A method and system for evaluating the effectiveness of online learning

Abstract

The present application relates to the technical field of learning effectiveness evaluation, and in particular to an effectiveness evaluation method and system for online learning. The method includes: obtaining an online operation behavior data set, analyzing the online operation behavior data set, and determining a weak subjective dimension effectiveness indicator set; obtaining a physiological data set, analyzing the physiological data set, and determining a strong objective dimension effectiveness indicator set; based on the strong objective dimension effectiveness indicator set, analyzing the weak subjective dimension effectiveness indicator set, and determining the deep learning effectiveness evaluation results; obtaining a learning feedback data set, and based on the deep learning effectiveness indicator evaluation results, analyzing the learning feedback data set, and determining and outputting an online learning effectiveness report. The present application uses multi-dimensional analysis methods to scientifically and comprehensively reflect the deep learning effectiveness of students in the online learning process, and provides scientific suggestions for improving students' subsequent learning plans through online learning effectiveness reports, so as to guide students to find online learning methods that suit them and improve their learning quality.

Description

Technical Field

[0001] The present application relates to the technical field of learning effectiveness evaluation, and in particular to an effectiveness evaluation method and system for online learning. Background Art

[0002] With the popularization of the Internet and the advancement of technology, online learning has become a common way of learning. This model is not only convenient and flexible, allowing learners to acquire knowledge at any time and any place, but also provides educational institutions with a wider range of teaching and learning opportunities. Online learning effectiveness evaluation is an important means to ensure the quality of online learning.

[0003] However, the existing online learning effectiveness evaluation technology is based on the analysis method of online learning representation data, which makes it difficult to comprehensively and accurately evaluate the deep learning effectiveness of students during the online learning process. As a result, it is difficult to significantly improve the quality of students' online learning, which in turn leads to the development of students' high-order thinking ability being restricted during the online learning process. Summary of the Invention

[0004] This application provides an online learning effectiveness evaluation method and system to solve the above technical problems.

[0005] In a first aspect, the present application provides a method for evaluating the effectiveness of online learning, the method comprising:

[0006] Acquire an online operation behavior dataset, analyze the online operation behavior dataset, and determine a weak subjective dimension performance indicator set;

[0007] Acquiring a physiological data set, analyzing the physiological data set, and determining a set of strong objective dimension performance indicators;

[0008] Based on the strong objective dimension effectiveness indicator set, the weak subjective dimension effectiveness indicator set is analyzed to determine the deep learning effectiveness evaluation result;

[0009] Acquire a learning feedback data set, analyze the learning feedback data set based on the deep learning effectiveness indicator evaluation results, and determine and output an online learning effectiveness report.

[0010] Through this scheme, the students' learning outcomes are analyzed from the weak subjective dimension and the strong objective dimension based on their online operation behavior dataset and physiological dataset, and the corresponding indicators reflecting the students' learning outcomes are quantified to construct a weak subjective dimension effectiveness indicator set and a strong objective dimension effectiveness indicator set. On this basis, the deep learning effectiveness evaluation results that fully reflect the students' deep learning outcomes are integrated, and combined with the learning feedback dataset, the online learning effectiveness report is determined and output. Through multi-dimensional analysis methods, the deep learning outcomes of the students in the online learning process are scientifically and comprehensively reflected. The online learning effectiveness report provides scientific suggestions for improving the students' subsequent learning plans, so as to guide students to find online learning methods that suit them and improve their learning quality.

[0011] Optionally, the online operation behavior dataset includes learning time, interaction frequency, number of pauses, and number of fast-forwards. Analyzing the online operation behavior dataset to determine the weak subjective dimension performance indicator set includes:

[0012] Obtaining course content segmentation information, and determining an ideal learning duration and a normal interaction frequency based on the course content segmentation information;

[0013] Analyze the learning time and the ideal learning time to determine a time impact index;

[0014] Analyze the interaction frequency and the normal interaction frequency to determine an interaction impact index;

[0015] Analyzing the number of pauses and the number of fast-forwards to determine a pause impact index and a fast-forward impact index, respectively;

[0016] Determining a weak subjective comprehensive evaluation index according to the duration impact index, the interaction impact index, the pause impact index, and the fast-forward impact index;

[0017] The weak subjective dimension effectiveness indicator set is constructed according to the duration impact index, the interaction impact index, the pause impact index, the fast-forward impact index and the weak subjective comprehensive evaluation index.

[0018] Through this plan, the four data dimensions of learning time, interaction frequency, number of pauses and number of fast-forwards in the students' operational behavior are analyzed respectively to analyze their impact on the students' learning level, and the time impact index, interaction impact index, pause impact index and fast-forward impact index are obtained. The impact indices corresponding to the four data dimensions are integrated to obtain a weak subjective comprehensive evaluation index that can comprehensively reflect the students' learning outcomes under the weak subjective dimension, and based on this, a weak subjective dimension effectiveness indicator set is constructed, which significantly improves the scientificity and comprehensiveness of the students' learning effectiveness evaluation process under the weak subjective dimension. At the same time, by incorporating multidimensional impact indicators into the weak subjective dimension effectiveness set, the subsequent deficiencies in the students' learning methods can be more accurately located.

[0019] Optionally, the weak subjective comprehensive evaluation index is determined according to the duration impact index, the interaction impact index, the pause impact index, and the fast-forward impact index, specifically as follows:

[0020]

[0021] Among them, E is the weak subjective comprehensive evaluation index, k1 is the preset time influence coefficient, T is the learning time, α1 is the preset time influence index, T d For the ideal study time, is the duration impact index, k2 is the preset interaction impact coefficient, I is the interaction frequency, α2 is the preset interaction impact index, I d For normal interaction frequency, is the interaction influence index, k3 is the preset pause influence coefficient, P is the number of pauses, β is the preset pause influence index, k3·P β is the pause impact index, k4 is the preset fast-forward impact coefficient, α3 is the preset fast-forward impact index, F is the fast-forward times, is the fast-forward impact index.

[0022] Through this plan, mathematical analysis methods are used to clearly and accurately quantify the duration impact index, interaction impact index, pause impact index and fast-forward impact index according to the learning time, ideal learning time, interaction frequency, normal interaction frequency, number of pauses and number of fast-forwards through corresponding mathematical formulas. Based on this, several subjective comprehensive evaluation indexes are scientifically calculated to improve the accuracy and scientificity of the learning effectiveness evaluation under the weak subjective dimension, and provide clear quantitative data support for the subsequent improvement analysis of students' learning effectiveness.

[0023] Optionally, the physiological data set includes eye movement data, facial expression change data, heart rate change data, and skin conductance change data. The analyzing the physiological data set to determine a strong objective dimension performance indicator set includes:

[0024] performing focus change frequency analysis on the eye movement data to determine an attention index;

[0025] Performing emotion change analysis on the expression change data to determine an emotion index;

[0026] Determining an attention impact index and an emotion impact index based on the attention index and the emotion index, and determining an index representing physiological effectiveness based on the index;

[0027] determining an average heart rate and an average skin conductance according to the heart rate variation data and the skin conductance variation data;

[0028] Determining a heart rate impact index and a skin conductance impact index based on the average heart rate and the average skin conductance, and determining a deep physiological effectiveness index accordingly;

[0029] The strong objective dimension performance indicator set is constructed based on the attention impact index, the emotion impact index, the physiological performance index, the heart rate impact index, the skin conductance impact index and the deep physiological performance index.

[0030] Through this plan, the students' physiological data are divided into representational dimensions and deep dimensions. Starting from the eye movement data, expression change data, heart rate change data and skin conductance change data corresponding to the two dimensions, the attention impact indicators, emotion impact indicators, heart rate impact indicators and skin conductance impact indicators that will directly affect learning outcomes are analyzed respectively. On this basis, the representational physiological effectiveness index and the deep physiological effectiveness index are determined respectively. Based on the above-mentioned several impact indicators and indices, a strong objective dimension effectiveness indicator set is constructed to improve the scientificity and comprehensiveness of the learning effectiveness evaluation under the strong objective dimension, and at the same time further improve the accuracy of the subsequent analysis of students' shortcomings.

