing target labels and target attributes, which is a first number, and the number of target labels in the first set to be trained, which is a second number;

[0037] A confidence level between the target attribute and the target label is determined based on the first quantity and the second quantity.

[0038] In a possible implementation, obtaining a second set to be trained includes:

[0039] Get courses that only contain attribute information, which are the courses to be screened out;

[0040] Determine the number of views, number of employees who viewed the courses to be screened out, course duration, and employee learning time within a preset time period;

[0041] Determine the fourth control index of the course to be screened out according to the number of views of the course to be screened out, the number of employees who have viewed the course, the course duration, and the employee learning duration; wherein the fourth control index represents the importance of the course to be screened out;

[0042] If the fourth control indicator is greater than a preset fourth threshold, the course to be screened out is determined to be the second course.

[0043] In a possible implementation, it further includes:

[0044] Get the attribute information of the course to be classified;

[0045] The attribute information of the courses to be classified is input into the course label determination model to obtain the output label information of the courses to be classified.

[0046] In a second aspect, an embodiment of the present application provides a training device for a course label determination model, comprising:

[0047] An acquisition unit is used to acquire a first set to be trained and a second set to be trained; wherein the first set to be trained includes attribute information of the first course and label information of the first course, and the second set to be trained includes attribute information of the second course, and the label information represents the category of the course;

[0048] A processing unit is used to screen the attribute information of the first course in the first set to be trained to obtain a first target set, and to determine the label information of the second course in the second set to be trained to obtain a second target set; wherein the first target set represents the first set to be trained after the attribute information is screened, and the second target set includes the attribute information of the second course and the label information of the second course;

[0049] The training unit is used to train the initial model according to the first target set and the second target set to obtain a trained course label determination model; wherein the course label determination model is used to determine the label information of the course.

[0050] In a third aspect, an embodiment of the present application provides a training device for a course label determination model, including: a memory, a processor;

[0051] The memory stores computer-executable instructions;

[0052] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and / or various possible implementations of the first aspect.

[0053] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the first aspect above and / or various possible implementations of the first aspect.

[0054] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above first aspect and / or various possible implementation methods of the first aspect.

[0055] The training method, electronic device and storage medium of the course label determination model provided in the embodiment of the present application are obtained by obtaining a first set to be trained containing the attribute information of the first course and the label information of the first course, and obtaining a second set to be trained containing the attribute information of the second course. The attribute information of the first course in the first set to be trained is screened to obtain a first target set; the label information of the second course in the second set to be trained is determined to obtain a second target set. The initial model is trained using the first target set and the second target set to obtain a trained course label determination model. By using the trained course label model, the label information corresponding to the course can be quickly generated for courses without label information. By screening the first set to be trained to obtain the first target set, the association between the attribute information and the label information of the first course can be made more accurate; by determining the course label information corresponding to the second course in the second set to be trained to obtain the second target set, the course label information corresponding to the second course can be completed; by using the first target set and the second target set, the initial model is trained to improve the training accuracy of the model, which can make the recognition of the model more accurate and improve the recognition efficiency of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0057] Figure 1 A flowchart of a training method for a course label determination model provided in an embodiment of the present disclosure;

[0058] Figure 2 A flowchart of a training method for a course label determination model provided in an embodiment of the present disclosure;

[0059] Figure 3 A flowchart of a training method for a course label determination model provided in an embodiment of the present disclosure;

[0060] Figure 4 A flowchart of a training method for a course label determination model provided in an embodiment of the present disclosure;

[0061] Figure 5 A flowchart of a training method for a course label determination model provided in an embodiment of the present disclosure;

[0062] Figure 6 A flowchart of a training method for a course label determination model provided in an embodiment of the present disclosure;

[0063] Figure 7 A structural block diagram of a training device for a course label determination model provided by an embodiment of the present disclosure;

[0064] Figure 8 A structural block diagram of a training device for a course label determination model provided by an embodiment of the present disclosure;

[0065] Fig. 9 A structural block diagram of an electronic device provided in an embodiment of the present disclosure;

[0066] Fig.10 It is a block diagram of an electronic device according to an exemplary embodiment.

[0067] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0068] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0069] First, the terms involved in this application are explained:

[0070] Enterprise online learning platform: Enterprise online learning platform refers to a platform that provides online learning resources and services for enterprise employees. It is usually built based on cloud computing and big data technology, and supports mobile learning, personalized learning path planning, rich and diverse learning resources, etc. These platforms are designed to help enterprises improve employee skills, promote knowledge sharing, and improve overall operational efficiency.

[0071] With the development of Internet technology, enterprise online learning platforms have become an important way to conduct training and education within enterprises. The characteristics of enterprise online learning platforms are that users are company employees. Since each employee has a different position, the employees who browse the courses on the platform are not fixed, and most training cycles are short. Therefore, course tags are needed to help users quickly filter so that users can quickly find the courses they need, and it is also convenient for staff to manage according to tags.

[0072] At present, with the development of enterprise online learning platforms and the progress of course development and other work, more and more courses are accumulated in the enterprise online learning platforms. Due to the lack of means to quickly determine the course labels corresponding to the courses, a large number of courses have missing course labels. When setting course labels for courses, course labels are usually set according to the content of the courses, and the perspective of the service objects of the courses is not considered, so that course labels cannot help the service objects of the courses quickly find the required courses.

[0073] The training method, electronic device and storage medium for the course label determination model provided in this application are intended to solve the above technical problems of the prior art.

[0074] The application scenario of the present application is an enterprise online learning platform, and is specifically used to train a course label determination model, which is used to determine the course label information of courses in the enterprise online learning platform. The present application obtains a first set to be trained that contains attribute information of a first course and label information of the first course, and obtains a second set to be trained that contains attribute information of a second course, and screens and processes the attribute information of the first course in the first set to be trained to obtain a first target set, determines the label information of the second course in the second set to be trained to obtain a second target set, and uses the first target set and the second target set to train the initial model to obtain a trained course label determination model.

[0075] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0076] Figure 1 A flowchart of a method for training a course label determination model provided in an embodiment of the present disclosure is provided. The method can be executed by a training device for a course label determination model.

[0077] like Figure 1 As shown, the method comprises the following steps:

[0078] S101. Obtain a first set to be trained and a second set to be trained; wherein the first set to be trained includes attribute information of a first course and label information of the first course, and the second set to be trained includes attribute information of a second course, and the label information represents the category of the course.

[0079] Exemplarily, the course is a multimodal content course in an enterprise learning platform, such as a video course or a text course.

[0080] The attribute information of a course may include course characteristic attributes, course historical learning behavior characteristic attributes, and course staff characteristic attributes. The course characteristic attributes represent the basic characteristics of the course, and the course characteristic attributes include at least one of the attributes such as the course number, course duration, course online time, and course keywords. The course number is the number of the course in the enterprise learning platform; the course duration is the length of the course; the course online time is the upload time of the course in the enterprise learning platform; and the course keywords are the keywords in the course name.

[0081] The historical learning behavior characteristic attributes of a course represent the characteristics of the learning behavior of employees who have studied the course within a preset time period, and the historical learning behavior characteristic attributes of the course include at least one of the attributes of total learning time, total learning times, completion rate, average learning time per person, and average learning times per person within the preset time period. Among them, the total learning time is the total browsing time of employees who have browsed the course; the total learning times is the number of views of all employees who have browsed the course; the completion rate is the proportion of employees who have browsed the course and browsed the entire course content; the average learning time per person is the average browsing time of employees who have browsed the course; and the average learning times per person is the average number of views of employees who have browsed the course.

[0082] The course employee characteristic attribute represents the basic information of the employees who have studied the course within a preset time period. The course employee characteristics include at least one of the following attributes: employee position type, employee number, and keywords of the courses studied by the employee. Among them, employee position type refers to the position type of the employees who have browsed the course; employee number refers to the number or sequence of the employees who have browsed the course; and keywords of the courses studied by the employee refers to the keywords of the names of other courses browsed by the employees who have browsed the course.

[0083] The label information of a course represents the category of the course, that is, the classification of the course. The course label can be a label set according to the course content, such as workplace etiquette; it can also be a label set according to the service object of the course, that is, according to the employees who study the course, such as computer position course. The correspondence between the course and the course label information is a one-to-many correspondence, that is, one course can correspond to multiple course label information.

[0084] The enterprise learning platform is provided with a historical database, which is used to store the attribute information and label information of the courses and courses in the enterprise learning platform. The first set to be trained and the second set to be trained are obtained from the historical database of the learning platform. For example, the course attribute information and label information of all courses can be obtained from the historical database of the learning platform, and the course with course label information is determined as the first course, and the attribute information and course label information of the first course are determined as the first set to be trained; the course without course label information is determined as the second course, and the attribute information of the second course is determined as the second set to be trained. The first set to be trained includes the attribute information of the first course and the label information of the first course, and the second set to be trained includes the attribute information of the second course.

[0085] S102. Filter the attribute information of the first course in the first set to be trained to obtain a first target set, and determine the label information of the second course in the second set to be trained to obtain a second target set; wherein the first target set represents the first set to be trained after the attribute information is filtered, and the second target set includes the attribute information of the second course and the label information of the second course.

