Teaching Quality Analysis Method and Related Equipment

Abstract

An embodiment of the present invention provides a teaching quality analysis method. According to the campus activity file of a target student, target activity characteristics of the target student within a first preset time period are extracted; historical activity characteristics and historical evaluation attributes corresponding to each student within a second preset time period are extracted from historical campus activity files to obtain a historical activity characteristic set and a historical evaluation attribute set; according to the target activity characteristics, the historical activity characteristic set and the historical evaluation attribute set, the evaluation attribute of the target student within a third preset time is predicted; based on the evaluation attribute of the target student, the teaching quality of the school where the target student is located within the third preset time is analyzed. The present invention improves the prediction accuracy of the evaluation attribute, can predict the evaluation attribute of students at any current or future time period, so as to accurately evaluate the current or future teaching quality and provide continuous and real-time reference for teaching quality improvement.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular, to a method for analyzing teaching quality and related devices. Background Art

[0002] With the continuous development of the education cause, the education work of students has received more and more attention from parents and educators. How to improve teaching quality is a common concern in education work, and the quality of teaching is also an important indicator to measure the teaching strength of schools. At present, the evaluation of teaching quality mainly relies on quantifiable assessment indicators, such as students' exam scores. However, exam scores are limited by the assessment cycle, which is relatively long, and the assessment results are unstable. Moreover, exam scores are only an assessment indicator of historical teaching quality, which leads to the lag in improving teaching quality. Therefore, the existing teaching quality analysis methods cannot evaluate the current teaching quality. Summary of the Invention

[0003] Embodiments of the present invention provide a method for analyzing teaching quality and related devices, which improve the prediction accuracy of evaluation attributes, can predict the evaluation attributes of students at any current or future time period, so as to accurately evaluate the current or future teaching quality and provide continuous and real-time reference for improving teaching quality.

[0004] In a first aspect, embodiments of the present invention provide a method for analyzing teaching quality, the method comprising:

[0005] Extracting target activity characteristics of a target student within a first preset time period according to the campus activity file of the target student;

[0006] Extracting historical activity characteristics and historical evaluation attributes corresponding to each student within a second preset time period from historical campus activity files to obtain a historical activity characteristic set and a historical evaluation attribute set;

[0007] Predicting the evaluation attributes of the target student within a third preset time according to the target activity characteristics, the historical activity characteristic set and the historical evaluation attribute set;

[0008] Analyzing the teaching quality of the school where the target student is located within a third preset time based on the evaluation attributes of the target student.

[0009] Optionally, the predicting the evaluation attributes of the target student within a third preset time according to the target activity characteristics, the historical activity characteristic set and the historical evaluation attribute set includes:

[0010] Calculating the distances between the target activity characteristics and each of the historical activity characteristics in the historical activity characteristic set, and determining K historical activity characteristics that are closest to the target activity characteristics;

[0011] Based on the K historical activity features, determine K historical evaluation attributes corresponding to the K historical activity features from the historical evaluation attribute set;

[0012] Determine the evaluation attributes of the target student within the third preset time according to the K historical evaluation attributes.

[0013] Optionally, the determining the evaluation attributes of the target student within the third preset time according to the K historical evaluation attributes includes:

[0014] Calculate the occurrence frequency of each type of historical evaluation attribute among the K historical evaluation attributes;

[0015] Determine the evaluation attributes of the target student within the third preset time according to the historical evaluation attribute corresponding to the type with the highest occurrence frequency.

[0016] Optionally, before the step of calculating the distance between the target activity feature and the historical activity features and determining the K historical activity features with the closest distance to the target activity feature, the method further includes:

[0017] Construct a first data set, where the first data set includes first sample activity features and first evaluation attribute labels corresponding to the first sample activity features;

[0018] Divide the first data set into a training data set and a validation data set;

[0019] Initialize a K value, and train and adjust the K value through the training data set to obtain an adjusted K value;

[0020] Perform a validation process on the adjusted K value through the validation data set to determine the final K value.