[0031] Optionally, the attention impact index and the emotion impact index are determined respectively according to the attention index and the emotion index, and the physiological effectiveness index is determined accordingly, specifically as follows:

[0032]

[0033] Wherein, C is the physiological efficacy index, k a is the preset attention influence coefficient, γ is the preset attention adjustment index, A is the attention index, is the attention impact index, k e is the preset emotion influence coefficient, δ is the preset emotion suppression parameter, ψ is the preset emotion enhancement parameter, S is the emotion index, k e ·(-δS 2 +ψS) is the emotion impact index.

[0034] Through this plan, mathematical analysis methods are used to quantify the attention impact index and the emotion impact index based on the attention index and the emotion index respectively. On this basis, the physiological effectiveness index is calculated through a clear mathematical formula to achieve scientific quantification of the attention impact index, emotion impact index and physiological effectiveness index, and improve the accuracy of the analysis of learning effectiveness in the physiological dimension.

[0035] Optionally, the heart rate impact index and the skin conductance impact index are determined based on the average heart rate and the average skin conductance, and the deep physiological effectiveness index is determined accordingly, specifically as follows:

[0036]

[0037] Where D is the deep physiological efficacy index, k H is the preset heart rate influence coefficient, λ is the preset heart rate influence adjustment parameter, H is the average heart rate, H t is the reference value of resting heart rate, is the heart rate impact index, k G is the preset skin conductance influence coefficient, G is the average skin conductivity, k G G is the skin conductance impact index.

[0038] Through this solution, mathematical analysis methods are used to quantify the heart rate impact index and the skin conductance impact index based on the average heart rate and the average skin conductivity, respectively. On this basis, the deep physiological effectiveness index is accurately calculated through a clear mathematical formula, thereby achieving scientific quantification of the heart rate impact index, skin conductance impact index and deep physiological effectiveness index, and improving the accuracy of the analysis of learning effectiveness in the deep physiological dimension.

[0039] Optionally, analyzing the weak subjective dimension effectiveness indicator set based on the strong objective dimension effectiveness indicator set to determine the deep learning effectiveness evaluation result includes:

[0040] Merging the strong objective dimension performance indicator set with the weak subjective dimension performance indicator set to determine a deep learning performance evaluation indicator set corresponding to each student;

[0041] Obtaining an indicator weight set, and constructing a multi-user deep learning effectiveness scoring matrix based on the indicator weight set and the deep learning effectiveness evaluation indicator set corresponding to each student;

[0042] Determining a deep learning effectiveness score for each student based on the multi-user deep learning effectiveness scoring matrix;

[0043] The multi-user deep learning effectiveness rating matrix and the deep learning effectiveness ratings of all students are used as the deep learning effectiveness evaluation results.

[0044] Through this solution, the strong objective dimension effectiveness indicator set and the weak subjective dimension effectiveness indicator set are merged to obtain the deep learning effectiveness evaluation indicator set. Combined with the indicator weight set, a scientific evaluation framework for students' deep learning effectiveness is constructed in the form of a mathematical matrix. The multi-user deep learning effectiveness scoring matrix and the deep learning effectiveness scores of all students are used as the deep learning effectiveness evaluation results. While improving the accuracy of learning effectiveness evaluation, it provides an important data framework for subsequent analysis of the direction of improvement of students' learning effectiveness.

[0045] Optionally, based on the indicator weight set and according to the deep learning effectiveness evaluation indicator set corresponding to each student, a multi-user deep learning effectiveness scoring matrix is ​​constructed, which is specifically the following matrix expression:

[0046]

[0047] Where I is the multi-user deep learning performance scoring matrix, Y mn is the nth evaluation indicator in the deep learning effectiveness evaluation indicator set of the mth student, W n is the weight value corresponding to the nth evaluation indicator in the indicator weight set;

[0048] The deep learning effectiveness score of each student is determined according to the multi-user deep learning effectiveness score matrix, specifically as follows:

[0049]

[0050] Among them, RI i Score the deep learning effectiveness of the i-th student, n is the total number of evaluation indicators, I j The effectiveness score of the current student corresponding to the jth evaluation indicator in the multi-user deep learning effectiveness scoring matrix.

[0051] Through this solution, mathematical analysis methods are used to clarify the multi-user deep learning effectiveness scoring matrix through mathematical expressions, and mathematical formulas are used to accurately calculate the deep learning effectiveness score of each student based on the multi-user deep learning effectiveness scoring matrix, further improving the scientificity and accuracy of the learning effectiveness evaluation process.

[0052] Optionally, analyzing the learning feedback dataset based on the deep learning effectiveness indicator evaluation results to determine and output an online learning effectiveness report includes:

[0053] Analyze the learning feedback data set to determine a multi-type feedback indicator information set corresponding to each student;

[0054] Extracting the corresponding deep learning effectiveness evaluation indicator set in the multi-user deep learning effectiveness scoring matrix according to the multi-type feedback indicator information set;

[0055] Analyzing the deep learning effectiveness evaluation indicator set based on the multi-type feedback indicator information set to determine a number of conflicting feedback indicators;

[0056] The online learning effectiveness report is constructed and outputted based on the deep learning effectiveness score of each student and a plurality of the conflicting feedback indicators.

[0057] Through this plan, on the basis of completing the evaluation of students' learning outcomes, we further combine the students' feedback data, extract a number of contradictory feedback indicators based on the multi-type feedback indicator information set and the deep learning effectiveness evaluation indicator set, and construct and output an online learning effectiveness report based on each student's deep learning effectiveness score and a number of contradictory feedback indicators to clarify the areas that students need to focus on in order to improve their learning outcomes in the future, and provide a significant positive effect on improving the quality of students' learning.

[0058] In a second aspect, the present application provides an online learning effectiveness evaluation system, the system comprising:

[0059] A weak subjective analysis module is used to obtain an online operation behavior data set, analyze the online operation behavior data set, and determine a weak subjective dimension performance indicator set;

[0060] A strong objective analysis module, configured to obtain a physiological data set, analyze the physiological data set, and determine a set of strong objective dimension performance indicators;

[0061] A deep analysis module, configured to analyze the weak subjective dimension effectiveness indicator set based on the strong objective dimension effectiveness indicator set to determine a deep learning effectiveness evaluation result;

[0062] The output module is used to obtain a learning feedback data set, analyze the learning feedback data set based on the deep learning effectiveness indicator evaluation results, and determine and output an online learning effectiveness report. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0064] Figure 1A schematic diagram of an application scenario provided in one embodiment of the present application;

[0065] Figure 2 A flowchart of an online learning effectiveness evaluation method provided in one embodiment of the present application;

[0066] Figure 3 This is a schematic diagram of the structure of an online learning effectiveness evaluation system provided in one embodiment of the present application. DETAILED DESCRIPTION

[0067] To make the purpose, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0068] In this document, the term "and / or" simply describes a relationship between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " in this document, unless otherwise specified, generally indicates an "or" relationship between the related objects.

[0069] The embodiments of the present application are described in further detail below with reference to the accompanying drawings.

[0070] The existing online learning effectiveness evaluation technology is based on the analysis method of online learning representation data, which makes it difficult to comprehensively and accurately evaluate the deep learning effectiveness of students during the online learning process, resulting in difficulty in significantly improving the quality of students' online learning, and further limiting the development of students' high-order thinking ability during the online learning process.

[0071] Based on this, the present application provides an online learning effectiveness evaluation method and system. Based on the students' online operation behavior dataset and physiological dataset, the students' learning effectiveness is analyzed from the weak subjective dimension and the strong objective dimension, and the corresponding indicators reflecting the students' learning effectiveness are quantified to construct a weak subjective dimension effectiveness indicator set and a strong objective dimension effectiveness indicator set. On this basis, the deep learning effectiveness evaluation results that fully reflect the students' deep learning effectiveness are integrated, and combined with the learning feedback dataset, an online learning effectiveness report is determined and output. Through multi-dimensional analysis methods, the deep learning effectiveness of the students in the online learning process is scientifically and comprehensively reflected, and the online learning effectiveness report provides scientific suggestions for the improvement of the students' subsequent learning plans, so as to guide the students to find online learning methods that suit them and improve the quality of their learning.

