Model training method, low-efficiency course identification method, device, equipment and medium

Abstract

The invention provides a model training method, a low-efficiency course identification method and device, equipment and a medium. According to the method, a plurality of sample courses are obtained, for each sample course, the course type of the sample course is determined according to the online time of the sample course, and then the sample label of the sample course is determined according to the course type, so that an evaluation result of whether the sample course is a low-efficiency course is obtained. And finally, according to the course information of the plurality of sample courses, the student information, the historical learning record information and the sample labels, performing model training on a pre-acquired to-be-trained model to obtain a low-efficiency course identification model, thereby improving the accuracy of low-efficiency course identification through the low-efficiency course identification model.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a model training method, an inefficient course identification method, a device, equipment, and a medium. Background Technology

[0002] With the popularization and in-depth application of the Internet, corporate online learning platforms have become an important channel for internal education and knowledge sharing.

[0003] However, with the continuous increase in learning resources, corporate online learning platforms are offering more and more courses, some of which are inefficient. Failure to clean them up in a timely manner will make it difficult for learners to focus on more efficient courses when making course selections.

[0004] Existing technologies suffer from inaccurate identification of inefficient courses. Summary of the Invention

[0005] This application provides a model training method, an inefficient course identification method, a device, equipment, and a medium to solve the problem of inaccurate identification of inefficient courses in the prior art.

[0006] Firstly, this application provides a model training method, including:

[0007] Obtain multiple sample courses;

[0008] For each sample course, the course type is determined based on its launch time. The course types include new courses and old courses.

[0009] Based on the course type, determine the sample label for the sample course. The sample label is used to indicate whether the sample course is an inefficient course.

[0010] Based on course information, student information, historical learning records, and sample labels from multiple sample courses, the pre-acquired training model is trained to obtain an inefficient course identification model.

[0011] Furthermore, for each sample course, the course type is determined based on its launch time, including:

[0012] For each sample course, determine the launch time and the preset time point;

[0013] If the sample course is launched later than the preset time, the course type of the sample course is a new course.

[0014] If the sample course is launched earlier than the preset time, the course type of the sample course is the old course type.

[0015] Furthermore, based on the course type, sample labels for the sample courses are determined, including:

[0016] Determine the initial sample labels for the sample courses based on the course type;

[0017] Based on the unit duration, completion rate, and number of citations of the sample courses within the preset learning period, the initial sample labels are corrected to obtain the sample labels of the sample courses. The unit duration is obtained by the ratio of the total learning time of the sample courses to the course duration.

[0018] Furthermore, when the sample course type is a new course,

[0019] Based on the course type, determine the initial sample labels for the sample courses, including:

[0020] Get the first number of times the sample course has been viewed since its launch;

[0021] If the first learning attempt is less than the preset first learning attempt threshold, the initial sample label of the sample course is an inefficient course.

[0022] If the first learning attempt exceeds the preset first learning attempt threshold, the initial sample label for the sample course is "non-inefficient course".

[0023] Furthermore, when the sample course type is an old course type,

[0024] Based on the course type, determine the initial sample labels for the sample courses, including:

[0025] Obtain the second learning session of the sample course within the first preset time period;

[0026] If the second learning count is less than the preset second learning count threshold, the initial sample label of the sample course is an inefficient course.

[0027] If the second learning count is greater than the preset second learning count threshold, the initial sample label for the sample course is "non-inefficient course".

[0028] Furthermore, when the initial sample label is "inefficient course,"

[0029] Based on the unit duration, completion rate, and citation count of the sample courses within the preset learning period, the initial sample labels are revised to obtain the sample labels for the sample courses, including:

[0030] Obtain the first sample set of courses whose initial sample label is "inefficient" from multiple sample courses;

[0031] Based on the unit duration, completion rate, and number of citations of each sample course in the first sample set within the preset learning period, determine the first average unit duration, first average completion rate, and first average number of citations for all sample courses in the first sample set.

[0032] The first indicator is determined based on the comparison results of the unit time of the sample course within the preset learning period with the first preset unit time threshold and the first average unit time, the comparison results of the completion rate of the sample course within the preset learning period with the first preset completion rate threshold and the first average completion rate, and the comparison results of the number of times the sample course is cited within the preset learning period with the first preset number of citations threshold and the first average number of citations.

[0033] If the first indicator is greater than the preset first indicator threshold, then the non-inefficient course will be used as the sample label for the sample course.

[0034] If the first indicator is less than or equal to the preset first indicator threshold, the initial sample label will be used as the sample label for the sample course.

[0035] Furthermore, based on the comparison results of the sample course's unit duration within the preset learning period with the first preset unit duration threshold and the first average unit duration, the comparison results of the sample course's completion rate within the preset learning period with the first preset completion rate threshold and the first average completion rate, and the comparison results of the sample course's citation count within the preset learning period with the first preset citation count threshold and the first average citation count, the first indicator is determined, satisfying the following:

[0036]

[0037] Among them, f1 j The primary metric is T, where T is the preset learning period, and C1 is the primary metric. j For the j-th sample course in the first sample set, C1 j The learning time within T, C1 j Course duration, C1 j Within a unit duration of T, F1 is the first preset unit duration threshold, and A1 is the first average unit duration. C1 j The completion rate within time T, where F2 is the first preset completion rate threshold and A2 is the first average completion rate. C1 j The number of citations within T, F3 is the first preset citation threshold, A3 is the first average citation count, and α1, β1 and γ1 are adjustment coefficients.