[0086] Exemplarily, the attribute information of the first course in the first set to be trained is screened to obtain the first target set. For example, it can be a preset number threshold, and based on the first set to be trained, the attribute information of all the first courses in the first set to be trained is determined, and the number of times each attribute in the attribute information appears is determined. For any attribute in the attribute information, if the number of times the attribute appears is less than the preset number threshold, the attribute is judged to be an irrelevant attribute, and the attribute is deleted, and the first set to be trained with all irrelevant attributes deleted is determined as the first target set. The preset number threshold can be, for example, 2 times; the first target set represents the first set to be trained after the attribute information is screened.

[0087] Determine the label information of the second course in the second set to be trained to obtain the second target set. For example, the course employee characteristic attributes of each second course in the second set to be trained can be determined based on the attribute information of the second set to be trained. For any second course, determine the professional categories of employees who have browsed the course and the number of views by employees of each professional category, that is, the number of times employees of each professional category have studied the second course based on the course employee characteristic attributes. Based on the number of views by employees of each professional category, determine the professional category with the most views as the label information of the second course. Determine the attribute information of the second course and the label information of the second course as the second target set.

[0088] S103. Train the initial model according to the first target set and the second target set to obtain a trained course label determination model; wherein the course label determination model is used to determine label information of the course.

[0089] Exemplarily, an initial model is pre-set, and the initial model is a preset neural network model.

[0090] The initial model is trained according to the first target set and the second target set. For example, a loss function is preset. The preset loss function is used to evaluate the difference between the prediction result of the initial model and the actual label information of the course. The preset loss function can be, for example, a mean square error formula. The attribute information of the courses in the first target set and the second target set can be used as the independent variable for model training, the label information of the courses in the first target set and the second target set can be used as the dependent variable, and the preset loss function can be used to train the initial model. The attribute information of a part of the courses is used as input, and the predicted label information is obtained through model calculation. The difference between the predicted label information and the actual label information in the first target set and the second target set is calculated using the preset loss function as the loss value.

[0091] According to the loss value, the gradient of each parameter in the initial model is calculated, and the parameters in the model are updated using the gradient descent method, and the initial model is iterated. When the error rate between the label information predicted by the model and the actual label information is lower than the preset error rate standard, the iterated initial model is determined as the trained initial model, that is, the trained course label determination model. The course label determination model is used to determine the label information of the course, that is, the course label determination model can determine the label information of the course according to the attribute information of the course.

[0092] In this embodiment, there is no specific limitation on the preset loss function, and there is no specific limitation on the type and model structure of the preset neural network model.

[0093] In this embodiment, it also includes: obtaining attribute information of the courses to be classified; inputting the attribute information of the courses to be classified into the course label determination model to obtain output label information of the courses to be classified.

[0094] Specifically, the courses to be classified are obtained, and the attribute information of the courses to be classified is obtained. The courses to be classified may be courses without label information, or courses whose label information quantity is lower than a preset label quantity threshold. The label quantity threshold may be, for example, 3.

[0095] The attribute information of the courses to be classified is input into the course label determination model, and the course label determination model determines the label according to the attribute information of the courses to be classified, and outputs the label information of the courses to be classified.

[0096] The beneficial effect of this arrangement is that the trained course label determination model is used to determine the course labels of the courses to be classified, and the label information of the courses to be classified is obtained, which makes it easier for employees to find the courses they need according to the course label information, and also makes it easier for staff to manage courses according to the course labels.

[0097] The training method, electronic device and storage medium of the course label determination model provided in the embodiment of the present application are obtained by obtaining a first set to be trained containing the attribute information of the first course and the label information of the first course, and obtaining a second set to be trained containing the attribute information of the second course. The attribute information of the first course in the first set to be trained is screened to obtain a first target set; the label information of the second course in the second set to be trained is determined to obtain a second target set. The initial model is trained using the first target set and the second target set to obtain a trained course label determination model. By using the trained course label model, the label information corresponding to the course can be quickly generated for courses without label information. By screening the first set to be trained to obtain the first target set, the association between the attribute information and the label information of the first course can be made more accurate; by determining the course label information corresponding to the second course in the second set to be trained to obtain the second target set, the course label information corresponding to the second course can be completed; by using the first target set and the second target set, the initial model is trained to improve the training accuracy of the model, which can make the recognition of the model more accurate and improve the recognition efficiency of the model.

[0098] Figure 2 A flowchart of a training method for a course label determination model provided in an embodiment of the present disclosure.

[0099] In this embodiment, the attribute information includes the total number of employees, employee majors, and the number of majors. The total number of employees represents the total number of employees who have browsed the second course, the number of majors represents the number of employees who belong to the employee major among the employees who have browsed the second course, and the label information of the second course represents the professional category label, and the professional category label represents the professional category to which the second course belongs; determining the label information of the second course in the second to-be-trained set includes: determining the professional category variance according to the number of majors corresponding to each employee major; wherein the professional category variance represents the degree of dispersion of the employee majors in the second course; if the professional category variance is greater than a preset variance threshold, determining the label information of the second course according to each employee major, the number of majors corresponding to each employee major, and the total number of employees.

[0100] like Figure 2 As shown, the method comprises the following steps:

[0101] S201. Obtain a first set to be trained and a second set to be trained; wherein the first set to be trained includes attribute information of a first course and label information of the first course, and the second set to be trained includes attribute information of a second course, and the label information represents the category of the course.

[0102] Exemplarily, this step may refer to the above-mentioned step S101 and will not be described in detail.

[0103] S202: Filter the attribute information of the first course in the first set to be trained to obtain a first target set.

[0104] Exemplarily, this step may refer to the above-mentioned step S102 and will not be described in detail.

[0105] S203. Determine the professional category variance according to the number of professionals corresponding to each employee's major; wherein the professional category variance represents the degree of dispersion of the employee's major in the second course.

[0106] Exemplarily, the attribute information of the second course includes the total number of employees, employee majors, and the number of majors. The total number of employees represents the total number of employees who have browsed the second course within a preset time period; the employee major represents the professional category of the employees who have browsed the second course, and the professional category can be, for example, digital professional line, network professional line, operation and maintenance professional line, customer service professional line, and marketing professional line; the number of majors represents the number of employees belonging to each employee major among the employees who have browsed the second course. The label information of the second course represents the professional category label, that is, the professional category to which the second course belongs.

[0107] According to the number of professionals corresponding to each employee's major, determine the professional category variance. For example, for any second course, determine that the number of employees majoring in the second course is K; the symbol of each employee major is k a , a is the number of the employee major in the total employee majors of the second course, for example, k1 represents the symbol of the first major in the employee majors of the second course, k K The symbol representing the last major in the second course of employee majors. The number is only used to distinguish majors. For any employee major k a , the number of professionals corresponding to this employee's major is kn i,a , that is, the number of professionals corresponding to the major of employees with number a in the course with number i, for example, kn 3,,1 is the number of students in the first major of the course numbered 3 in the second course; i is the course number in the second course, the minimum value of i is 1, that is, the first course in the second course, and the maximum value of i is I, that is, the last course in the second course. The number is only used to distinguish courses.

[0108] The professional category variance of the second course may be determined based on a preset variance determination formula. The preset variance determination formula may be, for example:

[0109]

[0110] Where D() is the symbol for variance calculation; E() is the symbol for expectation calculation; C i Represents the second course with course number i; this formula represents that the variance of the professional category of the second course is the variance of the number of majors corresponding to each major in the second course, and the variance of the number of majors corresponding to each major in the second course is equal to the expected square of the number of majors minus the square of the expected number of majors.

[0111] The beneficial effect of this setting is that by calculating the variance of the number of people in each major in the second course as the variance of the major category of the second course, the degree of dispersion of the number of people in each major of the second course can be effectively obtained, and the subsequent determination of the major category label of the second course can be determined based on the degree of dispersion, making the determined major category label more accurate.

[0112] S204: If the variance of the professional category is greater than a preset variance threshold, the label information of the second course is determined according to the major of each employee, the number of majors corresponding to each employee's major, and the total number of employees.

[0113] Exemplarily, a variance threshold is set in advance, and the variance threshold is used to determine whether the number of majors corresponding to the employee majors of the second course is discrete. If the professional category variance is greater than the preset variance threshold, the number of majors corresponding to the employee majors of the second course is sufficiently discrete, that is, the number of majors corresponding to the employee majors of the second course has a tendency, and the employees of the second course are mainly distributed in individual majors in the second course. The professional category label of the second course can be determined according to the number of majors corresponding to the employee majors of the second course, and then the label information of the second course can be determined.

[0114] The label information of the second course is determined based on each employee's major, the number of professionals corresponding to each employee's major, and the total number of employees. For example, the ratio of the number of professionals corresponding to each employee's major to the total number of employees can be determined. If the ratio of the number of professionals corresponding to each employee's major to the total number of employees exceeds the preset employee major ratio threshold, the corresponding employee major is determined as the second course professional category label, that is, the corresponding employee major is determined as the label information of the second course. The preset employee major ratio threshold is used to determine whether the employee major is the label information of the second course, and the preset employee major ratio threshold can be, for example, 80%.

[0115] The beneficial effect of this setting is that since the number of professionals corresponding to the employee majors is relatively balanced and the number of employee majors is not concentrated in individual majors, it is difficult to determine the professional category label corresponding to the course. According to the preset variance threshold, the professional category variance is verified, which can effectively filter out the second course with a relatively balanced distribution of professional personnel corresponding to the employee major, thereby improving the accuracy and reliability of the professional category label of the determined second course.