[0072] Figure 1 This is a schematic diagram of an application scenario provided by this application. In the process of evaluating students' online learning results, the method provided by this application is applied to accurately evaluate students' deep learning results to improve the quality of students' online learning.

[0073] Specifically, the method of the present application is applied to any server, which communicates with the background log system and the physiological data collection device respectively, and obtains the online operation behavior data set provided by the background log system and the physiological data set provided by the physiological data collection device through the server. According to the student's online operation behavior data set and physiological data set, the student's learning performance is analyzed from the weak subjective dimension and the strong objective dimension, and the corresponding indicators reflecting the student's learning performance are quantified to construct a weak subjective dimension performance indicator set and a strong objective dimension performance indicator set. On this basis, the deep learning performance evaluation results that fully reflect the student's deep learning performance are integrated, and combined with the learning feedback data set provided by the student, an online learning performance report is determined and output. Through multi-dimensional analysis methods, the deep learning performance of the student in the online learning process is scientifically and comprehensively reflected, and the online learning performance report provides scientific suggestions for the improvement of the student's subsequent learning plan, so as to guide the student to find an online learning method that suits him or her and improve the student's learning quality. The specific implementation method can refer to the following embodiments.

[0074] Figure 2 This is a flow chart of an online learning effectiveness evaluation method provided by an embodiment of the present application. The method of this embodiment can be applied to the server in the above scenario. Figure 2 As shown, the method includes:

[0075] S201. Obtain an online operation behavior dataset, analyze the online operation behavior dataset, and determine a weak subjective dimension effectiveness indicator set.

[0076] The online operation behavior dataset can be a data set corresponding to a series of operation behaviors performed by students in the online learning platform during the online learning process. The online operation behavior dataset can be obtained through the background log system of the online learning platform.

[0077] The weak subjective dimension effectiveness indicator set may be a collection of a series of indicators that can reflect the student's learning effectiveness in online operation behaviors with a weak subjective color of the student.

[0078] Specifically, online learning is often in a relatively autonomous and flexible learning environment. Students choose when, where and how to learn. This flexibility makes it easier for students to adjust their learning methods according to their own needs. The series of operational behaviors generated by students in the process of learning on the learning platform can reveal their learning strategies. The operational behaviors performed by students on the learning platform are usually generated unconsciously based on their learning habits and learning status. Although they have a certain degree of subjective consciousness, this subjective consciousness has a clear weakening trend compared with the objective feedback brought by the operational behavior. Therefore, the students' online operational behavior can reflect their learning effectiveness from the weak subjective dimension. Through mathematical analysis, the students' online operational behavior data set is scientifically processed, and several indicators that can reflect the students' learning effectiveness in the weak subjective dimension are extracted and quantified to construct a weak subjective dimension effectiveness indicator set.

[0079] S202: Acquire a physiological data set, analyze the physiological data set, and determine a set of strong objective dimension performance indicators.

[0080] The physiological data set can be a collection of a series of physiological data collected by students during the learning process, such as heart rate and eye movement data. The physiological data set can be obtained through different physiological data collection devices, such as wearable devices and eye trackers, after obtaining the student's authorization and consent.

[0081] The strong objective dimension performance indicator set can be a collection of a series of indicators in the students' physiological data with guaranteed objectivity that can reflect the students' learning outcomes.

[0082] Specifically, when students use online learning platforms to study, they are affected by a series of positive and negative influences such as attention, learning pressure, and learning achievement. Their physiological data, such as heart rate, eye movement data, and skin conductivity, will change significantly. The changes in physiological data can reflect the cognitive load and emotional state of students during the learning process, and thus directly affect the students' learning outcomes. In addition, the students' physiological data cannot be controlled by the students' subjective behavior and are highly objective. Therefore, physiological data can reflect the students' learning outcomes from a strongly objective dimension. Through mathematical analysis, the students' physiological data sets are scientifically processed, and several indicators that can reflect the students' learning outcomes in a strongly objective dimension are extracted and quantified to construct a set of effectiveness indicators in a strongly subjective dimension.

[0083] S203. Based on the strong objective dimension effectiveness indicator set, analyze the weak subjective dimension effectiveness indicator set to determine the deep learning effectiveness evaluation results.

[0084] The results of deep learning effectiveness evaluation can be the evaluation information obtained by analyzing the combination of weak subjective dimension effectiveness indicator set and strong objective dimension effectiveness indicator set, which can objectively and comprehensively reflect the learning effectiveness of students.

[0085] Specifically, the strong objective dimension effectiveness indicator set and the weak subjective dimension effectiveness indicator set obtained in the above steps reflect the students' learning effectiveness from the strong objective dimension and the weak subjective dimension respectively. By combining the two, we can have a deeper insight into the students' learning outcomes and improve the comprehensiveness of the evaluation of students' learning effectiveness. Using mathematical analysis methods, we integrate the strong objective dimension effectiveness indicator set and the weak subjective dimension effectiveness indicator set to obtain deep learning effectiveness evaluation results, and use scientific mathematical language to comprehensively and accurately describe the students' learning effectiveness.

[0086] S204: Obtain a learning feedback data set, analyze the learning feedback data set based on the deep learning effectiveness indicator evaluation results, and determine and output an online learning effectiveness report.

[0087] The learning feedback dataset can be a collection of learning experience information provided by students using electronic feedback tools, such as online questionnaires.

[0088] The online learning effectiveness report may be report information that includes various effectiveness indicators of students and effectiveness indicators that need to be focused on, and is used to reflect the online learning effectiveness of students.

[0089] Specifically, after obtaining the deep learning effectiveness indicator evaluation results that can objectively reflect the students' learning outcomes, although the students' learning outcomes can be accurately judged, it is difficult to clearly determine the direction in which the students need to improve in the future. The students' feedback information can reflect their subjective feelings during the learning process. These subjective feelings are based on the students' understanding of the learning content, the adaptability of the learning method, and the evaluation of the learning environment. These cognitive differences may lead to differences between their subjective evaluation of the learning effect and the actual learning outcomes. These differences are the directions that students need to improve in order to improve their learning outcomes. Therefore, by analyzing the deep learning effectiveness indicator evaluation results and the learning feedback data set, we find and extract the indicators that are contradictory due to the cognitive differences of the students. We use these indicators as the directions that need to be focused on in the subsequent learning process of the students. Combined with the data that can reflect the students' current learning outcomes, we use data visualization technology to integrate and generate the corresponding online learning effectiveness report, and provide the online learning effectiveness report to the students through human-computer interaction equipment, providing a scientific data reference for the students to improve their learning plans.

[0090] Through this scheme, the students' learning outcomes are analyzed from the weak subjective dimension and the strong objective dimension based on their online operation behavior dataset and physiological dataset, and the corresponding indicators reflecting the students' learning outcomes are quantified to construct a weak subjective dimension effectiveness indicator set and a strong objective dimension effectiveness indicator set. On this basis, the deep learning effectiveness evaluation results that fully reflect the students' deep learning outcomes are integrated, and combined with the learning feedback dataset, the online learning effectiveness report is determined and output. Through multi-dimensional analysis methods, the deep learning outcomes of the students in the online learning process are scientifically and comprehensively reflected. The online learning effectiveness report provides scientific suggestions for improving the students' subsequent learning plans, so as to guide students to find online learning methods that suit them and improve their learning quality.

[0091] In some embodiments, course content segmentation information is obtained, and the ideal learning time and normal interaction frequency are determined based on the course content segmentation information; the learning time and the ideal learning time are analyzed to determine the time impact index; the interaction frequency and the normal interaction frequency are analyzed to determine the interaction impact index; the number of pauses and the number of fast-forwards are analyzed to determine the pause impact index and the fast-forward impact index respectively; the weak subjective comprehensive evaluation index is determined based on the time impact index, the interaction impact index, the pause impact index and the fast-forward impact index; and a weak subjective dimension effectiveness indicator set is constructed based on the time impact index, the interaction impact index, the pause impact index, the fast-forward impact index and the weak subjective comprehensive evaluation index.