[0038] Furthermore, when the initial sample label is "non-inefficient course",

[0039] Based on the unit duration, completion rate, and citation count of the sample courses within the preset learning period, the initial sample labels are revised to obtain the sample labels for the sample courses, including:

[0040] Obtain a second set of samples from multiple sample courses whose initial sample labels are non-inefficient courses;

[0041] Based on the unit duration, completion rate, and number of citations of each sample course in the second sample set within the preset learning period, determine the second average unit duration, second average completion rate, and second average number of citations for all sample courses in the second sample set.

[0042] The second indicator is determined based on the comparison results of the unit time of the sample course within the preset learning period with the second preset unit time threshold and the second average unit time, the comparison results of the completion rate of the sample course within the preset learning period with the second preset completion rate threshold and the second average completion rate, and the comparison results of the number of times the sample course is cited within the preset learning period with the second preset number of citations threshold and the second average number of citations.

[0043] If the second indicator is greater than the preset second indicator threshold, then the inefficient course will be used as the sample label for the sample course.

[0044] If the second indicator is less than or equal to the preset second indicator threshold, the initial sample label will be used as the sample label for the sample course.

[0045] Furthermore, based on the comparison results of the sample course's unit duration within the preset learning period with the second preset unit duration threshold and the second average unit duration, the comparison results of the sample course's completion rate within the preset learning period with the second preset completion rate threshold and the second average completion rate, and the comparison results of the sample course's citation count within the preset learning period with the second preset citation count threshold and the second average citation count, the second indicator is determined, satisfying the following:

[0046]

[0047] Among them, f2 j As the second indicator, C2 j For the j-th sample course in the second sample set, For C2 j The learning time within T, For C2 j Course duration, For C2 j Within the unit duration T, F4 is the second preset unit duration threshold, and A4 is the second average unit duration. For C2 j The completion rate within T, where F5 is the second preset completion rate threshold and A5 is the second average completion rate. For C2 j The number of citations within T, F6 is the second preset citation threshold, A6 is the second average citation count, and α2, β2 and γ2 are adjustment coefficients.

[0048] Secondly, this application provides a method for identifying inefficient courses, including:

[0049] Obtain course information, student information, and historical learning records of the course to be identified;

[0050] The course information, student information, and historical learning record information of the course to be identified are input into the inefficient course identification model for identification processing to obtain the identification result. The inefficient course identification model is the inefficient course identification model provided in this application. The identification result is used to characterize whether the course to be identified is an inefficient course.

[0051] Furthermore, the method also includes:

[0052] When the course is identified as an inefficient course, it will be added to the list of courses to be removed from the platform.

[0053] Thirdly, this application provides a model training apparatus, comprising:

[0054] The acquisition module is used to acquire multiple sample courses;

[0055] The course type module is used to determine the course type of each sample course based on its launch time. Course types include new course type and old course type.

[0056] The sample label module is used to determine the sample label of a sample course based on the course type. The sample label is used to indicate whether the sample course is an inefficient course.

[0057] The training module is used to train the pre-acquired model to obtain an inefficient course recognition model based on course information, student information, historical learning record information, and sample labels from multiple sample courses.

[0058] Fourthly, this application provides an inefficient course identification device, comprising:

[0059] The course acquisition module is used to acquire course information, student information, and historical learning records of the courses to be identified.

[0060] The identification module is used to input the course information, student information, and historical learning record information of the course to be identified into the inefficient course identification model for identification processing, and to obtain the identification result. The inefficient course identification model is the inefficient course identification model provided in this application. The identification result is used to characterize whether the course to be identified is an inefficient course.

[0061] Fifthly, this application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0062] The memory stores the instructions that the computer executes;

[0063] The processor executes computer execution instructions stored in memory to implement the method provided in this application.

[0064] Sixthly, this application provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the method provided in this application.

[0065] The model training method, inefficient course identification method, device, equipment, and medium provided in this application acquire multiple sample courses. For each sample course, the course type is determined based on the course's online time, and then a sample label is determined based on the course type. This yields an evaluation result as to whether a sample course is an inefficient course. Finally, based on the course information, student information, historical learning record information, and sample labels of multiple sample courses, a pre-acquired model to be trained is trained to obtain an inefficient course identification model. This inefficient course identification model improves the accuracy of inefficient course identification. Attached Figure Description

[0066] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0067] Figure 1 A schematic diagram of a model training scenario provided in an embodiment of this application;

[0068] Figure 2 A schematic flowchart illustrating a model training method provided in an embodiment of this application;

[0069] Figure 3 A flowchart illustrating an inefficient course identification method provided in an embodiment of this application;

[0070] Figure 4 This is a schematic diagram of the structure of a model training device provided in an embodiment of this application;

[0071] Figure 5A schematic diagram of the structure of an inefficient course recognition device provided in an embodiment of this application;

[0072] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0073] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0074] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0075] To clearly understand the technical solution of this application, the solutions of the prior art will be described in detail first.