[0116] In this embodiment, according to each employee's major, the number of professionals corresponding to each employee's major, and the total number of employees, the label information of the second course is determined, including: sorting each employee's major according to the number of professionals corresponding to each employee's major to obtain a sorting result; determining an average value according to the total number of employees and the number of employee majors; wherein the average value represents the average value of the number of employees under each employee's major; determining the first control indicator corresponding to each employee's major according to the sorting result, the total number of employees, the average value, and the number of professionals corresponding to each employee's major; wherein the first control indicator represents whether the professional category label of the second course is an employee's major; if the first control indicator is a preset first threshold, the employee's major is determined as the professional category label of the second course.

[0117] Specifically, for any second course, the employee majors are sorted in descending order according to the number of majors corresponding to each employee major to obtain a sorting result. According to the total number of employees and the number of employee majors, the total number of employees is divided by the number of employee majors to obtain an average value, that is, the average number of employees under each employee major.

[0118] The first control index corresponding to each employee's major is determined based on the sorting results, the total number of employees, the average value, and the number of professionals corresponding to each employee's major. For example, based on the sorting results, the number of professionals corresponding to the employee's major with the largest number of professionals is divided by the total number of employees to obtain the ratio of the number of professionals in the employee's major with the largest number of professionals to the total number of employees, and the ratio is multiplied by the average value. If the product is greater than a preset product threshold, the value of the first control index is 1, otherwise it is 0.

[0119] If the first control indicator is 1, that is, the proportion of the number of employees' majors corresponding to the largest number of professionals in the total number of employees in the first course is sufficient to prove the degree of correlation between the majors of the largest number of professionals and the second course, then the employee major will be determined as the professional category label of the second course.

[0120] The beneficial effect of this setting is that it fully considers whether the degree of correlation between the employee's major and the second course meets the requirements based on the sorting results, the total number of employees, the average value and other values. If it meets the requirements, the corresponding employee's major is determined as the professional category label.

[0121] In this embodiment, the first control indicator corresponding to each employee's major is determined based on the sorting results, the total number of employees, the average value, and the number of professionals corresponding to each employee's major, including: if it is determined that the employee's major is located in a preset sorting position based on the sorting results, then the first control indicator corresponding to each employee's major is determined based on the total number of employees, the average value, and the number of professionals corresponding to each employee's major.

[0122] Specifically, for any second course, a sorting position is preset, and the preset sorting position is used to determine whether the number of professionals corresponding to the employee's major meets the requirements for determining the first control indicator. The preset sorting position can be, for example, 3, that is, the employee majors ranked in the top three positions in the sorting results.

[0123] If it is determined that the employee's major is in a preset ranking position according to the ranking result, the first control index corresponding to each employee's major is determined according to the total number of employees, the average value, and the number of professionals corresponding to each employee's major. For example, the first control index corresponding to each employee in the preset ranking position can be determined based on a preset first index determination formula, and the preset first index determination formula can be, for example:

[0124]

[0125] Among them, X i,a is the ranking index, that is, whether the number of employees with the corresponding major numbered a in the second course numbered i meets the requirements, that is, X i,a When it is equal to 1, in the second course numbered i, the number of employees with the major numbered a ranks in the top M of the number of employees with the major numbered a in the second course, X i,a When it is equal to 0, in the second course numbered i, the number of employees with the major numbered a is outside the Mth position of the number of employees with the major numbered a in the second course. M represents the preset sorting position, Top M (kn i,a ) represents the number of professionals in the preset sorting position of the employee major numbered a in the second course numbered i. For example, if M is 3, then Top3(kn i,a ) represents the number of professionals corresponding to the third-ranked employee major in the sorting results. This formula means that if the number of professionals is greater than the number of professionals at the preset sorting position, the sorting index is 1, otherwise it is 0. This formula is used to determine whether the employee major in the sorting results is at the preset sorting position.

[0126] f1 represents the first control indicator; J represents the total number of employees in the second course; A1 is the preset first calculation coefficient, which is used to adjust the calculation result of the first control indicator; E(kn i,a) is the average number of professionals in the employee major of the second course. This formula represents that the first control index is equal to the product of the ratio of the number of professionals to the total number of employees and the preset first calculated ratio and the average value, and the intersection of the product and 1 is calculated, and the result of the intersection is multiplied by the ranking index, which is the first control index.

[0127] The beneficial effect of such a setting is that, according to the sorting index, it is possible to quickly determine whether the employee's major is in the preset sorting position, and the sorting index is 0 or 1, which can simplify the subsequent determination formula of the first indicator. The first control indicator is determined according to the total number of employees in the second course, the average value, etc., so that the first control indicator can more accurately reflect the correspondence between the employee's major and the second course, making the determination of the course label information more accurate and reliable.

[0128] S205. Train the initial model according to the first target set and the second target set to obtain a trained course label determination model; wherein the course label determination model is used to determine label information of the course.

[0129] Exemplarily, this step may refer to the above-mentioned step S103 and will not be described in detail.

[0130] The training method, electronic device and storage medium of the course label determination model provided in the embodiment of the present application are obtained by obtaining a first set to be trained containing the attribute information of the first course and the label information of the first course, and obtaining a second set to be trained containing the attribute information of the second course. The attribute information of the first course in the first set to be trained is screened to obtain a first target set; the label information of the second course in the second set to be trained is determined to obtain a second target set. The initial model is trained using the first target set and the second target set to obtain a trained course label determination model. By using the trained course label model, the label information corresponding to the course can be quickly generated for courses without label information. By screening the first set to be trained to obtain the first target set, the association between the attribute information and the label information of the first course can be made more accurate; by determining the course label information corresponding to the second course in the second set to be trained to obtain the second target set, the course label information corresponding to the second course can be completed; by using the first target set and the second target set, the initial model is trained to improve the training accuracy of the model, which can make the recognition of the model more accurate and improve the recognition efficiency of the model.

[0131] Figure 3 A flowchart of a training method for a course label determination model provided in an embodiment of the present disclosure.

[0132] In this embodiment, the attribute information includes the total number of employees, the time when employees join the company, and the length of time when employees study. The total number of employees represents the total number of employees who have browsed the second course, the length of time when employees study represents the length of time when employees browse the course, and the label information of the second course represents the category label of new employees, and the category label of new employees represents that the second course is a course provided to new employees. Determining the label information of the second course in the second set to be trained includes: obtaining the current time; determining a second control indicator of the second course according to the current time, the total number of employees, the time when employees join the company, and the length of time when employees study; wherein the second control indicator represents whether the second course is a course provided to new employees; if the second control indicator is greater than a preset second threshold, determining that the category label of new employees is the label information of the second course.

[0133] like Figure 3 As shown, the method comprises the following steps:

[0134] S301, obtaining a first set to be trained and a second set to be trained; wherein the first set to be trained includes attribute information of a first course and label information of the first course, and the second set to be trained includes attribute information of a second course, and the label information represents the category of the course.

[0135] Exemplarily, this step may refer to the above-mentioned step S101 and will not be described in detail.

[0136] S302: Filter the attribute information of the first course in the first set to be trained to obtain a first target set.

[0137] Exemplarily, this step may refer to the above-mentioned step S102 and will not be described in detail.

[0138] S303: Get the current time.

[0139] Exemplarily, the time at the current time point is obtained as the current time.

[0140] The beneficial effect of such a setting is that it can determine in real time whether an employee is a new employee based on the current time and the employee's entry time, making the process of determining new employees real-time and making the determined new employee identity of the employee more accurate.

[0141] S304. Determine a second control indicator for the second course based on the current time, the total number of employees, the employee entry time, and the employee learning time; wherein the second control indicator indicates whether the second course is a course provided to new employees.

[0142] For example, for any second course, the attribute information of the second course includes the total number of employees, employee entry time, and employee learning time. The total number of employees is the total number of employees who have browsed the second course within a preset time period; the employee entry time is the entry time of each employee who has browsed the second course within a preset time period; and the employee learning time is the time that the employees who have browsed the second course within the preset time period have browsed the second course.

[0143] The label information of the second course represents a new employee category label, and the new employee category label represents that the second course is a course provided to new employees, that is, the second course is a course that new employees need to learn.

[0144] The second control indicator of the second course is determined based on the current time, the total number of employees, the employee joining time, and the employee learning time. For example, for any second course, the joining time of each employee who has browsed the course can be determined based on the joining time of each employee who has browsed the course and the current time, and employees whose joining time is less than the preset time threshold are determined as new employees, and employees whose joining time is greater than or equal to the preset time threshold are determined as old employees. The preset time threshold is used to distinguish whether the employee who has browsed the second course is a new employee, and the preset time threshold can be, for example, one year.

[0145] Determine the ratio of the number of new employees to the total number of employees, which is the new employee ratio. Based on the learning time of employees, determine the total learning time of new employees to the total learning time of employees. Based on the total learning time of new employees to the total learning time of employees, determine the ratio of the total learning time of new employees to the total learning time of employees, that is, the proportion of the total learning time of new employees in the total learning time of employees, which is the new employee learning ratio.