[0092] The online operation behavior dataset includes learning time, interaction frequency, number of pauses, and number of fast-forwards.

[0093] The learning time can be the time a student spends studying the current online course.

[0094] Interaction frequency can be the frequency with which students communicate and interact with teachers or other students while studying the current online course.

[0095] The number of pauses may be the number of pauses a student has experienced while watching the video material corresponding to the current online course.

[0096] The number of fast-forward times may be the number of fast-forward times a student has fast-forwarded while watching the corresponding video material of the current online course.

[0097] The course content segment information may be content segment information of the current online course studied by the student.

[0098] The ideal learning time can be the average time required for the current online course in the learning history of different students.

[0099] The normal interaction frequency may be the average interaction frequency corresponding to the current online course in the learning history of different students.

[0100] The duration impact index may be a quantitative value reflecting the impact of learning duration on learning outcomes.

[0101] The interaction impact index can be a quantitative value reflecting the impact of interaction frequency on learning outcomes.

[0102] The pause impact index may be a quantitative value reflecting the impact of the number of pauses on learning outcomes.

[0103] The fast-forward impact index may be a quantitative value reflecting the impact of the number of fast-forwards on learning outcomes.

[0104] The weak subjective comprehensive evaluation index can be a comprehensive evaluation value that reflects the learning outcomes of students under the weak subjective dimension.

[0105] Specifically, in the analysis of students' online operation behavior, the main focus is on learning time, interaction frequency, number of pauses and number of fast-forwards. Among them, learning time is usually directly related to the depth and breadth of learning. A longer learning time means that students have invested more time in a certain topic, which enhances their mastery and understanding of knowledge. However, if the learning time is too long but there is no substantial learning progress (such as lack of concentration), it will lead to fatigue and decreased learning results; interaction frequency reflects the degree of interaction between students and learning content, teachers and other students. A higher interaction frequency usually means that students actively participate in learning, which helps to better understand and master knowledge. However, if the interaction frequency is too high but lacks actual knowledge communication, it will This leads to superficial learning of the content, which has a negative impact on learning outcomes. The number of pauses can reflect the time and depth of students' thinking during the learning process. Too many pauses may indicate that students have great difficulties in certain knowledge points, which will affect the continuity of learning and lead to decreased learning outcomes. The number of fast-forwards usually indicates that students skip over certain content. Appropriate fast-forwarding can help students save time, quickly skip over mastered or repeated content, and concentrate on more challenging learning materials. However, if the number of fast-forwards is too many, it means that students lack interest in the course content or think that the content is not profound enough, which will affect the students' overall learning motivation and knowledge mastery, and thus have a negative impact on the students' learning outcomes.

[0106] Therefore, by analyzing the segmented information of course content, the learning time and interaction frequency of each student in different course segments are extracted, and the ideal learning time and normal interaction frequency that can have a positive impact on learning outcomes are determined. On this basis, the time impact index is determined according to the students' learning time and ideal learning time through mathematical analysis methods. At the same time, the interaction impact index is determined according to the students' interaction frequency and normal interaction frequency. Moreover, through mathematical analysis methods, the impact and mapping of pauses and fast-forwards on learning outcomes are quantified according to the number of pauses and fast-forwards, respectively, to obtain the pause impact index and fast-forward impact index. After combining the above four impact indices, further mathematical analysis is performed to determine the weak subjective comprehensive evaluation index that can comprehensively reflect the students' learning outcomes under the weak subjective dimension. After integrating the above four impact indices and the weak subjective comprehensive evaluation index, a weak subjective dimension effectiveness indicator set is constructed.

[0107] Through this plan, the four data dimensions of learning time, interaction frequency, number of pauses and number of fast-forwards in the students' operational behavior are analyzed respectively to analyze their impact on the students' learning level, and the time impact index, interaction impact index, pause impact index and fast-forward impact index are obtained. The impact indices corresponding to the four data dimensions are integrated to obtain a weak subjective comprehensive evaluation index that can comprehensively reflect the students' learning outcomes under the weak subjective dimension, and based on this, a weak subjective dimension effectiveness indicator set is constructed, which significantly improves the scientificity and comprehensiveness of the students' learning effectiveness evaluation process under the weak subjective dimension. At the same time, by incorporating multidimensional impact indicators into the weak subjective dimension effectiveness set, the subsequent deficiencies in the students' learning methods can be more accurately located.

[0108] In some embodiments, a weak subjective comprehensive evaluation index is determined based on the duration impact index, the interaction impact index, the pause impact index, and the fast-forward impact index, specifically the following formula (1):

[0109]

[0110] Among them, E is the weak subjective comprehensive evaluation index, k1 is the preset time influence coefficient, T is the learning time, α1 is the preset time influence index, T d For the ideal study time, is the duration impact index, k2 is the preset interaction impact coefficient, I is the interaction frequency, α2 is the preset interaction impact index, I d For normal interaction frequency, is the interaction influence index, k3 is the preset pause influence coefficient, P is the number of pauses, β is the preset pause influence index, k3·P β is the pause impact index, k4 is the preset fast-forward impact coefficient, α3 is the preset fast-forward impact index, F is the number of fast-forward times, is the fast-forward impact index.

[0111] The preset duration impact coefficient may be a preset value used to reflect the degree of linear impact of learning duration on learning outcomes. The preset duration impact coefficient may be obtained by performing linear fitting on historical learning data.

[0112] The preset duration impact index may be a preset value used to reflect the degree of nonlinear impact of learning duration on learning outcomes. The preset duration impact index may be obtained by performing polynomial regression on historical learning data.

[0113] The preset interaction influence coefficient may be a preset value used to reflect the linear influence of the interaction frequency on the learning outcome. The preset interaction influence coefficient may be obtained by performing linear fitting on historical learning data.

[0114] The preset interaction impact index may be a preset value used to reflect the nonlinear impact of interaction frequency on learning outcomes. The preset interaction impact index may be obtained by performing polynomial regression on historical learning data.

[0115] The preset pause influence coefficient may be a preset value used to reflect the linear influence of the number of pauses on learning outcomes, and the preset interaction influence coefficient may be obtained by performing linear fitting on historical learning data.

[0116] The preset pause impact index may be a preset value used to reflect the degree of nonlinear impact of the number of pauses on learning outcomes. The preset pause impact index may be obtained by performing polynomial regression on historical learning data.

[0117] The preset fast-forward influence coefficient may be a preset value used to reflect the linear influence of the number of fast-forward times on the learning outcome. The preset fast-forward influence coefficient may be obtained by performing linear fitting on historical learning data.

[0118] The preset fast-forward impact index can be a preset value used to reflect the degree of nonlinear influence of the number of fast-forwards on learning outcomes. The preset fast-forward impact index can be obtained by performing polynomial regression on historical learning data.

[0119] Specifically, through formula (1) It reflects the nonlinear relationship between the student's current learning time and their learning effect. As the learning time increases, the learning effect should gradually improve, but this improvement is not linear and will tend to stabilize after reaching a certain point. That is, when the learning time is close to the ideal learning time, the improvement of learning effect will gradually weaken. Similarly, through formula (1) Reflects the nonlinear relationship between the interaction frequency of students and their learning effect; through k3·P β Reflects the relationship between the number of pauses and learning outcomes. Since an increase in the number of pauses usually worsens learning outcomes, a power model is used here to describe the weakening effect of the number of pauses on learning outcomes. This function reflects the relationship between the number of fast-forwards and learning outcomes. Excessive fast-forwarding means that students lack interest in the learning content. An exponential function is used here to describe this relationship. Specifically, the shorter the number of fast-forwards, the less significant the improvement in learning outcomes.