[0076] With the widespread adoption and deep application of the internet, corporate online learning platforms have become an important channel for internal education and knowledge sharing. These platforms provide learning services to all employees, generating corresponding learning data as they access video courses and other content. However, as course development and system construction continue, the number of courses accumulated on these platforms is increasing. This makes it difficult for learners to focus on the courses they need. Furthermore, some older courses have extremely low recent learning frequency, becoming inefficient. Therefore, removing inefficient courses is a necessary step in optimizing the online learning platform's course system. Current technologies for identifying inefficient courses suffer from inaccurate identification.

[0077] To address the aforementioned issue of inaccurate identification, the inventors discovered in their research that sample courses can be categorized by their online launch time, distinguishing between old and new course types. Then, based on these categories, it can be determined whether a sample course is an inefficient course. By considering the correlation between old and new course types and inefficient courses, the resulting inefficient course identification model becomes more accurate.

[0078] The following describes the application scenarios of the model training method provided in the embodiments of this application.

[0079] Figure 1 This is a schematic diagram of a model training scenario provided in an embodiment of this application, such as... Figure 1As shown, the scenario includes a course platform and a server. The course platform includes multiple courses, and the server obtains multiple courses from the course platform as sample courses. For each sample course, the course type is determined based on the online time of the sample course. The course type includes new courses and old courses. Based on the course type, a sample tag is determined for the sample course. The sample tag is used to indicate whether the sample course is an inefficient course. Based on the course information, student information, historical learning record information, and sample tags of multiple sample courses, the pre-obtained model to be trained is trained to obtain an inefficient course identification model.

[0080] A course platform can refer to an electronic device or server that contains courses. For example, a course platform can be an enterprise online learning platform, which can be an application, program, or webpage that includes preset courses and is convenient for enterprise trainees to view. This application does not limit this.

[0081] Figure 2 This is a flowchart illustrating a model training method provided in an embodiment of this application, as shown below. Figure 2 As shown, the method includes:

[0082] S201. Obtain multiple sample courses.

[0083] In this embodiment, in order to train an inefficient course recognition model, a sample set is needed. Therefore, multiple courses can be obtained from the course platform as multiple sample courses to obtain a sample set, which is then used to train the model.

[0084] S202. For each sample course, determine the course type based on the course's launch time. Course types include new course types and old course types.

[0085] In this embodiment, the performance of inefficient courses is often related to the number of times a sample course is studied. The number of times a sample course is studied increases with the online time of the sample course. In order to reduce the impact of the online time of the sample course on the determination of whether it is an inefficient course, it is necessary to distinguish the online time of the sample courses, so as to determine the sample label of the sample course based on the situation of the sample courses with different online times.

[0086] Optionally, a preset time point can be determined first, and then the online time of the sample courses can be compared with the preset time point. The course type of the sample courses can be classified by the preset time point. If the online time of the sample courses is later than the preset time point, the course type of the sample courses is new courses; if the online time of the sample courses is earlier than the preset time point, the course type of the sample courses is old courses.

[0087] For example, the preset time point can be the starting point of a duration of 3 years before the current time. Sample courses within 3 years before the current time are classified as new courses, and sample courses more than 3 years before the current time are classified as old courses.

[0088] S203. Based on the course type, determine the sample label for the sample course. The sample label is used to indicate whether the sample course is an inefficient course.

[0089] In this embodiment, since the sample courses have already been distinguished according to course type, the sample labels of the sample courses can be determined separately according to course type.

[0090] When the sample course is a new course, it means that the sample course has been online for a short time. The number of times the sample course is studied can be used to determine whether the learning value brought by the sample course is inefficient. Therefore, the first number of times the sample course has been studied since it was launched is obtained. If the first number of times of study is less than the preset first number of times of study threshold, the sample course is labeled as an inefficient course. If the first number of times of study is greater than the preset first number of times of study threshold, the sample course is labeled as a non-inefficient course.

[0091] For example, if the preset threshold for the first number of learning sessions is 200, and sample course A, which is a new course, has been learned 220 times since its launch, then sample course A is labeled as a non-inefficient course. If sample course B, which is also a new course, has been learned 180 times since its launch, then sample course B is labeled as an inefficient course.

[0092] When the sample course is an old course, the number of times the sample course is studied increases with the increase of its online duration. Judging the inefficiency of the sample course directly by the total number of studies cannot reflect the inefficiency of the sample course. For the above reasons, the number of studies in a period of time before the current time is used for evaluation. Specifically, the second number of studies of the sample course in the first preset time period is obtained. If the second number of studies is less than the preset second number of studies threshold, the sample course is labeled as an inefficient course. If the second number of studies is greater than the preset second number of studies threshold, the sample course is labeled as a non-inefficient course.

[0093] For example, if the first preset time period is one year prior to the current time, and the preset second learning frequency threshold is 50 times, then sample course C, which is an older course, has been learned 60 times since its launch. Therefore, sample course C is labeled as a non-inefficient course. Conversely, sample course D, which is a new course, has been learned 40 times since its launch. Therefore, sample course D is labeled as an inefficient course.

[0094] Based on historical operational experience, two types of interfering information are relatively common: pseudo-inefficient courses and pseudo-non-inefficient courses. This is because the assignment of the "whether it is an inefficient course" attribute to the sample label may be incorrect when constructing training samples. For example, for sample courses of the "old course" type, although the second learning count within the first preset time period may exceed the second learning count threshold, the learning duration and completion rate are extremely low. This indicates that the sample course was merely opened but not actually studied, thus the course can be considered a pseudo-inefficient course.