[0021] Optionally, before the step of analyzing the teaching quality of the school where the target student is located within the third preset time based on the evaluation attributes of the target student, the method further includes:

[0022] Input the current target activity feature and the current evaluation attribute into a preset confidence network to calculate the confidence of the current evaluation attribute;

[0023] If the confidence meets the preset condition, add the current target activity feature to the historical activity feature set and add the current evaluation attribute to the historical evaluation attribute set;

[0024] If the confidence level does not meet the preset condition, add the current target activity feature to the secondary prediction queue according to the preset rule for waiting. The secondary prediction queue is used to recalculate the target activity features in the secondary prediction queue with the historical activity feature set after all target activity features are processed.

[0025] Optionally, both the target activity feature and the historical activity feature are obtained by splicing each regional activity feature according to a preset rule. A regional activity feature corresponds to a campus area. Before the step of inputting the current target activity feature and the current evaluation attribute into the preset confidence network to calculate the confidence level of the current evaluation attribute, the method further includes:

[0026] Construct a confidence network and a second data set. The confidence network includes weight parameters and bias parameters corresponding to the regional activity features, and the second data set includes second sample activity features and second evaluation attribute labels corresponding to the second sample activity features;

[0027] Train the confidence network through the second data set, and make the confidence network converge by continuously adjusting the weight parameters and the bias parameters to obtain a trained confidence network, and use the trained confidence network as the preset confidence network.

[0028] Optionally, the teaching quality analysis based on the evaluation attributes of the target student includes:

[0029] Evaluate the campus environment according to the weight parameters in the trained confidence network to obtain a campus environment evaluation result;

[0030] Conduct teaching quality analysis based on the campus environment evaluation result and the evaluation attributes of the target student.

[0031] In a second aspect, an embodiment of the present invention provides a teaching quality analysis device, and the device includes:

[0032] A first extraction module, configured to extract the target activity features of the target student within a first preset time period according to the campus activity file of the target student;

[0033] A second extraction module, configured to extract the historical activity features and historical evaluation attributes corresponding to each student within a second preset time period from the historical campus activity file to obtain a historical activity feature set and a historical evaluation attribute set;

[0034] A prediction module, configured to predict the evaluation attributes of the target student within a third preset time according to the target activity features, the historical activity feature set, and the historical evaluation attribute set;

[0035] An analysis module for analyzing the teaching quality of the school where the target student is located within a third preset time based on the assessment attributes of the target student.

[0036] In a third aspect, an embodiment of the present invention provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the steps in the teaching quality analysis method provided by the embodiment of the present invention are implemented.

[0037] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in the teaching quality analysis method provided by the embodiment of the invention are implemented.

[0038] In the embodiment of the present invention, according to the campus activity file of the target student, the target activity characteristics of the target student within a first preset time period are extracted; in the historical campus activity file, the historical activity characteristics and historical assessment attributes corresponding to each student within a second preset time period are extracted to obtain a historical activity characteristic set and a historical assessment attribute set; according to the target activity characteristics, the historical activity characteristic set and the historical assessment attribute set, the assessment attributes of the target student within a third preset time are predicted; based on the assessment attributes of the target student, the teaching quality of the school where the target student is located within a third preset time is analyzed. By creating files for students' campus activities, extracting the target activity characteristics of students within a first preset time period from the students' campus activity files, and extracting the historical activity characteristics and historical assessment attributes of each student within a second preset time period from the historical campus activity files, using the target activity characteristics, historical activity characteristics, and historical assessment attributes to predict the assessment attributes of students, and conducting teaching quality analysis based on the predicted assessment attributes, without waiting for a long assessment cycle, using the potential common connection between students' campus activities and assessment attributes to predict assessment attributes, improving the prediction accuracy of assessment attributes, and being able to predict the assessment attributes of students at any current or future time period, so as to accurately evaluate the current or future teaching quality and provide continuous and real-time reference for teaching quality improvement. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0040] Figure 1 is a flowchart of a teaching quality analysis method provided by an embodiment of the present invention;

[0041] Figure 2 It is a schematic structural diagram of a teaching quality analysis device provided by an embodiment of the present invention;

[0042] Figure 3 It is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. Detailed implementation manners

[0043] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0044] Please refer to Figure 1 , Figure 1 which is a flowchart of a teaching quality analysis method provided by an embodiment of the present invention. As Figure 1 shown, the teaching quality analysis method includes the following steps:

[0045] 101. Extract the target activity characteristics of the target student within the first preset time period according to the campus activity file of the target student.