[0120] Through this plan, mathematical analysis methods are used to clearly and accurately quantify the duration impact index, interaction impact index, pause impact index and fast-forward impact index according to the learning time, ideal learning time, interaction frequency, normal interaction frequency, number of pauses and number of fast-forwards through corresponding mathematical formulas. Based on this, several subjective comprehensive evaluation indexes are scientifically calculated to improve the accuracy and scientificity of the learning effectiveness evaluation under the weak subjective dimension, and provide clear quantitative data support for the subsequent improvement analysis of students' learning effectiveness.

[0121] In some embodiments, the focus change frequency analysis is performed on the eye movement data to determine the attention index; the emotion change analysis is performed on the expression change data to determine the emotion index; based on the attention index and the emotion index, the attention influence index and the emotion influence index are respectively determined, and the physiological effectiveness index is determined accordingly; based on the heart rate change data and the skin conductance change data, the average heart rate and the average skin conductivity are respectively determined; based on the average heart rate and the average skin conductivity, the heart rate influence index and the skin conductance influence index are respectively determined, and the deep physiological effectiveness index is determined accordingly; based on the attention influence index, the emotion influence index, the physiological effectiveness index, the heart rate influence index, the skin conductance influence index and the deep physiological effectiveness index, a strong objective dimension effectiveness index set is constructed.

[0122] The physiological data set includes eye movement data, facial expression change data, heart rate change data, and skin conductance change data.

[0123] Eye movement data can be data on changes in a student's eye gaze behavior during the learning process.

[0124] The expression change data may be the facial expression change data of the learner during the learning process.

[0125] The heart rate change data may be the heart rate change data of the student during the learning process.

[0126] The skin conductance change data may be skin conductance change data of a student during the learning process.

[0127] The attention index may be a value reflecting the degree of concentration of a student during the learning process.

[0128] The emotional index can be a value that reflects whether the student's emotional state tends to be positive or negative during the learning process.

[0129] The representative physiological effectiveness index may be a quantitative value obtained by analyzing the representative physiological data of the trainees, reflecting the degree of their influence on the learning effectiveness.

[0130] The average heart rate may be an average heart rate value of different students in the process of learning the current online course obtained from statistics of historical learning data.

[0131] The average skin conductance may be the average skin conductance of different students in the process of learning the current online course obtained from historical learning data.

[0132] The deep physiological effectiveness index may be a quantitative value obtained by analyzing the trainees' deep physiological data and reflecting the degree of its influence on learning effectiveness.

[0133] Specifically, changes in students' physiological data during the learning process will directly affect and reflect their learning outcomes. In the process of analyzing physiological data, the physiological data are divided into two dimensions. One is the representation dimension. The physiological data under this dimension reflects the students' direct and explicit physiological responses, such as changes in eye gaze area and facial expressions. The other is the deep dimension. The physiological data under this dimension covers the students' complex and non-explicit physiological mechanisms during the learning process, such as changes in heart rate and skin conductivity. Through analysis of these two dimensions, the mapping relationship between changes in students' physiological data and learning outcomes can be more comprehensively reflected.

[0134] Among them, the students' eye movement data reflects the changes in their attention during the learning process. Frequent eye movements indicate that students have difficulty concentrating or have difficulty understanding the course, which will directly have a negative impact on the students' learning outcomes. Changes in panel expressions can reflect the students' emotional state during the learning process, such as joy, confusion, and frustration. The more positive the students' emotions are, the better their learning outcomes. Heart rate changes reflect the students' psychological pressure and tension during the learning process. When students feel anxious, their information processing ability will decline significantly, which will have a negative impact on their learning outcomes. Skin conductivity can also reflect changes in students' psychological pressure. Therefore, the frequency of students' eye gaze changes reflected in the statisticians' eye movement data is used to determine the students' attention index. The higher the frequency of eye gaze changes, the lower the concentration, and the lower the corresponding attention index. Conversely, the higher the attention index, the lower the concentration. High; through the emotional classification model, such as support vector machine or random forest, the data of students' facial expression changes are analyzed to determine the main emotional types of students in the learning process. According to the emotional type, it is judged whether the students' emotions tend to be positive or negative, and the degree of tendency is determined. The higher the degree to which students' emotions tend to be positive, the higher the corresponding emotional index, and vice versa, the lower the emotional index; according to the attention index and the emotional index, the attention impact index and the emotional impact index are quantified through mathematical analysis; at the same time, based on the average heart rate and the average skin conductivity, the heart rate impact index and the skin conductance impact index are calculated using the mathematical analysis process, and the deep physiological effectiveness index is quantified; further, the attention impact index, emotional impact index, representative physiological effectiveness index, heart rate impact index, skin conductance impact index and deep physiological effectiveness index are included in the strong objective dimension effectiveness index set.

[0135] Through this plan, the students' physiological data are divided into representational dimensions and deep dimensions. Starting from the eye movement data, expression change data, heart rate change data and skin conductance change data corresponding to the two dimensions, the attention impact indicators, emotion impact indicators, heart rate impact indicators and skin conductance impact indicators that will directly affect learning outcomes are analyzed respectively. On this basis, the representational physiological effectiveness index and the deep physiological effectiveness index are determined respectively. Based on the above-mentioned several impact indicators and indices, a strong objective dimension effectiveness indicator set is constructed to improve the scientificity and comprehensiveness of the learning effectiveness evaluation under the strong objective dimension, and at the same time further improve the accuracy of the subsequent analysis of students' shortcomings.

[0136] In some embodiments, based on the attention index and the emotion index, an attention impact index and an emotion impact index are determined respectively, and the physiological effectiveness index is determined accordingly, specifically as follows:

[0137]

[0138] Among them, C is the index representing physiological effectiveness, k a is the preset attention influence coefficient, γ is the preset attention adjustment index, A is the attention index, is the attention impact indicator, k e is the preset emotion influence coefficient, δ is the preset emotion suppression parameter, ψ is the preset emotion enhancement parameter, S is the emotion index, k e ·(-δS 2 +ψS) is the emotional impact indicator.

[0139] The preset attention influence coefficient may be a numerical value reflecting the linear influence of the attention index on the learning outcome. The preset attention influence coefficient may be obtained by linear fitting of historical learning data.

[0140] The preset attention adjustment index may be a value reflecting the nonlinear effect of the attention index on learning outcomes. The preset attention adjustment index may be obtained by polynomial regression fitting of historical learning data.

[0141] The preset emotion suppression parameter may be a numerical value reflecting the degree of influence of negative emotions on learning outcomes. The preset emotion suppression parameter may be obtained by linear fitting of historical learning data.

[0142] The preset emotion enhancement parameter may be a numerical value reflecting the degree of influence of positive emotions on learning outcomes. The preset emotion enhancement parameter may be obtained by linear fitting of historical learning data.

[0143] Specifically, through formula (2) Describe the impact of student attention changes on their learning outcomes, and use an exponential function to reflect the nonlinear effect of attention on physiological outcomes, that is, an increase in attention may lead to an exponential improvement in physiological outcomes; at the same time, through k e ·(-δS 2 +ψS) describes the impact of students’ emotional changes on their learning outcomes, where -δS 2 Reflecting the asymmetric impact of negative emotions on learning outcomes, the negative impact of emotions increases with the increase of emotion intensity. Therefore, a negative quadratic term is introduced, and the degree of this negative impact is controlled by the emotion suppression parameter. The linear term ψS reflects the reinforcing effect of positive emotions on learning outcomes. The higher the emotion index, the greater the positive impact. Through the coordination of the above items, the physiological performance index is accurately calculated.

[0144] Through this plan, mathematical analysis methods are used to quantify the attention impact index and the emotion impact index based on the attention index and the emotion index respectively. On this basis, the physiological effectiveness index is calculated through a clear mathematical formula to achieve scientific quantification of the attention impact index, emotion impact index and physiological effectiveness index, and improve the accuracy of the analysis of learning effectiveness in the physiological dimension.