[0095] Therefore, to improve the model training effect, this embodiment will modify the aforementioned sample labels. For distinction, the sample labels obtained by the aforementioned method of determining sample labels based on course type are called initial sample labels. That is, the initial sample labels of the sample courses are determined according to the course type, and then the initial sample labels are modified according to the unit duration, completion rate, and number of citations of the sample courses within a preset learning period to obtain the sample labels of the sample courses. The unit duration is obtained by the ratio of the total learning time of the sample courses to the course duration.

[0096] In this embodiment, the unit duration represents the ratio of the total learning time to the course duration of the sample course. For example, if the total learning time of the sample course is 100 hours and the course duration is 1 hour, then the unit duration of the sample course is 100. Since the total learning time of each sample course is affected by the course duration, and different courses have different learning times, the total learning time varies. Judging whether a sample course is inefficient solely based on the total learning time would result in a large error. Therefore, to reduce the influence of course duration, the unit duration is used as a parameter for judging whether a sample course is inefficient.

[0097] The completion rate represents the average ratio of each student's study time in the sample course to the total course duration in the sample course. in, Sample Course C j The completion rate, I j Sample Course C j The number of students For student S i Learning Sample Course C j Study time Sample Course C j The course duration is indicated by ∧, which represents the minimum value between the two sides of the symbol, used to control the completion rate of each student to within 1.

[0098] The number of times a sample course is cited refers to the number of times it is cited by other sections of the course platform, such as a specific learning section. The number of times a sample course is cited reflects its value and function; the more times it is cited, the greater its value.

[0099] In some implementations, when the initial sample label is "inefficient course," the initial sample label is modified based on the sample course's unit duration, completion rate, and number of citations within a preset learning period to obtain the sample label for the sample course. This modification may include:

[0100] Obtain the first sample set of courses whose initial sample label is "inefficient" from multiple sample courses;

[0101] Based on the unit duration, completion rate, and number of citations of each sample course in the first sample set within the preset learning period, determine the first average unit duration, first average completion rate, and first average number of citations for all sample courses in the first sample set.

[0102] The first indicator is determined based on the comparison results of the unit time of the sample course within the preset learning period with the first preset unit time threshold and the first average unit time, the comparison results of the completion rate of the sample course within the preset learning period with the first preset completion rate threshold and the first average completion rate, and the comparison results of the number of times the sample course is cited within the preset learning period with the first preset number of citations threshold and the first average number of citations.

[0103] If the first indicator is greater than the preset first indicator threshold, then the non-inefficient course will be used as the sample label for the sample course.

[0104] If the first indicator is less than or equal to the preset first indicator threshold, the initial sample label will be used as the sample label for the sample course.

[0105] In this embodiment, in order to determine the level of a sample course's unit duration, completion rate, and number of citations within a preset learning period, all inefficient courses are first aggregated to obtain a first sample set. Then, based on the first sample set, the first average unit duration, the first average completion rate, and the first average number of citations are obtained. Considering that determining the sample course solely based on the average value of the first sample set may be incomplete, a preset threshold dimension is also used to determine the sample course's status, thereby ensuring the accuracy of the evaluation of the sample course.

[0106] Specifically, the first indicator can satisfy the following formula:

[0107]

[0108] Among them, f1 j The primary metric is T, where T is the preset learning period, and C1 is the primary metric. j For the j-th sample course in the first sample set, C1 j The learning time within T, C1 j Course duration, C1 j Within a unit duration of T, F1 is the first preset unit duration threshold, and A1 is the first average unit duration. C1 j The completion rate within time T, where F2 is the first preset completion rate threshold and A2 is the first average completion rate. C1 j The number of citations within T, F3 is the first preset citation threshold, A3 is the first average citation count, and α1, β1 and γ1 are adjustment coefficients.

[0109] In this embodiment, ∧ represents taking the smaller value between the two sides for comparing the two parameters, and ∨ represents taking the larger value between the two sides for selecting the better result from the two comparison results. F1, F2, F3, α1, β1, and γ1 can be set as needed, or default values ​​can be used, for example, let α1 = β1 = γ1 = 1.

[0110] The first part of the formula for the first indicator mentioned above This represents the larger of the comparison results between the unit duration of the sample course within the preset learning period and the first preset unit duration threshold, and the comparison results between the unit duration of the sample course within the preset learning period and the first average unit duration. When, it means that the unit duration of the sample course within the preset learning cycle is higher than the first preset unit duration threshold and / or the first average unit duration.

[0111] The second part of the formula for the first indicator mentioned above This represents the larger of the following comparisons: the completion rate of the sample course within the preset learning period compared to the first preset completion rate threshold, and the completion rate of the sample course within the preset learning period compared to the first average completion rate. When the completion rate of the sample course within the preset learning period is higher than the first preset completion rate threshold and / or the first average completion rate, it indicates that the completion rate of the sample course is higher than the first preset completion rate threshold and / or the first average completion rate.

[0112] The third part of the formula for the first indicator mentioned above This represents the larger of the following comparisons: the number of times a sample course is cited within a preset learning period compared to a first preset citation threshold, and the number of times a sample course is cited within a preset learning period compared to a first average citation count. This indicates that the number of times the sample course is cited within the preset learning period is higher than the first preset citation threshold and / or the first average citation.