[0046] In the embodiment of the present invention, the above-mentioned campus activity file refers to a data set recorded when students generate activities on campus. The above-mentioned campus activity file can also be understood as a daily life file on campus centered around students.

[0047] Cameras with face recognition functions can be installed in preset campus areas, and each campus area can be captured and monitored by these cameras with face recognition functions, so as to obtain the activity data of each student in each campus area. The activity data of each student in each campus area and the corresponding evaluation attributes are filed as campus activity files. Further, if the assessment has not been carried out, only the activity data of each student in each campus area needs to be filed into the corresponding campus activity file. If the assessment has been carried out, the action data of each student in each campus area before the assessment and the corresponding evaluation attributes are filed as historical campus activity files.

[0048] The above-mentioned campus areas refer to areas such as classrooms, corridors, stairs, libraries, gymnasiums, various campus sports venues (such as badminton courts, basketball courts, football fields, track and field fields, etc.), offices, canteens, small stores, and campus streets on campus.

[0049] The above-mentioned campus activity files can be divided by semester, academic year, or assessment period. For example, at the beginning of each semester, a campus activity file for the current semester can be re-maintained for students and updated in real time according to the captured activity data, and the campus activity file corresponding to the previous semester can be maintained as a historical campus file. It is also possible to re-maintain a campus activity file for the current academic year for students at the beginning of each academic year and update it in real time according to the captured activity data, and maintain the campus activity file corresponding to the previous academic year as a historical campus file. It is also possible to maintain a campus activity file for the current assessment period for students after the end of the previous assessment period, update it in real time according to the captured activity data, and maintain the campus activity file corresponding to the previous assessment period as a historical campus file.

[0050] The data collection dimensions in the campus activity file include activity data and basic file data. Among them, the activity data can be collected and recognized by the above-mentioned cameras with face recognition function, and the above-mentioned basic file data can be collected and sorted from the registration materials. The activity data can include face capture attributes, face capture locations, face capture times, etc., and the basic file data can include student class data, learning data, class course arrangement data, work and rest data, and holiday time arrangement data, etc.

[0051] The above-mentioned target students refer to the students for whom the evaluation attribute prediction needs to be carried out. Specifically, the above-mentioned target students can be the captured students. When the number of times a student is captured in a campus area reaches the preset value, the evaluation attribute prediction can be carried out for the student, and the student is the target student. It is also possible to carry out the evaluation attribute prediction for all students after receiving the user's prediction instruction. The above-mentioned evaluation attributes can be grades, personalities, etc.

[0052] The above-mentioned first preset time period can be a time period with a time unit of week, month, assessment period, quarter, semester, academic year. For example, the first preset time period can be the first quarter of the second grade in the current academic year. Then, the activity data of the target students in the first quarter of the second grade in the current academic year can be extracted from the campus activity files of the target students, and the target activity features can be extracted from the activity data. The activity data can be as shown in Table 1 below:

[0053] Table 1

[0054]

[0055] Among them, Table 2 shows the activity data and evaluation attributes of the first quarter of the second grade in the current academic year. In the activity data in Table 1, the number of times each student is captured in different campus areas is recorded, and the grades and personalities are the corresponding evaluation attributes, which are unknown because no assessment has been carried out.

[0056] The above-mentioned target activity features can be features in vector form. For example, the target activity features corresponding to Chen Dan can be (468, 220, 322, 10, 34, 16, …).

[0057] 102. Extract the historical activity features and historical evaluation attributes corresponding to each student within the second preset time period from the historical campus activity archives to obtain a historical activity feature set and a historical evaluation attribute set.

[0058] In the embodiment of the present invention, the historical campus activity archives include the historical activity data and historical evaluation attributes of each student.