[0145] In some embodiments, the heart rate impact index and the skin conductance impact index are determined based on the average heart rate and the average skin conductance, respectively, and the deep physiological effectiveness index is determined accordingly, specifically as follows:

[0146]

[0147] Among them, D is the deep physiological effectiveness index, k H is the preset heart rate influence coefficient, λ is the preset heart rate influence adjustment parameter, H is the average heart rate, H t is the reference value of resting heart rate, is the heart rate impact index, k G is the preset skin conductance influence coefficient, G is the average skin conductivity, k G G is the skin conductance impact index.

[0148] The preset heart rate influence coefficient may be a numerical value reflecting the linear influence of heart rate on learning outcomes. The preset heart rate influence coefficient may be obtained by performing linear fitting on historical learning data.

[0149] The preset heart rate impact adjustment parameter may be a value reflecting the degree of nonlinear impact of heart rate on learning outcomes. The preset heart rate impact adjustment parameter may be obtained by performing polynomial regression fitting on historical learning data.

[0150] The calm heart rate reference value may be a heart rate reference value under normal learning pressure, and the calm heart rate reference value may be obtained by statistically analyzing heart rate change data of different students.

[0151] The preset skin conductance influence coefficient may be a numerical value reflecting the degree of linear influence of skin conductivity on learning outcomes. The preset skin conductance influence coefficient may be obtained by linear fitting of historical learning data.

[0152] Specifically, through formula (3) The deviation between the average heart rate of the students and the reference value of the resting heart rate is converted into an influence factor to indicate that when the average heart rate of the students is higher than the resting heart rate, the learning effect is inhibited; G G describes the relationship between skin conductivity and learning outcomes, and expresses the negative correlation between skin conductivity and learning outcomes. By combining the above items, the deep physiological effectiveness index is calculated using formula (3).

[0153] Through this solution, mathematical analysis methods are used to quantify the heart rate impact index and the skin conductance impact index based on the average heart rate and the average skin conductivity, respectively. On this basis, the deep physiological effectiveness index is accurately calculated through a clear mathematical formula, thereby achieving scientific quantification of the heart rate impact index, skin conductance impact index and deep physiological effectiveness index, and improving the accuracy of the analysis of learning effectiveness in the deep physiological dimension.

[0154] In some embodiments, a strong objective dimension performance indicator set and a weak subjective dimension performance indicator set are merged to determine a deep learning performance evaluation indicator set corresponding to each student; an indicator weight set is obtained, and based on the indicator weight set, a multi-user deep learning performance scoring matrix is ​​constructed according to the deep learning performance evaluation indicator set corresponding to each student; based on the multi-user deep learning performance scoring matrix, the deep learning performance score of each student is determined; the multi-user deep learning performance scoring matrix and the deep learning performance scores of all students are used as the deep learning performance evaluation results.

[0155] The deep learning effectiveness evaluation indicator set may be a set of all indicators that reflect the deep learning effectiveness of students.

[0156] The indicator weight set can be a collection of the weights of each evaluation indicator in the deep learning effectiveness evaluation indicator set in the process of measuring the deep learning effectiveness of students. The indicator weight set can be obtained through expert evaluation.

[0157] The multi-user deep learning effectiveness scoring matrix can be a mathematical matrix with a two-dimensional array structure, which integrates all learning effectiveness evaluation indicators and corresponding weights corresponding to different students.

[0158] The deep learning effectiveness score may be a numerical score reflecting the comprehensive learning effectiveness of the students. The higher the deep learning effectiveness score, the better the learning effectiveness of the corresponding students.

[0159] Specifically, through set splicing, the strong objective dimension effectiveness indicator set and the weak subjective dimension effectiveness indicator set are merged, and the relative order of different indicators in the original sets is maintained during the merging process to obtain the deep learning effectiveness evaluation indicator set. Through mathematical analysis, according to the deep learning effectiveness evaluation indicator set and indicator weight set corresponding to each student, a multi-user deep learning effectiveness scoring matrix is ​​constructed. Based on the multi-user deep learning effectiveness scoring matrix, a one-step mathematical analysis process is carried out to quantify the students' deep learning effectiveness scores, so as to obtain scientific and accurate deep learning effectiveness evaluation results for each student.

[0160] Through this solution, the strong objective dimension effectiveness indicator set and the weak subjective dimension effectiveness indicator set are merged to obtain the deep learning effectiveness evaluation indicator set. Combined with the indicator weight set, a scientific evaluation framework for students' deep learning effectiveness is constructed in the form of a mathematical matrix. The multi-user deep learning effectiveness scoring matrix and the deep learning effectiveness scores of all students are used as the deep learning effectiveness evaluation results. While improving the accuracy of learning effectiveness evaluation, it provides an important data framework for subsequent analysis of the direction of improvement of students' learning effectiveness.

[0161] In some embodiments, based on the indicator weight set and the deep learning effectiveness evaluation indicator set corresponding to each student, a multi-user deep learning effectiveness scoring matrix is ​​constructed, specifically the following matrix expression (4):

[0162]

[0163] Among them, I is the multi-user deep learning performance rating matrix, Y mn is the nth evaluation indicator in the deep learning effectiveness evaluation indicator set of the mth student, W n is the weight value corresponding to the nth evaluation indicator in the indicator weight set; according to the multi-user deep learning effectiveness scoring matrix, the deep learning effectiveness score of each student is determined, specifically as follows:

[0164]

[0165] Among them, RI i is the score of the deep learning effectiveness of the i-th student, n is the total number of evaluation indicators, I j It is the effectiveness score of the current student corresponding to the jth evaluation indicator in the multi-user deep learning effectiveness scoring matrix.

[0166] Specifically, through the matrix structure described by matrix expression (4), the deep learning effectiveness evaluation indicator set and indicator weight set corresponding to each student are automatically integrated to construct a multi-user deep learning effectiveness scoring matrix, and through formula (5), the corresponding products of the evaluation indicators and weights corresponding to each student in the matrix are automatically summed to accurately obtain the deep learning effectiveness score corresponding to each student under the multi-user deep learning effectiveness scoring matrix.

[0167] Through this solution, mathematical analysis methods are used to clarify the multi-user deep learning effectiveness scoring matrix through mathematical expressions, and mathematical formulas are used to accurately calculate the deep learning effectiveness score of each student based on the multi-user deep learning effectiveness scoring matrix, further improving the scientificity and accuracy of the learning effectiveness evaluation process.

[0168] In some embodiments, the learning feedback data set is analyzed to determine a multi-type feedback indicator information set corresponding to each student; based on the multi-type feedback indicator information set, the corresponding deep learning effectiveness evaluation indicator set in the multi-user deep learning effectiveness scoring matrix is ​​extracted; based on the multi-type feedback indicator information set, the deep learning effectiveness evaluation indicator set is analyzed to determine a number of contradictory feedback indicators; based on each student's deep learning effectiveness score and the number of contradictory feedback indicators, an online learning effectiveness report is constructed and output.

[0169] The multi-type feedback indicator information set may be a set of feedback information given by trainees in response to different types of questionnaire questions.

[0170] Contradictory feedback indicators can be indicators in the feedback information given by students based on their self-cognition that contradict the objective evaluation indicators in the deep learning effectiveness evaluation indicator set.

[0171] Specifically, after completing a scientific assessment of students' learning outcomes, in order to further improve their online learning outcomes, students can provide feedback on their self-perceptions during the learning process through electronic feedback tools, such as questionnaires. However, students' self-perceptions are often highly subjective and easily conflict with the objective effectiveness evaluation indicators in the collected deep learning effectiveness evaluation indicator set. This indicates that students have self-cognitive biases in the learning process, and it is necessary to focus on students' cognitive biases to improve their subsequent learning outcomes. In the process of determining several contradictory feedback indicators, data standardization techniques, such as Z-score standardization, are used to standardize the indicator information in the multi-type feedback indicator information set and the deep learning effectiveness evaluation indicator set to make different indicators comparable. Then, correlation quantification algorithms, such as the Pearson correlation coefficient, are used to quantify the correlations between different indicators. Several feedback indicators with strong correlations and information contradictions are regarded as contradictory feedback indicators. Data visualization technology is used to integrate and visualize each student's deep learning effectiveness score and several contradictory feedback indicators to construct and output an online learning effectiveness report.