[0113] As the above analysis shows, the value of the first indicator is 0-3. After obtaining the first indicator, it is compared with a preset first indicator threshold. If the first indicator is greater than the preset first indicator threshold, the initial sample label of the sample course is adjusted based on the comparison result. For example, the preset first indicator threshold is 2.

[0114] In other implementations, when the initial sample label is a non-inefficient course, the initial sample label is modified based on the sample course's unit duration, completion rate, and number of citations within a preset learning period to obtain the sample label for the sample course, which may include:

[0115] Obtain a second set of samples from multiple sample courses whose initial sample labels are non-inefficient courses;

[0116] Based on the unit duration, completion rate, and number of citations of each sample course in the second sample set within the preset learning period, determine the second average unit duration, second average completion rate, and second average number of citations for all sample courses in the second sample set.

[0117] The second indicator is determined based on the comparison results of the unit time of the sample course within the preset learning period with the second preset unit time threshold and the second average unit time, the comparison results of the completion rate of the sample course within the preset learning period with the second preset completion rate threshold and the second average completion rate, and the comparison results of the number of times the sample course is cited within the preset learning period with the second preset number of citations threshold and the second average number of citations.

[0118] If the second indicator is greater than the preset second indicator threshold, then the inefficient course will be used as the sample label for the sample course.

[0119] If the second indicator is less than or equal to the preset second indicator threshold, the initial sample label will be used as the sample label for the sample course.

[0120] In this embodiment, in order to determine the level of a sample course's unit duration, completion rate, and number of citations within a preset learning period, all non-inefficient courses are first aggregated to obtain a second sample set. Then, based on the second sample set, the second average unit duration, second average completion rate, and second average number of citations are obtained. Considering that determining the sample course solely based on the average value of the second sample set may be incomplete, a preset threshold dimension is also used to determine the sample course's status, thereby ensuring the accuracy of the evaluation of the sample course.

[0121] Specifically, the second indicator can satisfy the following formula:

[0122]

[0123] Among them, f2 j As the second indicator, C2 j For the j-th sample course in the second sample set, For C2 j The learning time within T, For C2 j Course duration, For C2 j Within the unit duration T, F4 is the second preset unit duration threshold, and A4 is the second average unit duration. For C2 j The completion rate within T, where F5 is the second preset completion rate threshold and A5 is the second average completion rate. For C2 j The number of citations within T, F6 is the second preset citation threshold, A6 is the second average citation count, and α2, β2 and γ2 are adjustment coefficients.

[0124] In this embodiment, F4, F5, F6, α2, β2, and γ2 can be set as needed, or default values ​​can be used, for example, α2 = β2 = γ2 = 1. Based on the aforementioned analysis of the first indicator, the value of the second indicator is also 0-3. After obtaining the second indicator, it is compared with a preset second indicator threshold. If the second indicator is greater than the preset second indicator threshold, it is determined whether to correct the initial sample label of the sample course based on the comparison result. For example, the preset second indicator threshold is 2.

[0125] S204. Based on the course information, student information, historical learning record information, and sample labels of multiple sample courses, train the pre-acquired model to obtain an inefficient course recognition model.

[0126] In this embodiment, course information may include attributes such as course ID, course duration, course category, course launch time, course keywords, and number of citations, as well as attributes such as total learning duration, total number of learning sessions, completion rate, average learning duration per student, and average number of learning sessions per student within a time period T. Student information may include attributes such as student ID, age, gender, job title, major, and education level, as well as attributes such as total learning duration, total number of learning sessions, completion rate, average learning interval, time of the most recent learning session, and percentage of learning time for special category courses within a time period T. Historical learning records may include attributes such as each student's total learning duration, total number of learning sessions, total learning progress, last learning time, current learning time, current learning duration, and learning portal (special zone, training course, open course, live stream, etc.) within a time period T.

[0127] The model to be trained can learn the feature information of the sample courses to learn the changing relationship between the course information, student information, historical learning record information and sample labels, thereby predicting whether the course to be predicted is an inefficient course. The model to be trained can refer to any possible model, such as neural network models, data mining software, etc., without restriction.

[0128] The model training method provided in this application obtains multiple sample courses. For each sample course, the course type is determined based on the course's online time. Then, the sample label is determined based on the course type, thereby obtaining an evaluation result of whether the sample course is an inefficient course. Finally, based on the course information, student information, historical learning record information, and sample labels of multiple sample courses, the pre-obtained model to be trained is trained to obtain an inefficient course identification model. Through the inefficient course identification model, the accuracy of inefficient course identification is improved.

[0129] Figure 3 This is a flowchart illustrating an inefficient course identification method provided in an embodiment of this application, as shown below. Figure 3 As shown, the method includes:

[0130] S301. Obtain course information, student information, and historical learning record information of the course to be identified.

[0131] In this embodiment, when it is necessary to perform inefficient course identification on the course platform, the course information, student information, and historical learning record information of the course to be identified are obtained from the course platform.

[0132] S302. Input the course information, student information, and historical learning record information of the course to be identified into the inefficient course identification model for identification processing to obtain the identification result.

[0133] In some implementations, when an inefficient course is identified, the course to be identified is added to a list of courses to be removed from the platform, providing data for optimizing the platform's resources.