[0146] A new employee ratio threshold and a new employee learning ratio threshold are preset. The new employee ratio threshold is used to determine whether the proportion of the number of new employees in the number of employees in the second course meets the requirement of adding a new employee category label to the second course; the new employee learning ratio threshold is used to determine whether the proportion of the total learning time of new employees in the total learning time of employees in the second course meets the requirement of adding a new employee category label to the second course. If the new employee ratio is greater than the preset new employee ratio threshold, and the new employee learning ratio is greater than the preset new employee learning ratio threshold, the second control index of the second course is determined to be 2; if the new employee ratio is less than or equal to the preset new employee ratio threshold, and the new employee learning ratio is greater than the new employee learning ratio threshold of the fool, the second control index of the second course is determined to be 1; if the new employee ratio is greater than the preset new employee ratio threshold, and the new employee learning ratio is less than or equal to the preset new employee learning ratio threshold, the second control index of the second course is determined to be 1; in other cases, the second control index of the second course is determined to be 0.

[0147] The beneficial effect of such a setting is that, based on the two indicators of the new employee ratio and the new employee learning ratio, it is determined whether the second course should be labeled with a new employee category, making the addition of the new employee category label more accurate. By using the new employee ratio and the new employee ratio threshold, courses where the number of new employees exceeds a certain proportion of the total number of employees can be fully screened out; by using the learning time of new employees and the total learning time of employees, courses where the learning time of new employees is longer can be fully screened out, fully eliminating the possibility that the number of new employees exceeds the ratio due to the new employee's mistake, but the learning time is short, and the course is not a new employee's learning, and the new employee category label is added, making the determined new employee label more accurate.

[0148] In this embodiment, the second control indicator of the second course is determined based on the current time, the total number of employees, the employee's entry time, and the employee's learning time, including: determining the time difference between the current time and the employee's entry time, if the time difference is less than a preset difference threshold, determining that the employee is a new employee; determining the second control indicator of the second course based on the total number of employees, the employee learning time of each employee, the number of new employees, and the employee learning time of each new employee.

[0149] Specifically, the difference between the current time and the employee's employment time is determined as the time difference. If the time difference is less than a preset difference threshold, the employee is determined to be a new employee. The preset difference threshold may be, for example, 1 year.

[0150] For any second course, the second control index of the second course is determined based on the total number of employees, the learning time of each employee, the number of new employees, and the learning time of each new employee. For example, for any second course, the second control index of the second course can be determined based on a preset second index determination formula, and the preset second index determination formula can be, for example:

[0151]

[0152] Among them, i is the number of the second course, the minimum value of i is 1, and the maximum value is I, that is, the second training set includes I second courses; j is the number of employees who have browsed the course numbered i, and the maximum value of j is J, that is, the total number of employees who have browsed the course numbered i is J.

[0153] Y i,j is the new employee indicator, that is, when Y i,j When Y is 1, the employee with number j who has browsed the course with number i is a new employee. i,j When CSJT is 0, the employee with ID j who has browsed the course with ID i is an old employee; i,j CSJT is the employee’s entry time.i,j Indicates the entry time of employee number j who has browsed course number i; TN is the current time; A2 is the preset difference threshold, that is, the preset time difference, for example, 1 year. This formula indicates that the new employee indicator is 1 when the time difference is less than the preset difference threshold, and 0 in other cases.

[0154] f2 is the second control indicator, which consists of two parts. The square brackets and their contents close to f2 are the first part of the second control indicator, and the square brackets and their contents far from f2 are the second part of the second control indicator. The first part is used to detect whether the proportion of new employees in the total number of employees in the second course reaches the preset proportion; the second part is used to detect whether the proportion of the total learning time of new employees in the second course to the total learning time of employees reaches the preset proportion. A3 is the preset threshold value for the proportion of new employees, which can be, for example, 0.5; A4 is the preset threshold value for the proportion of new employee learning time, which can be, for example, 0.5.

[0155] The first part of the second control indicator represents the sum of the new employee indicators of all new employees of the second course to obtain the number of new employees, divide the number of new employees by the total number of employees to obtain the proportion of the number of new employees in the total number of employees, multiply the proportion by the inverse of the new employee proportion threshold, and take the intersection of the obtained product and 1, which is the verification result of the first part of the second control indicator. The maximum value of the verification result of the first part of the second control indicator is 1. The closer the verification result of the first part of the second control indicator is to 1, the greater the proportion of new employees in the total number of employees.

[0156] The second part of the second control indicator represents the new employee indicator of all new employees of the second course multiplied by their respective employee learning time and then added together to obtain the total learning time of the new employees. The total learning time of the new employees is divided by the total learning time of the employees to obtain the proportion of the total learning time of the new employees in the total learning time of the employees. The proportion is multiplied by the inverse of the threshold of the proportion of new employee learning time. The intersection of the obtained product and 1 is the verification result of the second part of the second control indicator. The maximum value of the verification result of the second part of the second control indicator is 1. The closer the verification result of the second part of the second control indicator is to 1, the greater the proportion of the total learning time of the new employees in the total learning time of the employees.

[0157] The verification result of the first part of the second control index is added to the verification result of the second part of the second control index to obtain the second control index f2.

[0158] The beneficial effect of this setting is that the two indicators are used to respectively determine whether the proportion of new employees and the proportion of new employees' learning time have reached the preset thresholds, and comprehensively consider whether the second course should add new employee category labels, which can make the determination of new employee category labels more accurate and more credible.

[0159] S305: If the second control indicator is greater than a preset second threshold, determine that the new employee category label is label information of the second course.

[0160] Exemplarily, a second threshold is pre-set, and the second threshold is used to determine whether the second control indicator reaches a preset standard, that is, the second threshold is used to determine whether a new employee category label should be added to the second course. The preset second threshold can be any value greater than or equal to 1 and less than or equal to 2.

[0161] For any second course, if the second control indicator of the second course is greater than a preset second threshold, that is, the second course is a course that new employees need to learn, then it is determined that the new employee category label is label information of the second course.

[0162] The beneficial effect of such a setting is that whether a new employee label should be added is judged according to the preset second control index and the preset second threshold, so that the label of the new employee is determined accurately.

[0163] S306. Train the initial model according to the first target set and the second target set to obtain a trained course label determination model; wherein the course label determination model is used to determine label information of the course.

[0164] Exemplarily, this step may refer to the above-mentioned step S103 and will not be described in detail.

[0165] The training method, electronic device and storage medium of the course label determination model provided in the embodiment of the present application are obtained by obtaining a first set to be trained containing the attribute information of the first course and the label information of the first course, and obtaining a second set to be trained containing the attribute information of the second course. The attribute information of the first course in the first set to be trained is screened to obtain a first target set; the label information of the second course in the second set to be trained is determined to obtain a second target set. The initial model is trained using the first target set and the second target set to obtain a trained course label determination model. By using the trained course label model, the label information corresponding to the course can be quickly generated for courses without label information. By screening the first set to be trained to obtain the first target set, the association between the attribute information and the label information of the first course can be made more accurate; by determining the course label information corresponding to the second course in the second set to be trained to obtain the second target set, the course label information corresponding to the second course can be completed; by using the first target set and the second target set, the initial model is trained to improve the training accuracy of the model, which can make the recognition of the model more accurate and improve the recognition efficiency of the model.

[0166] Figure 4 A flowchart of a training method for a course label determination model provided in an embodiment of the present disclosure.

[0167] In this embodiment, the attribute information includes the total number of employees, employee positions, and employee learning time. The total number of employees represents the total number of employees who have browsed the second course, the employee learning time represents the time employees browsed the course, and the label information of the second course represents the management position category label; the management position category label represents that the second course is a course provided to employees in management positions; determining the label information of the second course in the second training set includes: if the employee is determined to be a management position based on the employee's employee position, then determining the sum of the learning time of employees in all management positions as time and; determining the third control indicator of the second course based on the number of employees in management positions, the total number of employees, the employee learning time of each employee, and time; wherein the third control indicator represents whether the second course is a course provided to employees in management positions; if the third control indicator is greater than a preset third threshold, then determining the management position category label as the label information of the second course.

[0168] like Figure 4 As shown, the method comprises the following steps:

[0169] S401, obtaining a first set to be trained and a second set to be trained; wherein the first set to be trained includes attribute information of a first course and label information of the first course, and the second set to be trained includes attribute information of a second course, and the label information represents the category of the course.

[0170] Exemplarily, this step may refer to the above-mentioned step S101 and will not be described in detail.

[0171] S402: Filter the attribute information of the first course in the first set to be trained to obtain a first target set.

[0172] Exemplarily, this step may refer to the above-mentioned step S102 and will not be described in detail.

[0173] S403. If the employee is determined to be in a management position according to his / her employee position, the sum of the study hours of all employees in the management position is determined as the time and.

[0174] For example, for any second course, the attribute information of the second course includes the total number of employees, employee positions, and employee learning time. Among them, the employee position is used to determine whether the employee is a management position employee, and the employee position symbol is CSM. i,j , which represents the employee position of employee number j who has browsed the course number i. If CSM i,j is 1, then the employee with number j who has browsed the course with number i is a management employee; if CSM i,j If it is 0, then the employee with number j who has browsed the course with number i is not a management employee.