[0172] Through this plan, on the basis of completing the evaluation of students' learning outcomes, we further combine the students' feedback data, extract a number of contradictory feedback indicators based on the multi-type feedback indicator information set and the deep learning effectiveness evaluation indicator set, and construct and output an online learning effectiveness report based on each student's deep learning effectiveness score and a number of contradictory feedback indicators to clarify the areas that students need to focus on in order to improve their learning outcomes in the future, and provide a significant positive effect on improving the quality of students' learning.

[0173] Figure 3 This is a structural diagram of an online learning effectiveness evaluation system provided by an embodiment of the present application, such as Figure 3As shown, an online learning effectiveness evaluation system 300 of this embodiment includes a weak subjective analysis module 301 , a strong objective analysis module 302 , a deep analysis module 303 and an output module 304 .

[0174] A weak subjective analysis module 301 is used to obtain an online operation behavior dataset, analyze the online operation behavior dataset, and determine a weak subjective dimension performance indicator set;

[0175] A strong objective analysis module 302 is used to obtain a physiological data set, analyze the physiological data set, and determine a strong objective dimension performance indicator set;

[0176] A deep analysis module 303 is configured to analyze the weak subjective dimension effectiveness indicator set based on the strong objective dimension effectiveness indicator set to determine a deep learning effectiveness evaluation result;

[0177] The output module 304 is configured to obtain a learning feedback data set, analyze the learning feedback data set based on the deep learning effectiveness indicator evaluation results, and determine and output an online learning effectiveness report.

[0178] Optionally, the weak subjective analysis module 301 is specifically configured to:

[0179] Obtain course content segmentation information, and determine the ideal learning time and normal interaction frequency based on the course content segmentation information; analyze the learning time and the ideal learning time to determine the time impact index; analyze the interaction frequency and the normal interaction frequency to determine the interaction impact index; analyze the number of pauses and the number of fast-forwards to determine the pause impact index and the fast-forward impact index respectively; determine a weak subjective comprehensive evaluation index based on the time impact index, the interaction impact index, the pause impact index, and the fast-forward impact index; construct the weak subjective dimension effectiveness indicator set based on the time impact index, the interaction impact index, the pause impact index, the fast-forward impact index, and the weak subjective comprehensive evaluation index.

[0180] Optionally, when the weak subjective analysis module 301 determines the weak subjective comprehensive evaluation index based on the duration impact index, the interaction impact index, the pause impact index, and the fast-forward impact index, the formula is as follows:

[0181] Among them, E is the weak subjective comprehensive evaluation index, k1 is the preset time influence coefficient, T is the learning time, α1 is the preset time influence index, T d For the ideal study time, is the duration impact index, k2 is the preset interaction impact coefficient, I is the interaction frequency, α2 is the preset interaction impact index, I d For normal interaction frequency, is the interaction influence index, k3 is the preset pause influence coefficient, P is the number of pauses, β is the preset pause influence index, k3·P β is the pause impact index, k4 is the preset fast-forward impact coefficient, α3 is the preset fast-forward impact index, F is the fast-forward times, is the fast-forward impact index.

[0182] Optionally, the strong objective analysis module 302 is specifically configured to:

[0183] Perform focus change frequency analysis on the eye movement data to determine the attention index; perform emotion change analysis on the expression change data to determine the emotion index; determine the attention influence index and the emotion influence index respectively based on the attention index and the emotion index, and determine the physiological effectiveness index accordingly; determine the average heart rate and the average skin conductivity respectively based on the heart rate change data and the skin conductance change data; determine the heart rate influence index and the skin conductance influence index respectively based on the average heart rate and the average skin conductance, and determine the deep physiological effectiveness index accordingly; construct the strong objective dimension effectiveness index set based on the attention influence index, the emotion influence index, the physiological effectiveness index, the heart rate influence index, the skin conductance influence index and the deep physiological effectiveness index.

[0184] Optionally, when the strong objective analysis module 302 determines the attention impact index and the emotion impact index according to the attention index and the emotion index, and determines the physiological effectiveness index accordingly, the formula is specifically as follows:

[0185] Wherein, C is the physiological efficacy index, k a is the preset attention influence coefficient, γ is the preset attention adjustment index, A is the attention index, is the attention impact index, k e is the preset emotion influence coefficient, δ is the preset emotion suppression parameter, ψ is the preset emotion enhancement parameter, S is the emotion index, k e ·(-δS 2 +ψS) is the emotion impact index.

[0186] Optionally, the strong objective analysis module 302 determines the heart rate impact index and the skin conductance impact index based on the average heart rate and the average skin conductance, and determines the deep physiological effectiveness index based on the heart rate impact index and the skin conductance impact index, specifically using the following formula:

[0187] Where D is the deep physiological efficacy index, k His the preset heart rate influence coefficient, λ is the preset heart rate influence adjustment parameter, H is the average heart rate, H t is the reference value of resting heart rate, is the heart rate impact index, k G is the preset skin conductance influence coefficient, G is the average skin conductivity, k G G is the skin conductance impact index.

[0188] Optionally, the deep analysis module 303 is specifically configured to:

[0189] The strong objective dimension performance indicator set and the weak subjective dimension performance indicator set are merged to determine the deep learning performance evaluation indicator set corresponding to each student; an indicator weight set is obtained, and based on the indicator weight set and according to the deep learning performance evaluation indicator set corresponding to each student, a multi-user deep learning performance scoring matrix is ​​constructed; based on the multi-user deep learning performance scoring matrix, the deep learning performance score of each student is determined; the multi-user deep learning performance scoring matrix and the deep learning performance scores of all students are used as the deep learning performance evaluation results.

[0190] Optionally, when the deep analysis module 303 constructs a multi-user deep learning effectiveness scoring matrix based on the indicator weight set and the deep learning effectiveness evaluation indicator set corresponding to each student, the matrix expression is specifically as follows:

[0191] Where I is the multi-user deep learning performance scoring matrix, Y mn is the nth evaluation indicator in the deep learning effectiveness evaluation indicator set of the mth student, W n is the weight value corresponding to the nth evaluation indicator in the indicator weight set; the deep learning effectiveness score of each student is determined according to the multi-user deep learning effectiveness scoring matrix, specifically the following formula:

[0192] Among them, RI i Score the deep learning effectiveness of the i-th student, n is the total number of evaluation indicators, I j The effectiveness score of the current student corresponding to the jth evaluation indicator in the multi-user deep learning effectiveness scoring matrix.

[0193] Optionally, the output module 304 is specifically configured to:

[0194] Analyze the learning feedback data set to determine a multi-type feedback indicator information set corresponding to each student; extract the corresponding deep learning effectiveness evaluation indicator set in the multi-user deep learning effectiveness scoring matrix based on the multi-type feedback indicator information set; analyze the deep learning effectiveness evaluation indicator set based on the multi-type feedback indicator information set to determine a number of contradictory feedback indicators; construct and output the online learning effectiveness report based on the deep learning effectiveness score of each student and the several contradictory feedback indicators.

[0195] The system of this embodiment can be used to execute the method of any of the above embodiments. Its implementation principles and technical effects are similar and will not be described in detail here.