[0134] For example, when the identification result is an inefficient course, the course ID of the course to be identified can be obtained and written into a preset list of courses to be removed from the platform. The list of courses to be removed from the platform can be preset in the database.

[0135] In this embodiment, the inefficient course identification model is a model obtained according to the model training method of this application, and the identification result is used to characterize whether the course to be identified is an inefficient course.

[0136] The inefficient course identification method provided in this application obtains course information, student information, and historical learning record information of the course to be identified, and then inputs the course information, student information, and historical learning record information of the course to be identified into the inefficient course identification model for identification processing to obtain the identification result, thereby improving the accuracy of inefficient course identification.

[0137] Figure 4 This is a schematic diagram of the structure of a model training device provided in an embodiment of this application, as shown below. Figure 4 As shown, the device includes:

[0138] Module 401 is used to acquire multiple sample courses;

[0139] The course type module 402 is used to determine the course type of each sample course based on its launch time. The course types include new course type and old course type.

[0140] The sample label module 403 is used to determine the sample label of the sample course based on the course type. The sample label is used to indicate whether the sample course is an inefficient course.

[0141] Training module 404 is used to train a pre-acquired model based on course information, student information, historical learning record information, and sample labels from multiple sample courses to obtain an inefficient course recognition model.

[0142] In some implementations, the course type module 402 is also used for:

[0143] For each sample course, determine the launch time and the preset time point;

[0144] If the sample course is launched later than the preset time, the course type of the sample course is a new course.

[0145] If the sample course is launched earlier than the preset time, the course type of the sample course is the old course type.

[0146] In some implementations, the sample label module 403 is also used for:

[0147] Determine the initial sample labels for the sample courses based on the course type;

[0148] Based on the unit duration, completion rate, and number of citations of the sample courses within the preset learning period, the initial sample labels are corrected to obtain the sample labels of the sample courses. The unit duration is obtained by the ratio of the total learning time of the sample courses to the course duration.

[0149] In some implementations, when the sample course type is a new course, the sample tag module 403 is also used for:

[0150] Based on the course type, determine the initial sample labels for the sample courses, including:

[0151] Get the first number of times the sample course has been viewed since its launch;

[0152] If the first learning attempt is less than the preset first learning attempt threshold, the initial sample label of the sample course is an inefficient course.

[0153] If the first learning attempt exceeds the preset first learning attempt threshold, the initial sample label for the sample course is "non-inefficient course".

[0154] In some implementations, when the sample course type is an old course type, the sample tag module 403 is also used for:

[0155] Based on the course type, determine the initial sample labels for the sample courses, including:

[0156] Obtain the second learning session of the sample course within the first preset time period;

[0157] If the second learning count is less than the preset second learning count threshold, the initial sample label of the sample course is an inefficient course.

[0158] If the second learning count is greater than the preset second learning count threshold, the initial sample label for the sample course is "non-inefficient course".

[0159] In some implementations, when the initial sample label is an inefficient course, the sample label module 403 is also used for:

[0160] Obtain the first sample set of courses whose initial sample label is "inefficient" from multiple sample courses;

[0161] Based on the unit duration, completion rate, and number of citations of each sample course in the first sample set within the preset learning period, determine the first average unit duration, first average completion rate, and first average number of citations for all sample courses in the first sample set.

[0162] The first indicator is determined based on the comparison results of the unit time of the sample course within the preset learning period with the first preset unit time threshold and the first average unit time, the comparison results of the completion rate of the sample course within the preset learning period with the first preset completion rate threshold and the first average completion rate, and the comparison results of the number of times the sample course is cited within the preset learning period with the first preset number of citations threshold and the first average number of citations.

[0163] If the first indicator is greater than the preset first indicator threshold, then the non-inefficient course will be used as the sample label for the sample course.

[0164] If the first indicator is less than or equal to the preset first indicator threshold, the initial sample label will be used as the sample label for the sample course.

[0165] In some implementations, when the initial sample label is an inefficient course, the sample label module 403 is also used for:

[0166] satisfy:

[0167]

[0168] Among them, f1 j The primary metric is T, where T is the preset learning period, and C1 is the primary metric. j For the j-th sample course in the first sample set, C1 j The learning time within T, C1 j Course duration, C1 j Within a unit duration of T, F1 is the first preset unit duration threshold, and A1 is the first average unit duration. C1 j The completion rate within time T, where F2 is the first preset completion rate threshold and A2 is the first average completion rate. C1 j The number of citations within T, F3 is the first preset citation threshold, A3 is the first average citation count, and α1, β1 and γ1 are adjustment coefficients.

[0169] In some implementations, when the initial sample label is a non-inefficient course, the sample label module 403 is also used for:

[0170] Obtain a second set of samples from multiple sample courses whose initial sample labels are non-inefficient courses;

[0171] Based on the unit duration, completion rate, and number of citations of each sample course in the second sample set within the preset learning period, determine the second average unit duration, second average completion rate, and second average number of citations for all sample courses in the second sample set.

[0172] The second indicator is determined based on the comparison results of the unit time of the sample course within the preset learning period with the second preset unit time threshold and the second average unit time, the comparison results of the completion rate of the sample course within the preset learning period with the second preset completion rate threshold and the second average completion rate, and the comparison results of the number of times the sample course is cited within the preset learning period with the second preset number of citations threshold and the second average number of citations.