[0175] For any second course, determine whether the employee position of the employees who have browsed the course is a management position, determine the number of management position employees in the second course, and determine the sum of the learning time of management position employees based on the employee position and the employee learning time, which is the time sum.

[0176] In one example, the employee position may also be any preset employee position, such as a financial position, etc. The employee position may be set as needed.

[0177] The beneficial effect of this arrangement is that by judging whether an employee is in a management position based on the employee position identification, the number of management position employees who have browsed any second course can be quickly determined, and the sum of the study time of management position employees can be determined based on the employee position identification and the employee's study time, thereby fully obtaining the information of management position employees and facilitating subsequent processing.

[0178] S404. Determine a third control indicator for the second course based on the number of employees in management positions, the total number of employees, the learning hours of each employee, and the time; wherein the third control indicator indicates whether the second course is a course provided to employees in management positions.

[0179] Exemplarily, the third control indicator represents whether the second course is a course provided to employees in management positions.

[0180] The third control indicator of the second course is determined based on the number of employees in management positions, the total number of employees, the learning time of each employee, and the time. For example, the third control indicator of the second course can be determined based on a preset third determination formula, and the preset third determination formula can be, for example:

[0181]

[0182] Among them, f3 is the third control indicator, and f3 consists of two parts. The square brackets and their contents close to f3 are the first part of the third control indicator, and the square brackets and their contents far from f3 are the second part of the third control indicator. Among them, the first part is used to detect whether the proportion of management employees in the total number of employees in the second course reaches the preset proportion; the second part is used to detect whether the proportion of the total learning time of management employees in the second course reaches the preset proportion. A5 is the preset threshold for the proportion of management employees, which can be, for example, 0.5; A6 is the preset threshold for the proportion of learning time of management employees, which can be, for example, 0.5.

[0183] The first part of the third control indicator represents the sum of the employee positions of all employees in management positions of the second course to obtain the number of employees in management positions, divide the number of employees in management positions by the total number of employees to obtain the proportion of the number of employees in management positions in the total number of employees, multiply the proportion by the inverse of the threshold of the proportion of employees in management positions, and take the intersection of the obtained product and 1, which is the verification result of the first part of the third control indicator. The maximum value of the verification result of the first part of the third control indicator is 1. The closer the verification result of the first part of the third control indicator is to 1, the greater the proportion of employees in management positions in the total number of employees.

[0184] The second part of the third control indicator represents the total learning time of the management employees, that is, time, by multiplying the employee positions of all management employees of the second course by their respective employee learning hours and then adding them together. The total learning time of the management employees is divided by the total learning time of the employees to obtain the proportion of the total learning time of the management employees in the total learning time of the employees. The proportion is multiplied by the inverse of the threshold of the proportion of learning time of the management employees. The intersection of the obtained product and 1 is the verification result of the second part of the third control indicator. The maximum value of the verification result of the second part of the third control indicator is 1. The closer the verification result of the second part of the third control indicator is to 1, the greater the proportion of the total learning time of the management employees in the total learning time of the employees.

[0185] The verification result of the first part of the third control index is added to the verification result of the second part of the third control index to obtain the third control index f3.

[0186] The beneficial effect of this setting is that the number of employees in management positions is determined according to the employee positions, and the third control indicator is determined according to the proportion of employees in management positions and the proportion of their learning time. This makes the determined third control indicator more accurate and can better reflect whether the second course should add a management position category label.

[0187] S405: If the third control indicator is greater than a preset third threshold, determine that the management position category label is label information of the second course.

[0188] Exemplarily, a third threshold is set in advance, and the third threshold is used to determine whether the third control indicator reaches a preset standard, that is, the third threshold is used to determine whether the third course should add a management position category label. The preset second threshold can be any value greater than or equal to 1 and less than or equal to 2.

[0189] For any second course, if the third control indicator of the second course is greater than the preset third threshold, that is, the second course is a course that management employees need to learn, then the management position category label is determined to be the label information of the second course.

[0190] The beneficial effect of such a setting is that whether a management position label should be added is judged according to the preset third control index and the preset third threshold value, so that the determined management position category label is more accurate.

[0191] S406. Train the initial model according to the first target set and the second target set to obtain a trained course label determination model; wherein the course label determination model is used to determine label information of the course.

[0192] Exemplarily, this step may refer to the above-mentioned step S103 and will not be described in detail.

[0193] The training method, electronic device and storage medium of the course label determination model provided in the embodiment of the present application are obtained by obtaining a first set to be trained containing the attribute information of the first course and the label information of the first course, and obtaining a second set to be trained containing the attribute information of the second course. The attribute information of the first course in the first set to be trained is screened to obtain a first target set; the label information of the second course in the second set to be trained is determined to obtain a second target set. The initial model is trained using the first target set and the second target set to obtain a trained course label determination model. By using the trained course label model, the label information corresponding to the course can be quickly generated for courses without label information. By screening the first set to be trained to obtain the first target set, the association between the attribute information and the label information of the first course can be made more accurate; by determining the course label information corresponding to the second course in the second set to be trained to obtain the second target set, the course label information corresponding to the second course can be completed; by using the first target set and the second target set, the initial model is trained to improve the training accuracy of the model, which can make the recognition of the model more accurate and improve the recognition efficiency of the model.

[0194] Figure 5 A flowchart of a training method for a course label determination model provided in an embodiment of the present disclosure.

[0195] In this embodiment, the attribute information of the first course in the first set to be trained is screened to obtain the first target set, including: determining the confidence between each attribute information and each label information based on the attribute information and label information of all the first courses in the first set to be trained; wherein the confidence represents the degree of association between the attribute information and the label information; if the confidence is less than a preset confidence threshold, the attribute information is deleted to obtain the first target set.

[0196] like Figure 5 As shown, the method comprises the following steps:

[0197] S501. Obtain a first set to be trained and a second set to be trained; wherein the first set to be trained includes attribute information of a first course and label information of the first course, and the second set to be trained includes attribute information of a second course, and the label information represents the category of the course.

[0198] Exemplarily, this step may refer to the above-mentioned step S101 and will not be described in detail.

[0199] S502. Determine the confidence between each attribute information and each label information based on the attribute information and label information of all the first courses in the first set to be trained; wherein the confidence represents the degree of association between the attribute information and the label information.

[0200] Exemplarily, the confidence level represents the degree of association between the attribute information and the label information, that is, the higher the confidence level, the higher the degree of association between the attribute information and the label information.

[0201] According to the attribute information and label information of all the first courses in the first set to be trained, the confidence between each attribute information and each label information is determined. For example, a confidence determination model may be preset, and the preset confidence determination model may be a large model. The preset confidence determination model may determine the confidence between any attribute information and any label information in the first set to be trained according to the attribute information and label information of all the first courses in the first set to be trained. According to the attribute information and label information of all the first courses in the first set to be trained, the confidence between each attribute information and each label information is determined based on the preset confidence determination model.

[0202] The beneficial effect of this setting is that since there are multiple first courses in the first training set, as well as attribute information and label information of the first courses, the attribute information and label information may be repeated. Therefore, by calculating the confidence between the attribute information and label information of all the first courses in the first training set, the confidence relationship between the attribute information and the label information can be more comprehensively determined, and the degree of correlation between the attribute information and the label information can be better determined.

[0203] In this embodiment, the confidence between each attribute information and each label information is determined based on the attribute information and label information of all the first courses in the first set to be trained, including: arranging and combining the attribute information and label information of all the first courses in the first set to be trained to obtain at least one information group; wherein the information group includes one attribute information and one label information; determining each label information of the first set to be trained as a target label, and determining each attribute information of the first set to be trained as a target attribute; determining the number of information groups containing target labels and target attributes as a first number, and the number of target labels in the first set to be trained as a second number; determining the confidence between the target attribute and the target label based on the first number and the second number.

[0204] Specifically, the attribute information and label information of all the first courses in the first training set are arranged and combined to obtain at least one information group, that is, for any attribute information in the first course, the attribute information is combined with all the label information to obtain the result of the arrangement and combination of all the attribute information and the label information, and any of the arrangement results is determined as an information group, and the number of all information groups is determined as the number of information groups. Among them, the information group contains one attribute information and one label information, and the information group can be, for example, {new employee, workplace etiquette}, that is, the attribute information in the information group is new employee, and the label information in the information group is workplace etiquette.

[0205] Each piece of label information in the first set to be trained is determined as a target label, and each piece of attribute information in the first set to be trained is determined as a target attribute. For the current target label and target attribute, the number of information groups containing the target label and the target attribute is determined as a first number; the number of target labels in the first set to be trained is determined as a second number.

[0206] The support is determined according to the first number and the number of information groups, wherein the support represents the percentage of information groups containing the target label and the target attribute in all information groups, that is, the probability of the target label and the target attribute appearing at the same time.

[0207] According to the second number and the number of all label information in the first to-be-trained set, the percentage of the target label in all label information, that is, the probability of the target label appearing is determined.

[0208] The confidence level is determined based on the probability and support of the target label. That is, the confidence level is equal to the probability of the target label and the target attribute appearing at the same time in the probability of the target label appearing.

[0209] Compute the confidence between each target attribute and the target label.