[0059] The above-mentioned second preset time period can be determined according to the first preset time period and the prediction target. For example, if the first preset time period is the first quarter of the second grade and the prediction target is the evaluation data of the target student in the third grade, then the second preset time period can be the first quarter of the third grade in each historical school year. At this time, the historical activity data and historical evaluation attributes corresponding to each student within the first quarter of the third grade can be extracted from the historical campus activity archives, so as to extract the corresponding historical activity features from the historical activity data and construct the corresponding historical activity feature set and historical evaluation attribute set according to the historical activity features of each student. The above-mentioned historical activity data and historical evaluation attributes can be as shown in Table 2 below:

[0060] Table 2

[0061]

[0062] Among them, Table 2 shows the historical activity data and historical evaluation attributes in the first quarter of the third grade in each historical school year. In the historical activity data in Table 2, the number of times each student is captured in different campus areas is recorded, and the grades and personalities are the corresponding evaluation attributes. Since the assessment has been carried out, they are known.

[0063] The above-mentioned target activity features can be features in vector form. For example, the target activity features corresponding to Huang Dan can be (300, 120, 222, 50, 44, 6, …), and the historical evaluation attribute is grade A and the personality is introverted.

[0064] 103. Predict the evaluation attributes of the target student within the third preset time according to the target activity features, the historical activity feature set and the historical evaluation attribute set.

[0065] In the embodiment of the present invention, the above-mentioned third preset time can be determined according to the prediction target. For example, if the prediction target is the evaluation attributes of the target student in the next school year, and the first preset time is the first quarter of the current school year, then the third preset time can be the evaluation attributes in the first quarter of the next school year.

[0066] Specifically, the similarity between the target activity features and the historical activity features can be utilized to predict the evaluation attributes of the target student. That is, the more similar the activity data of the target student is to a certain historical activity data, the greater the probability that their evaluation attributes are the same.

[0067] Optionally, the distances between the above-mentioned target activity features and each of the above-mentioned historical activity features in the above-mentioned historical activity feature set can be calculated, and the K historical activity features with the closest distance to the above-mentioned target activity features can be determined; based on the above-mentioned K historical activity features, the K historical evaluation attributes corresponding to the above-mentioned K historical activity features can be determined from the above-mentioned historical evaluation attribute set; according to the above-mentioned K historical evaluation attributes, the evaluation attributes of the above-mentioned target student within the third preset time can be determined.

[0068] The distance between the above-mentioned target activity features and the historical activity features can be the Euclidean distance, and specifically, it can be calculated according to the following formula:

[0069] d = sqrt((x1 - x2) 2 +(y1 - y2) 2 +(z1 - z2) 2 +…)

[0070] where d is the distance, (x1, y1, z1, …) are the coordinates of the target activity features in the Euclidean space, and (x2, y2, z2, …) are the coordinates of the historical activity features in the Euclidean space.

[0071] The Euclidean distances between the target activity features and all the historical activity features in the historical activity feature set can be traversed to obtain the distance values between the target activity features and each historical activity feature, and the K historical activity features with the smallest distance values are determined as the K historical activity features with the closest distance to the target activity features.

[0072] The corresponding K historical evaluation attributes are determined from the above-mentioned K historical activity features, and the above-mentioned historical activity features and historical evaluation attributes are associated through the student name or student number. Specifically, after obtaining the K historical activity features, the corresponding K student information can be determined, and the K historical evaluation attributes corresponding to the K student information can be determined from the historical evaluation attribute set according to the K student information.

[0073] The historical evaluation attribute with the highest occurrence frequency among the K historical evaluation attributes can be selected as the evaluation attribute of the target student to complete the prediction of the evaluation attributes of the target student.

[0074] Optionally, the occurrence frequencies of each type of historical evaluation attribute among the K historical evaluation attributes can be calculated; according to the historical evaluation attribute corresponding to the type with the highest occurrence frequency, the evaluation attributes of the target student within the third preset time can be determined.