[0173] If the second indicator is greater than the preset second indicator threshold, then the inefficient course will be used as the sample label for the sample course.

[0174] If the second indicator is less than or equal to the preset second indicator threshold, the initial sample label will be used as the sample label for the sample course.

[0175] In some implementations, when the initial sample label is a non-inefficient course, the sample label module 403 is also used for:

[0176] satisfy:

[0177]

[0178] Among them, f2 j As the second indicator, C2 j For the j-th sample course in the second sample set, For C2 j The learning time within T, For C2 j Course duration, For C2 j Within the unit duration T, F4 is the second preset unit duration threshold, and A4 is the second average unit duration. For C2 j The completion rate within T, where F5 is the second preset completion rate threshold and A5 is the second average completion rate. For C2 j The number of citations within T, F6 is the second preset citation threshold, A6 is the second average citation count, and α2, β2 and γ2 are adjustment coefficients.

[0179] Figure 5 This is a schematic diagram of the structure of an inefficient course recognition device provided in an embodiment of this application, as shown below. Figure 5 As shown, the device includes:

[0180] The course acquisition module 501 is used to acquire course information, student information, and historical learning record information of the course to be identified;

[0181] The identification module 502 is used to input the course information, student information, and historical learning record information of the course to be identified into the inefficient course identification model for identification processing to obtain the identification result. The inefficient course identification model is the inefficient course identification model provided in this application. The identification result is used to characterize whether the course to be identified is an inefficient course.

[0182] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 6 As shown, the electronic device 60 includes:

[0183] The electronic device 60 may include a processor 601 with one or more processing cores, a memory 602 with one or more computer-readable storage media, a communication component 603, and other components. The processor 601, memory 602, and communication component 603 are connected via a bus 604.

[0184] In the specific implementation process, at least one processor 601 executes computer execution instructions stored in memory 602, causing at least one processor 601 to execute the above-mentioned model training method or inefficient course recognition method.

[0185] The specific implementation process of processor 601 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0186] In the above Figure 6 In the illustrated embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0187] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0188] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0189] In some embodiments, a computer program product is also provided, comprising a computer program or instructions that, when executed by a processor, implement the steps in any of the model training methods or inefficient curriculum identification methods described above.

[0190] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0191] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0192] Therefore, embodiments of this application provide a computer-readable storage medium storing a plurality of instructions that can be loaded by a processor to execute steps in any of the model training methods or inefficient course recognition methods provided in embodiments of this application.

[0193] The storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0194] According to one aspect of this application, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium.

[0195] Since the instructions stored in the storage medium can execute the steps of any of the model training methods or inefficient course recognition methods provided in the embodiments of this application, the beneficial effects that any of the model training methods or inefficient course recognition methods provided in the embodiments of this application can achieve can be realized. For details, please refer to the previous embodiments, which will not be repeated here.

[0196] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0197] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A model training method, characterized in that, The method comprises the following steps: acquiring a plurality of sample courses; for each sample course, determining a course type of the sample course according to an online time of the sample course, the course type comprising a new course type and an old course type; determining a sample label of the sample course according to the course type, the sample label being used to indicate whether the sample course is an inefficient course; performing model training on a pre-acquired to-be-trained model according to course information, student information, historical learning record information and the sample label of the plurality of sample courses, to obtain an inefficient course identification model.

2. The method of claim 1, wherein, The step of determining the course type of the sample course according to the online time of the sample course comprises the following steps: for each sample course, determining an online time of the sample course and a preset time point; if the online time of the sample course is later than the preset time point, the course type of the sample course is the new course type; if the online time of the sample course is earlier than the preset time point, the course type of the sample course is the old course type.

3. The method of claim 1, wherein, The step of determining the sample label of the sample course according to the course type comprises the following steps: determining an initial sample label of the sample course according to the course type; correcting the initial sample label according to a unit time length, a completion rate and a reference count of the sample course within a preset learning period, to obtain the sample label of the sample course, the unit time length being obtained according to a ratio of a total learning time length and a course time length.

4. The method of claim 3, wherein, When the course type of the sample course is the new course type, the step of determining the initial sample label of the sample course according to the course type comprises the following steps: acquiring a first learning count of the sample course since being online; if the first learning count is less than a preset first learning count threshold, the initial sample label of the sample course is an inefficient course; if the first learning count is greater than the preset first learning count threshold, the initial sample label of the sample course is a non-inefficient course.

5. The method of claim 3, wherein, When the course type of the sample course is the old course type, the step of determining the initial sample label of the sample course according to the course type comprises the following steps: acquiring a second learning count of the sample course within a first preset time period; if the second learning count is less than a preset second learning count threshold, the initial sample label of the sample course is an inefficient course; if the second learning count is greater than the preset second learning count threshold, the initial sample label of the sample course is a non-inefficient course.

6. The method according to any of claims 3-5, characterized by, When the initial sample label is an inefficient course, the step of correcting the initial sample label according to the unit time length, the completion rate and the reference count of the sample course within the preset learning period, to obtain the sample label of the sample course, comprises the following steps: acquiring a first sample set in which the initial sample label of each sample course in the plurality of sample courses is an inefficient course; determining a first average unit time length, a first average completion rate and a first average reference count of all sample courses in the first sample set according to the unit time length, the completion rate and the reference count of each sample course in the first sample set within the preset learning period; determining a first index according to comparison results of the unit length of the sample course within the preset learning period with a first preset unit length threshold and the first average unit length, comparison results of the completion rate of the sample course within the preset learning period with a first preset completion rate threshold and the first average completion rate, and comparison results of the number of citations of the sample course within the preset learning period with a first preset number of citations threshold and the first average number of citations; if the first index is greater than a preset first index threshold, regarding a non-inefficient course as a sample label of the sample course; if the first index is less than or equal to the preset first index threshold, regarding the initial sample label as the sample label of the sample course.