[0210] The beneficial effect of such a setting is that, based on all the attribute information and label information of the first set to be trained, an information group of the arrangement and combination of the attribute information and the label information is determined, and by calculating the probability of the information group in which the target label and the target attribute appear at the same time in the information group, the support of the attribute information and the label information is accurately determined, and the confidence is calculated, so that the degree of correlation between the attribute information and the label information can be quantified, and a numerical value representing the degree of correlation can be efficiently obtained, which helps to quickly screen the first set to be trained to remove attribute information with a low degree of correlation and improve the efficiency of training.

[0211] S503: If the confidence level is less than a preset confidence level threshold, the attribute information is deleted to obtain a first target set.

[0212] Exemplarily, a confidence threshold is preset. For example, the confidence threshold formula may be set as:

[0213]

[0214] Among them, z represents the confidence threshold; H is the amount of attribute information in the first training set; α is the adjustment coefficient, which can be set to a value less than or equal to 1. For example, α can be 0.2, that is, the confidence level needs to reach one-fifth of the average probability.

[0215] For the confidence of any attribute information, if the confidence is less than the preset confidence threshold, it is judged that there is a weak correlation between the attribute information and the course label, and the attribute information is deleted from the first to-be-trained set, and the training set is updated, and the first to-be-trained set with all weakly correlated attribute information deleted is determined as the first target set.

[0216] The beneficial effect of this setting is that, since there is label information in the first training set, but there are some attributes in the attribute information that have a low correlation with the course label information, in order to improve the calculation efficiency, the attribute information with a low correlation is deleted according to the confidence level to improve the training efficiency of the course label determination model.

[0217] S504: Determine label information of the second course in the second set to be trained to obtain a second target set.

[0218] Exemplarily, this step may refer to the above-mentioned step S102 and will not be described in detail.

[0219] S505. Train the initial model according to the first target set and the second target set to obtain a trained course label determination model; wherein the course label determination model is used to determine label information of the course.

[0220] Exemplarily, this step may refer to the above-mentioned step S103 and will not be described in detail.

[0221] The training method, electronic device and storage medium of the course label determination model provided in the embodiment of the present application are obtained by obtaining a first set to be trained containing the attribute information of the first course and the label information of the first course, and obtaining a second set to be trained containing the attribute information of the second course. The attribute information of the first course in the first set to be trained is screened to obtain a first target set; the label information of the second course in the second set to be trained is determined to obtain a second target set. The initial model is trained using the first target set and the second target set to obtain a trained course label determination model. By using the trained course label model, the label information corresponding to the course can be quickly generated for courses without label information. By screening the first set to be trained to obtain the first target set, the association between the attribute information and the label information of the first course can be made more accurate; by determining the course label information corresponding to the second course in the second set to be trained to obtain the second target set, the course label information corresponding to the second course can be completed; by using the first target set and the second target set, the initial model is trained to improve the training accuracy of the model, which can make the recognition of the model more accurate and improve the recognition efficiency of the model.

[0222] Figure 6 A flowchart of a training method for a course label determination model provided in an embodiment of the present disclosure.

[0223] In this embodiment, obtaining the second training set includes: obtaining courses that only contain attribute information as courses to be screened out; determining the number of views, the number of employees who view the courses, the course duration, and the employee learning duration of the courses to be screened out within a preset time period; determining the fourth control index of the courses to be screened out according to the number of views, the number of employees who view the courses, the course duration, and the employee learning duration of the courses to be screened out; wherein the fourth control index represents the importance of the courses to be screened out; if the fourth control index is greater than a preset fourth threshold value, the courses to be screened out are determined as the second courses.

[0224] like Figure 6 As shown, the method comprises the following steps:

[0225] S601: Obtain a first set to be trained.

[0226] Exemplarily, this step may refer to the above-mentioned step S101 and will not be described in detail.

[0227] S602, obtaining courses that only contain attribute information, which are courses to be screened out.

[0228] Exemplarily, courses containing only attribute information are obtained from the enterprise learning platform, which are the courses to be screened out, and attribute information of the courses to be screened out is obtained. The courses to be screened out are all the courses containing only attribute information in the enterprise platform, that is, the courses to be screened out do not have course label information.

[0229] The beneficial effect of this setting is that, for courses without label information, all courses without label information are obtained as courses to be screened out, so that the courses to be screened out can be screened out later, the data quality of the courses to be screened out can be improved, and the courses to be screened out can be effectively used in the training course label determination model.

[0230] S603: Determine the number of views of the to-be-screened courses, the number of employees who have viewed the courses, the course duration, and the employee study duration within a preset time period.

[0231] For example, the number of views of the course to be screened out, the number of employees who have viewed the course, the course duration, and the employee learning duration within a preset time period are determined. The number of views of the course to be screened out is the number of times the course to be screened out has been viewed within the preset time period.

[0232] The beneficial effect of such a setting is that it can determine the effective information of the courses to be screened out, facilitate the screening out of some courses whose course label information is difficult to determine, retain the courses whose course labels are easier to determine, and improve the quality of the training set.

[0233] S604, determining a fourth control indicator of the course to be screened out according to the number of views of the course to be screened out, the number of employees who have viewed the course, the course duration, and the employee learning duration; wherein the fourth control indicator represents the importance of the course to be screened out.

[0234] Exemplarily, the fourth control indicator represents the importance of the courses to be screened out, that is, the larger the fourth control indicator, the more important the course to be screened out is; the larger the fourth control indicator, the higher the overall learning amount and learning progress of the course to be screened out, and the higher the quality of the attribute information of the course to be screened out as a training set.

[0235] The fourth control index of the courses to be screened out is determined according to the number of views of the courses to be screened out, the number of employees who have viewed the courses, the course duration, and the learning duration of the employees. For example, the fourth control index may be determined based on a preset fourth determination formula, and the preset fourth determination formula may be, for example:

[0236]

[0237] Wherein, f4 is the fourth control index; T is the preset time period; CST i,T The number of views of the course numbered i within the preset time period. Each time an employee clicks on the course, the number of views increases by one. CSD i,T is the total learning time of the course numbered i within the preset time period, that is, the browsing time of the course by all employees who browsed the course within the preset time period; CT iis the course duration of the course numbered i, that is, the teaching time length of the course itself. A7 is the preset learning times threshold, and A8 is the preset learning progress threshold.

[0238] f4 consists of two parts. The square brackets and their contents near f4 are the first part of the fourth control indicator, and the square brackets and their contents far from f4 are the second part of the fourth control indicator. The first part of the fourth control indicator is used to detect whether the number of learning times of the course to be screened out has reached the preset learning times threshold; the second part of the fourth control indicator is used to detect whether the learning progress of the course to be screened out has reached the preset learning progress threshold.

[0239] The first part of the fourth control indicator represents the ratio of the number of views of the course numbered i to the average number of views of the courses to be screened out within a preset time period, multiplied by the inverse of the learning times threshold, and the intersection of the obtained product and 1 is the verification result of the first part of the fourth control indicator. The maximum value of the verification result of the first part of the fourth indicator is 1. The closer the verification result of the first part of the fourth control indicator is to 1, the more the number of learning times of the course exceeds the average number of learning times of all the courses to be screened out.

[0240] The second part of the fourth control indicator represents the ratio of the total learning time of the course numbered i to the product of the number of course views and the course time of the course numbered i within a preset time period, that is, the learning progress of the course to be screened out. The ratio is multiplied by the inverse of the preset learning progress threshold, and the product is intersected with 1. The result obtained is the verification result of the second part of the fourth control indicator. The maximum value of the verification result of the second part of the fourth indicator is 1. The closer the verification result of the second part of the fourth control indicator is to 1, the greater the learning progress of all employees who have browsed the course.

[0241] The fourth control index is obtained by summing the first part of the fourth control index and the second part of the fourth control index.

[0242] The beneficial effect of this setting is that the fourth control indicator is determined based on the number of learning times and the learning progress, so that the fourth control indicator of each course to be screened out can more effectively reflect the overall learning situation of the course, thereby facilitating the subsequent deletion of courses to be screened out with poor overall learning conditions.

[0243] S605: If the fourth control indicator is greater than a preset fourth threshold, the course to be screened out is determined to be the second course.

[0244] Exemplarily, a fourth threshold is preset, and the fourth threshold is used to determine whether the fourth control indicator meets the requirements, that is, whether the learning situation of the to-be-screened course meets the requirements of being a training set. The preset fourth threshold can be any value greater than 1 and less than or equal to 2.

[0245] If the fourth control index is greater than the preset fourth threshold, the course to be screened out is determined as the second course. All courses to be screened out are determined to obtain the second course, and the attribute information of the second course is determined as the second training set.

[0246] The beneficial effect of such a setting is that, according to the fourth control indicator, the course information with poor learning amount or low learning progress will be deleted, thereby improving the quality of the training set data and improving the efficiency and accuracy of subsequent training of the course label determination model.

[0247] S606. Filter the attribute information of the first course in the first set to be trained to obtain a first target set, and determine the label information of the second course in the second set to be trained to obtain a second target set; wherein the first target set represents the first set to be trained after the attribute information is filtered, and the second target set includes the attribute information of the second course and the label information of the second course.

[0248] Exemplarily, this step may refer to the above-mentioned step S102 and will not be described in detail.

[0249] S607. Train the initial model according to the first target set and the second target set to obtain a trained course label determination model; wherein the course label determination model is used to determine label information of the course.

[0250] Exemplarily, this step may refer to the above-mentioned step S103 and will not be described in detail.