Claims

1. A method for evaluating the effectiveness of online learning, characterized in that: include: Acquire an online operation behavior dataset, analyze the online operation behavior dataset, and determine a weak subjective dimension performance indicator set; Acquiring a physiological data set, analyzing the physiological data set, and determining a set of strong objective dimension performance indicators; Based on the strong objective dimension effectiveness indicator set, the weak subjective dimension effectiveness indicator set is analyzed to determine the deep learning effectiveness evaluation result; Obtaining a learning feedback data set, analyzing the learning feedback data set based on the deep learning effectiveness indicator evaluation results, and determining and outputting an online learning effectiveness report; The online operation behavior dataset includes learning time, interaction frequency, pause times, and fast-forward times. The analysis of the online operation behavior dataset to determine a weak subjective dimension performance indicator set includes: Obtaining course content segmentation information, and determining an ideal learning duration and a normal interaction frequency based on the course content segmentation information; Analyze the learning time and the ideal learning time to determine a time impact index; Analyze the interaction frequency and the normal interaction frequency to determine an interaction impact index; Analyzing the number of pauses and the number of fast-forwards to determine a pause impact index and a fast-forward impact index, respectively; Determining a weak subjective comprehensive evaluation index according to the duration impact index, the interaction impact index, the pause impact index, and the fast-forward impact index; Constructing the weak subjective dimension effectiveness indicator set according to the duration impact index, the interaction impact index, the pause impact index, the fast-forward impact index, and the weak subjective comprehensive evaluation index; The weak subjective comprehensive evaluation index is determined according to the duration impact index, the interaction impact index, the pause impact index, and the fast-forward impact index, specifically as follows: Among them, E is the weak subjective comprehensive evaluation index, k1 is the preset time influence coefficient, T is the learning time, α1 is the preset time influence index, T d For the ideal study time, is the duration impact index, k2 is the preset interaction impact coefficient, I is the interaction frequency, α2 is the preset interaction impact index, I d For normal interaction frequency, is the interaction influence index, k3 is the preset pause influence coefficient, P is the number of pauses, β is the preset pause influence index, k3·P β is the pause impact index, k4 is the preset fast-forward impact coefficient, α3 is the preset fast-forward impact index, F is the fast-forward times, is the fast-forward impact index.

2. The method according to claim 1, characterized in that The physiological data set includes eye movement data, facial expression change data, heart rate change data, and skin conductance change data. The analysis of the physiological data set to determine a strong objective dimension performance indicator set includes: performing focus change frequency analysis on the eye movement data to determine an attention index; Performing emotion change analysis on the expression change data to determine an emotion index; Determining an attention impact index and an emotion impact index based on the attention index and the emotion index, and determining an index representing physiological effectiveness based on the index; determining an average heart rate and an average skin conductance according to the heart rate variation data and the skin conductance variation data; Determining a heart rate impact index and a skin conductance impact index based on the average heart rate and the average skin conductance, and determining a deep physiological effectiveness index accordingly; The strong objective dimension performance indicator set is constructed based on the attention impact index, the emotion impact index, the physiological performance index, the heart rate impact index, the skin conductance impact index and the deep physiological performance index.

3. The method according to claim 2, characterized in that According to the attention index and the emotion index, an attention impact index and an emotion impact index are determined respectively, and the physiological effectiveness index is determined accordingly, which is specifically the following formula: C=k a ·(e γA -1)+k e ·(-δS 2 +ψS); Wherein, C is the physiological efficacy index, k a is the preset attention influence coefficient, γ is the preset attention adjustment index, A is the attention index, k a ·(e γA -1) is the attention impact index, k e is the preset emotion influence coefficient, δ is the preset emotion suppression parameter, ψ is the preset emotion enhancement parameter, S is the emotion index, k e ·(-δS 2 +ψS) is the emotion impact index.

4. The method according to claim 2, characterized in that The heart rate impact index and the skin conductance impact index are determined based on the average heart rate and the average skin conductance, and the deep physiological effectiveness index is determined based on the above. Specifically, the formula is as follows: Where D is the deep physiological efficacy index, k H is the preset heart rate influence coefficient, λ is the preset heart rate influence adjustment parameter, H is the average heart rate, H t is the reference value of resting heart rate, is the heart rate impact index, k G is the preset skin conductance influence coefficient, G is the average skin conductivity, k G G is the skin conductance impact index.

5. The method according to claim 2, characterized in that The step of analyzing the weak subjective dimension effectiveness indicator set based on the strong objective dimension effectiveness indicator set to determine the deep learning effectiveness evaluation result includes: Merging the strong objective dimension performance indicator set with the weak subjective dimension performance indicator set to determine a deep learning performance evaluation indicator set corresponding to each student; Obtaining an indicator weight set, and constructing a multi-user deep learning effectiveness scoring matrix based on the indicator weight set and the deep learning effectiveness evaluation indicator set corresponding to each student; Determining a deep learning effectiveness score for each student based on the multi-user deep learning effectiveness scoring matrix; The multi-user deep learning effectiveness rating matrix and the deep learning effectiveness ratings of all students are used as the deep learning effectiveness evaluation results.

6. The method according to claim 5, characterized in that Based on the indicator weight set and according to the deep learning effectiveness evaluation indicator set corresponding to each student, a multi-user deep learning effectiveness scoring matrix is ​​constructed, which is specifically the following matrix expression: Where I is the multi-user deep learning performance scoring matrix, Y mn is the nth evaluation indicator in the deep learning effectiveness evaluation indicator set of the mth student, W n is the weight value corresponding to the nth evaluation indicator in the indicator weight set; The deep learning effectiveness score of each student is determined according to the multi-user deep learning effectiveness score matrix, specifically as follows: Among them, RI i Score the deep learning effectiveness of the i-th student, n is the total number of evaluation indicators, I j The effectiveness score of the current student corresponding to the jth evaluation indicator in the multi-user deep learning effectiveness scoring matrix.

7. The method according to claim 6, characterized in that The step of analyzing the learning feedback dataset based on the deep learning effectiveness indicator evaluation results to determine and output an online learning effectiveness report includes: Analyze the learning feedback data set to determine a multi-type feedback indicator information set corresponding to each student; Extracting the corresponding deep learning effectiveness evaluation indicator set in the multi-user deep learning effectiveness scoring matrix according to the multi-type feedback indicator information set; Analyzing the deep learning effectiveness evaluation indicator set based on the multi-type feedback indicator information set to determine a number of conflicting feedback indicators; The online learning effectiveness report is constructed and outputted based on the deep learning effectiveness score of each student and a plurality of the conflicting feedback indicators.

8. An online learning effectiveness evaluation system, characterized in that: include: A weak subjective analysis module is used to obtain an online operation behavior data set, analyze the online operation behavior data set, and determine a weak subjective dimension performance indicator set; A strong objective analysis module, configured to obtain a physiological data set, analyze the physiological data set, and determine a set of strong objective dimension performance indicators; A deep analysis module, configured to analyze the weak subjective dimension effectiveness indicator set based on the strong objective dimension effectiveness indicator set to determine a deep learning effectiveness evaluation result; An output module is used to obtain a learning feedback data set, analyze the learning feedback data set based on the deep learning effectiveness indicator evaluation results, and determine and output an online learning effectiveness report; The online operation behavior dataset includes learning time, interaction frequency, pause times, and fast-forward times; The weak subjective analysis module is specifically used to: Obtaining course content segmentation information, and determining an ideal learning duration and a normal interaction frequency based on the course content segmentation information; Analyze the learning time and the ideal learning time to determine a time impact index; Analyze the interaction frequency and the normal interaction frequency to determine an interaction impact index; Analyzing the number of pauses and the number of fast-forwards to determine a pause impact index and a fast-forward impact index, respectively; Determining a weak subjective comprehensive evaluation index according to the duration impact index, the interaction impact index, the pause impact index, and the fast-forward impact index; Constructing the weak subjective dimension effectiveness indicator set according to the duration impact index, the interaction impact index, the pause impact index, the fast-forward impact index, and the weak subjective comprehensive evaluation index; When the weak subjective analysis module determines the weak subjective comprehensive evaluation index based on the duration impact index, the interaction impact index, the pause impact index, and the fast-forward impact index, the weak subjective comprehensive evaluation index is specifically the following formula: Among them, E is the weak subjective comprehensive evaluation index, k1 is the preset time influence coefficient, T is the learning time, α1 is the preset time influence index, T d For the ideal study time, is the duration impact index, k2 is the preset interaction impact coefficient, I is the interaction frequency, α2 is the preset interaction impact index, I d For normal interaction frequency, is the interaction influence index, k3 is the preset pause influence coefficient, P is the number of pauses, β is the preset pause influence index, k3·P β is the pause impact index, k4 is the preset fast-forward impact coefficient, α3 is the preset fast-forward impact index, F is the fast-forward times, is the fast-forward impact index.

Images