7. The method of claim 6, wherein, The determination of the first index according to the comparison results of the unit length of the sample course within the preset learning period with the first preset unit length threshold and the first average unit length, the comparison results of the completion rate of the sample course within the preset learning period with the first preset completion rate threshold and the first average completion rate, and the comparison results of the number of citations of the sample course within the preset learning period with the first preset number of citations threshold and the first average number of citations satisfies: wherein f1 j is the first index, T is the preset learning period, C1 j is the jth sample course in the first sample set, is the course length of C1 j in T, is the course length of C1 j in T, is the unit length of C1 j in T, F1 is the first preset unit length threshold, and A1 is the first average unit length, is the completion rate of C1 j in T, F2 is the first preset completion rate threshold, and A2 is the first average completion rate, is the number of citations of C1 j in T, F3 is the first preset number of citations threshold, A3 is the first average number of citations, and α1, β1, and γ1 are adjustment coefficients.

8. The method according to any of claims 3-5, characterized by, when the initial sample label is a non-inefficient course, the correction of the initial sample label according to the unit length, the completion rate and the number of citations of the sample course within the preset learning period to obtain the sample label of the sample course includes: obtaining a second sample set in which the initial sample label of each sample course in the plurality of sample courses is a non-inefficient course; determining a second average unit length, a second average completion rate and a second average number of citations of all sample courses in the second sample set according to the unit length, the completion rate and the number of citations of each sample course in the second sample set within the preset learning period; determining a second index according to comparison results of the unit length of the sample course within the preset learning period with a second preset unit length threshold and the second average unit length, comparison results of the completion rate of the sample course within the preset learning period with a second preset completion rate threshold and the second average completion rate, and comparison results of the number of citations of the sample course within the preset learning period with a second preset number of citations threshold and the second average number of citations; if the second index is greater than a preset second index threshold, regarding an inefficient course as a sample label of the sample course; if the second index is less than or equal to the preset second index threshold, regarding the initial sample label as the sample label of the sample course.

9. The method of claim 8, wherein, The second indicators are determined according to comparison results of unit time lengths of the sample courses in the preset learning period with second preset unit time length thresholds and the second average unit time length, comparison results of completion rates of the sample courses in the preset learning period with second preset completion rate thresholds and the second average completion rate, and comparison results of cited times of the sample courses in the preset learning period with second preset cited time thresholds and the second average cited time, and meet: wherein f2 j is the second index, C2 j is the jth sample course in the second sample set, is C2 j the learning duration within T, is C2 j the course duration, is C2 j the unit duration within T, F4 is the second preset unit duration threshold, and A4 is the second average unit duration. is C2 j the completion rate within T, F5 is the second preset completion rate threshold, A5 is the second average completion rate, and CT C2j,T is C2 j the number of citations within T, F6 is the second preset number of citations threshold, A6 is the second average number of citations, and a2, b2, and g2 are adjustment coefficients.

10. A method for identifying low-efficiency courses, the method comprising: Comprise: Obtaining course information, student information, and historical learning record information of a to-be-identified course; Inputting the course information, student information, and historical learning record information of the to-be-identified course into the low-efficiency course identification model for identification processing to obtain an identification result, the low-efficiency course identification model being the low-efficiency course identification model in any one of claims 1-9, and the identification result being used to represent whether the to-be-identified course is a low-efficiency course.

11. The method of claim 10, wherein, The method further comprises: When the identification result is a low-efficiency course, adding the to-be-identified course to a list of courses to be delisted.

12. A model training apparatus, comprising: Comprise: An obtaining module is configured to obtain a plurality of sample courses; A course type module is configured to, for each sample course, determine a course type of the sample course according to an online time of the sample course, the course type comprising a new course type and an old course type; A sample label module is configured to determine a sample label of the sample course according to the course type, the sample label being used to indicate whether the sample course is a low-efficiency course; A training module is configured to perform model training on a to-be-trained model that is obtained in advance according to course information, student information, historical learning record information, and sample labels of the plurality of sample courses to obtain a low-efficiency course identification model.

13. A low-effort course identification apparatus, comprising: Comprise: A course obtaining module is configured to obtain course information, student information, and historical learning record information of a to-be-identified course; An identification module is configured to input the course information, student information, and historical learning record information of the to-be-identified course into a low-efficiency course identification model for identification processing to obtain an identification result, the low-efficiency course identification model being the low-efficiency course identification model in any one of claims 1-9, and the identification result being used to represent whether the to-be-identified course is a low-efficiency course.

14. An electronic device, comprising: Comprise: A processor and a memory in communication connection with the processor; The memory stores computer execution instructions; The processor executes the computer execution instructions stored in the memory to implement the method in any one of claims 1-11.

15. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer execution instructions, and the computer execution instructions are executed by the processor to implement the method in any one of claims 1-11.

Images