[0251] The training method, electronic device and storage medium of the course label determination model provided in the embodiment of the present application are obtained by obtaining a first set to be trained containing the attribute information of the first course and the label information of the first course, and obtaining a second set to be trained containing the attribute information of the second course. The attribute information of the first course in the first set to be trained is screened to obtain a first target set; the label information of the second course in the second set to be trained is determined to obtain a second target set. The initial model is trained using the first target set and the second target set to obtain a trained course label determination model. By using the trained course label model, the label information corresponding to the course can be quickly generated for courses without label information. By screening the first set to be trained to obtain the first target set, the association between the attribute information and the label information of the first course can be made more accurate; by determining the course label information corresponding to the second course in the second set to be trained to obtain the second target set, the course label information corresponding to the second course can be completed; by using the first target set and the second target set, the initial model is trained to improve the training accuracy of the model, which can make the recognition of the model more accurate and improve the recognition efficiency of the model.

[0252] Figure 7A structural block diagram of a training device for a course label determination model provided in an embodiment of the present disclosure.

[0253] For the sake of convenience, only the parts related to the embodiments of the present disclosure are shown. Figure 7 The training device 700 for determining the course label model includes: an acquisition unit 701, a processing unit 702 and a training unit 703.

[0254] The acquisition unit 701 is used to acquire a first set to be trained and a second set to be trained; wherein the first set to be trained includes attribute information of the first course and label information of the first course, and the second set to be trained includes attribute information of the second course, and the label information represents the category of the course;

[0255] The processing unit 702 is used to filter the attribute information of the first course in the first set to be trained to obtain a first target set, and determine the label information of the second course in the second set to be trained to obtain a second target set; wherein the first target set represents the first set to be trained after the attribute information is filtered, and the second target set includes the attribute information of the second course and the label information of the second course;

[0256] The training unit 703 is used to train the initial model according to the first target set and the second target set to obtain a trained course label determination model; wherein the course label determination model is used to determine the label information of the course.

[0257] Figure 8 A structural block diagram of a training device for a course label determination model provided in an embodiment of the present disclosure.

[0258] exist Figure 7 Based on the embodiment shown, Figure 8 As shown, the processing unit 702 includes a screening module 7021 , a variance module 7022 and a first determination module 7023 .

[0259] A screening module 7021 is used to screen the attribute information of the first course in the first to-be-trained set to obtain a first target set;

[0260] The variance module 7022 is used to determine the variance of the major category according to the number of majors corresponding to each employee's major; wherein the variance of the major category represents the degree of dispersion of the employee's major in the second course;

[0261] The first determination module 7023 is used to determine the label information of the second course according to the major of each employee, the number of majors corresponding to each employee's major, and the total number of employees if the variance of the major category is greater than a preset variance threshold.

[0262] In one example, the first determining module 7023 includes:

[0263] The sorting submodule is used to sort the majors of each employee according to the number of professionals corresponding to each employee's major and obtain the sorting results;

[0264] The average value submodule is used to determine the average value according to the total number of employees and the number of employee specialties; wherein the average value represents the average value of the number of employees under each employee specialty;

[0265] The first indicator submodule is used to determine the first control indicator corresponding to each employee's major according to the ranking result, the total number of employees, the average value, and the number of professionals corresponding to each employee's major; wherein the first control indicator represents whether the professional category label of the second course is an employee's major;

[0266] The label submodule is used to determine the employee's major as the professional category label of the second course if the first control indicator is a preset first threshold.

[0267] In one example, the indicator submodule is specifically used for:

[0268] If it is determined according to the sorting result that the employee's major is located in a preset sorting position, the first control indicator corresponding to each employee's major is determined according to the total number of employees, the average value, and the number of professionals corresponding to each employee's major.

[0269] In one example, the processing unit 702 includes:

[0270] Time module, used to obtain the current time;

[0271] A second determination module is used to determine a second control indicator of the second course according to the current time, the total number of employees, the employee's entry time, and the employee's learning time; wherein the second control indicator indicates whether the second course is a course provided to new employees;

[0272] The label module is used to determine that the new employee category label is the label information of the second course if the second control indicator is greater than a preset second threshold.

[0273] In one example, the second determination module includes:

[0274] The difference submodule is used to determine the time difference between the current time and the employee's employee entry time. If the time difference is less than a preset difference threshold, the employee is determined to be a new employee.

[0275] The second indicator submodule is used to determine the second control indicator of the second course according to the total number of employees, the employee learning time of each employee, the number of new employees, and the employee learning time of each new employee.

[0276] In one example, the processing unit 702 includes:

[0277] The sum module is used to determine the sum of the learning hours of all employees in management positions, which is the time and if the employee is determined to be a management position based on the employee's position;

[0278] An indicator module, used to determine a third control indicator of the second course according to the number of employees in management positions, the total number of employees, the learning time of each employee, and the time; wherein the third control indicator indicates whether the second course is a course provided to employees in management positions;

[0279] The third determination module is used to determine that the management position category label is the label information of the second course if the third control indicator is greater than a preset third threshold.

[0280] In one example, the processing unit 702 includes:

[0281] A confidence module, used to determine the confidence between each attribute information and each label information according to the attribute information and label information of all the first courses in the first to-be-trained set; wherein the confidence represents the degree of association between the attribute information and the label information;

[0282] The deletion module is used to delete the attribute information if the confidence is less than a preset confidence threshold, so as to obtain a first target set.

[0283] In one example, the confidence module includes:

[0284] An arrangement submodule, used for arranging and combining the attribute information and label information of all the first courses in the first set to be trained to obtain at least one information group; wherein the information group includes one attribute information and one label information;

[0285] A target determination submodule, used to determine each piece of label information of the first to-be-trained set as a target label, and to determine each piece of attribute information of the first to-be-trained set as a target attribute;

[0286] A quantity determination submodule, used to determine the number of information groups containing target labels and target attributes, which is a first number, and the number of target labels in the first set to be trained, which is a second number;

[0287] The confidence submodule is used to determine the confidence between the target attribute and the target label according to the first quantity and the second quantity.

[0288] In one example, the acquisition unit 701 includes:

[0289] The acquisition module is used to obtain courses that only contain attribute information, which are the courses to be screened out;

[0290] An information module is used to determine the number of views of the courses to be screened out, the number of employees who have viewed the courses, the course duration, and the employee learning time within a preset time period;

[0291] A screening indicator module is used to determine the fourth control indicator of the course to be screened out according to the number of views of the course to be screened out, the number of employees who have viewed the course, the course duration, and the employee learning duration; wherein the fourth control indicator represents the importance of the course to be screened out;

[0292] The fourth determination module is used to determine the course to be screened out as the second course if the fourth control indicator is greater than a preset fourth threshold.

[0293] In one example, it also includes:

[0294] Course unit, used to obtain courses to be classified;

[0295] The determination unit is used to input the courses to be classified into the course label determination model to obtain the output label information of the courses to be classified.

[0296] Fig. 9 A structural block diagram of an electronic device provided in an embodiment of the present disclosure, the electronic device may be a terminal device or a server, such as Fig. 9 As shown, the electronic device 900 includes: at least one processor 902; and a memory 901 communicatively connected to the at least one processor 902; wherein the memory stores instructions executable by the at least one processor 902, and the instructions are executed by the at least one processor 902 so that the at least one processor 902 can execute the training method of the course label determination model disclosed in the present invention.

[0297] The electronic device 900 further includes a receiver 903 and a transmitter 904. The receiver 903 is used to receive instructions and data sent by other devices, and the transmitter 904 is used to send instructions and data to external devices.

[0298] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0299] According to an embodiment of the present disclosure, the present disclosure also provides a computer program product, which includes: a computer program, the computer program is stored in a readable storage medium, at least one processor of an electronic device can read the computer program from the readable storage medium, and at least one processor executes the computer program so that the electronic device executes the solution provided by any of the above embodiments.

[0300] Fig.10It is a block diagram of an electronic device according to an exemplary embodiment, which device may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

[0301] Device 1000 may include one or more of the following components: a processing component 1002 , a memory 1004 , a power component 1006 , a multimedia component 1008 , an audio component 1011 , an input / output (I / O) interface 1012 , a sensor component 1014 , and a communication component 1016 .

[0302] The processing component 1002 generally controls the overall operation of the device 1000, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 1002 may include one or more processors 1020 to execute instructions to complete all or part of the steps of the above-mentioned method. In addition, the processing component 1002 may include one or more modules to facilitate the interaction between the processing component 1002 and other components. For example, the processing component 1002 may include a multimedia module to facilitate the interaction between the multimedia component 1008 and the processing component 1002.

[0303] The memory 1004 is configured to store various types of data to support the operation of the device 1000. Examples of such data include instructions for any application or method operating on the device 1000, contact data, phone book data, messages, pictures, videos, etc. The memory 1004 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0304] The power supply component 1006 provides power to the various components of the device 1000. The power supply component 1006 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device 1000.

[0305] The multimedia component 1008 includes a screen that provides an output interface between the device 1000 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 1008 includes a front camera and / or a rear camera. When the device 1000 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera may receive external multimedia data. Each front camera and the rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

[0306] The audio component 1010 is configured to output and / or input audio signals. For example, the audio component 1010 includes a microphone (MIC), and when the device 1000 is in an operating mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 1004 or sent via the communication component 1016. In some embodiments, the audio component 1010 also includes a speaker for outputting audio signals.

[0307] I / O interface 1012 provides an interface between processing component 1002 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: home button, volume button, start button, and lock button.

[0308] The sensor assembly 1014 includes one or more sensors for providing various aspects of status assessment for the device 1000. For example, the sensor assembly 1014 can detect the open / closed state of the device 1000, the relative positioning of components, such as the display and keypad of the device 1000, and the sensor assembly 1014 can also detect the position change of the device 1000 or a component of the device 1000, the presence or absence of user contact with the device 1000, the orientation or acceleration / deceleration of the device 1000, and the temperature change of the device 1000. The sensor assembly 1014 may include a proximity sensor configured to detect the presence of a nearby object without any physical contact. The sensor assembly 1014 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 1014 may also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0309] The communication component 1016 is configured to facilitate wired or wireless communication between the device 1000 and other devices. The device 1000 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 1016 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 1016 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0310] In an exemplary embodiment, the device 1000 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above methods.

[0311] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 1004 including instructions, and the instructions can be executed by the processor 1020 of the device 1000 to perform the above method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0312] A non-temporary computer-readable storage medium, when the instructions in the storage medium are executed by the processor of a terminal device, enables the terminal device to execute the training method of the above-mentioned course label determination model.

[0313] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.

[0314] It should be further noted that, although the various steps in the flowchart are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowchart may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0315] It should be understood that the above-mentioned device embodiments are only illustrative, and the device of the present application can also be implemented in other ways. For example, the division of units / modules in the above-mentioned embodiments is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units, modules or components can be combined, or can be integrated into another system, or some features can be ignored or not executed.

[0316] In addition, unless otherwise specified, each functional unit / module in each embodiment of the present application may be integrated into one unit / module, each unit / module may exist physically separately, or two or more units / modules may be integrated together. The above-mentioned integrated unit / module may be implemented in the form of hardware or in the form of a software program module.

[0317] If the integrated unit / module is implemented in the form of hardware, the hardware may be a digital circuit, an analog circuit, etc. The physical implementation of the hardware structure includes but is not limited to transistors, memristors, etc. If not specifically stated, the processor may be any appropriate hardware processor, such as a CPU, a GPU, an FPGA, a DSP, an ASIC, etc. If not specifically stated, the storage unit may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc.

[0318] If the integrated unit / module is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0319] In the above embodiments, the description of each embodiment has its own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant description of other embodiments. The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0320] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0321] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A training method for a course label determination model, characterized in that: include: Obtain a first set to be trained and a second set to be trained; wherein the first set to be trained includes attribute information of the first course and label information of the first course, and the second set to be trained includes attribute information of the second course, and the label information represents the category of the course; Filtering the attribute information of the first course in the first set to be trained to obtain a first target set, and determining the label information of the second course in the second set to be trained to obtain a second target set; wherein the first target set represents the first set to be trained after the attribute information is filtered, and the second target set includes the attribute information of the second course and the label information of the second course; The initial model is trained according to the first target set and the second target set to obtain a trained course label determination model; wherein the course label determination model is used to determine the label information of the course.

2. The method according to claim 1, characterized in that The attribute information includes the total number of employees, employee majors, and the number of majors, wherein the total number of employees represents the total number of employees who have browsed the second course, the number of majors represents the number of employees who belong to the employee major among the employees who have browsed the second course, and the label information of the second course represents a professional category label, and the professional category label represents the professional category to which the second course belongs; Determining label information of the second course in the second to-be-trained set includes: Determine the professional category variance according to the number of professionals corresponding to each employee's major; wherein the professional category variance represents the degree of dispersion of the employee's major in the second course; If the professional category variance is greater than a preset variance threshold, the label information of the second course is determined according to the major of each employee, the number of majors corresponding to each employee's major, and the total number of employees.

3. The method according to claim 2, characterized in that According to the major of each employee, the number of professionals corresponding to each employee's major, and the total number of employees, the label information of the second course is determined, including: According to the number of professionals corresponding to each employee's major, sort the majors of each employee to obtain the sorting results; Determine an average value according to the total number of employees and the number of employee specialties; wherein the average value represents the average value of the number of employees under each employee specialty; Determine a first control indicator corresponding to each employee's major according to the ranking result, the total number of employees, the average value, and the number of professionals corresponding to each employee's major; wherein the first control indicator represents whether the professional category label of the second course is an employee's major; If the first control indicator is a preset first threshold, the employee's major is determined as the professional category label of the second course.

4. The method according to claim 3, characterized in that Determine the first control indicator corresponding to each employee's major according to the ranking result, the total number of employees, the average value, and the number of professionals corresponding to each employee's major, including: If it is determined according to the sorting result that the employee's major is located in a preset sorting position, the first control indicator corresponding to each employee's major is determined according to the total number of employees, the average value, and the number of professionals corresponding to each employee's major.

5. The method according to claim 1, characterized in that: The attribute information includes the total number of employees, the time of joining the employees, and the length of time the employees have studied. The total number of employees represents the total number of employees who have browsed the second course. The length of time the employees have studied represents the length of time the employees have browsed the course. The label information of the second course represents a new employee category label. The new employee category label represents that the second course is a course provided to new employees. Determining label information of the second course in the second to-be-trained set includes: Get the current time; Determine a second control indicator for the second course according to the current time, the total number of employees, the employee entry time, and the employee learning time; wherein the second control indicator indicates whether the second course is a course provided to new employees; If the second control indicator is greater than a preset second threshold, it is determined that the new employee category label is the label information of the second course.

6. The method according to claim 5, characterized in that Determine the second control indicator of the second course according to the current time, the total number of employees, the employee entry time, and the employee learning time, including: Determine the time difference between the current time and the employee's employment time, and if the time difference is less than a preset difference threshold, determine that the employee is a new employee; A second control indicator for the second course is determined based on the total number of employees, the employee learning time of each employee, the number of new employees, and the employee learning time of each new employee.

7. The method according to claim 1, characterized in that The attribute information includes the total number of employees, employee positions, and employee learning time. The total number of employees represents the total number of employees who have browsed the second course. The employee learning time represents the time employees have browsed the course. The label information of the second course represents the management position category label. The management position category label indicates that the second course is a course provided to employees in management positions; Determining label information of the second course in the second to-be-trained set includes: If the employee is determined to be in a management position based on his / her position, the sum of the learning hours of all employees in management positions is determined as the time; Determine a third control indicator of the second course according to the number of employees in management positions, the total number of employees, the learning time of each employee, and the time; wherein the third control indicator indicates whether the second course is a course provided to employees in management positions; If the third control indicator is greater than a preset third threshold, it is determined that the management position category label is the label information of the second course.

8. The method according to claim 1, characterized in that The attribute information of the first course in the first set to be trained is screened to obtain a first target set, including: Determine the confidence between each attribute information and each label information according to the attribute information and label information of all the first courses in the first set to be trained; wherein the confidence represents the degree of association between the attribute information and the label information; If the confidence is less than a preset confidence threshold, the attribute information is deleted to obtain the first target set.

9. The method according to claim 8, characterized in that Determining the confidence between each attribute information and each label information according to the attribute information and label information of all the first courses in the first to-be-trained set includes: Arrange and combine the attribute information and label information of all the first courses in the first set to be trained to obtain at least one information group; wherein the information group includes one attribute information and one label information; Determine each piece of label information of the first to-be-trained set as a target label, and determine each piece of attribute information of the first to-be-trained set as a target attribute; Determine the number of information groups containing the target label and the target attribute as a first number, and the number of the target label in the first to-be-trained set as a second number; A confidence level between the target attribute and the target label is determined according to the first number and the second number.

10. The method according to claim 1, characterized in that Get the second training set, including: Get courses that only contain attribute information, which are the courses to be screened out; Determine the number of views of the to-be-screened course, the number of employees who have viewed the course, the course duration, and the employee study duration within a preset time period; Determine the fourth control index of the course to be screened out according to the number of views of the course to be screened out, the number of employees who have viewed the course, the course duration, and the employee learning duration; wherein the fourth control index represents the importance of the course to be screened out; If the fourth control indicator is greater than a preset fourth threshold, the course to be screened out is determined to be the second course.

11. The method according to claim 1, characterized in that: Also includes: Get the attribute information of the course to be classified; The attribute information of the course to be classified is input into the course label determination model to obtain the output label information of the course to be classified.

12. A training device for a course label determination model, characterized in that: include: An acquisition unit, configured to acquire a first set to be trained and a second set to be trained; wherein the first set to be trained includes attribute information of a first course and label information of the first course, and the second set to be trained includes attribute information of a second course, and the label information represents a category of the course; A processing unit, configured to filter the attribute information of the first course in the first set to be trained to obtain a first target set, and determine the label information of the second course in the second set to be trained to obtain a second target set; wherein the first target set represents the first set to be trained after the attribute information is filtered, and the second target set includes the attribute information of the second course and the label information of the second course; A training unit is used to train the initial model according to the first target set and the second target set to obtain a trained course label determination model; wherein the course label determination model is used to determine the label information of the course.

13. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 11.

14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 11 when executed by a processor.

15. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 11 when being executed by a processor.

Images