[0075] In an embodiment of the present invention, the above-mentioned historical evaluation attributes include a performance dimension and a personality dimension. Specifically, the types of the performance dimension may include types such as performance A, performance B, performance C, performance D, etc., and the types of the personality dimension may include types such as extroverted or introverted. If among the K historical evaluation attributes, the historical evaluation attributes of the performance A type have the highest occurrence frequency in the performance dimension, and the historical evaluation attributes of the introverted type have the highest occurrence frequency in the personality dimension, then the evaluation attributes of the target student can be determined as performance A and introverted.

[0076] For example, taking K as 3, it can be as shown in Table 3 below:

[0077] Table 3

[0078]

[0079] In Table 3, the historical activity characteristics of Huang Dan, Qu Hong, and Ye Ming are the closest to the target activity characteristics of Chen Dan. Among them, in the performance dimension, the frequency of performance A among Huang Dan, Qu Hong, and Ye Ming is 2, and the frequency of performance B is 1. Then, it can be determined that Chen Dan's evaluation attribute in the performance dimension is performance A; in the personality dimension, the frequency of extroverted among Huang Dan, Qu Hong, and Ye Ming is 2, and the frequency of introverted is 1. Then, it can be determined that Chen Dan's evaluation attribute in the personality dimension is extroverted.

[0080] Optionally, a first data set can be constructed. The first data set includes first sample activity characteristics and corresponding first evaluation attribute labels; the first data set is divided into a training data set and a validation data set; a K value is initialized, and the K value is trained and adjusted through the training data set to obtain an adjusted K value; the adjusted K value is verified through the validation data set to determine the final K value. The first sample activity characteristics can be historical activity characteristics.

[0081] Specifically, a relatively small K value can be initialized, such as 1, and the K value is increased during the process of training and adjusting the K value through the training data set. The final K value can be determined by the variance of the validation data set, and the K value corresponding to the lowest variance of the validation data set can be taken as the final K value.

[0082] The variance of the validation data set is determined according to the difference degree between the training data set error and the validation data set error. The greater the difference between the training data set error and the validation data set error, the higher the variance of the validation data set. For example, if the error on the training data set is 1% and the error on the validation set is 11%, the difference between the error on the training data set and the error on the validation set is relatively large, indicating a high variance of the validation data set. By determining a suitable K value, the prediction accuracy of the evaluation attributes can be improved.

[0083] 104. Analyze the teaching quality of the school where the target student is located within the third preset time based on the assessment attributes of the target student.

[0084] In the embodiment of the present invention, by predicting the assessment attributes of the target students, the assessment attributes of all target students can be predicted before the assessment, so as to analyze the teaching quality of the school where the target students are located according to the predicted assessment attributes. Specifically, the teaching quality of the school where the target student is located within the third preset time can be analyzed according to the predicted assessment attributes, so as to obtain the teaching quality of the school where the target student is located within the third preset time.

[0085] Optionally, before step 104, the current target activity feature and the current assessment attribute can also be input into a preset confidence network to calculate the confidence of the current assessment attribute; if the confidence meets the preset conditions, the current target activity feature is added to the historical activity feature set, and the current assessment attribute is added to the historical assessment attribute set; if the confidence does not meet the preset conditions, the current target activity feature is added to the secondary prediction queue according to a preset rule for waiting, where the secondary prediction queue is used to recalculate the target activity features in the secondary prediction queue and the historical activity feature set after all target activity features are processed.

[0086] A primary confidence prediction can be performed on the predicted target activity features and the predicted assessment attributes through a preset confidence network. The above confidence is the credibility of the predicted assessment attribute. The higher the confidence, the more credible the predicted assessment attribute; the lower the confidence, the less credible the predicted assessment attribute.

[0087] When the confidence of the current assessment attribute meets the preset conditions, the current assessment attribute can be considered credible; when the confidence of the current assessment attribute does not meet the preset conditions, the current assessment attribute can be considered not credible. When the current assessment attribute is credible, the corresponding current target activity feature can be temporarily added to the historical activity feature set, and the current assessment attribute can be temporarily added to the historical assessment attribute set. In this way, the historical activity feature set and the historical assessment attribute set can be expanded to calculate the target activity features in the secondary prediction queue with more historical activity features. When the current assessment attribute is not credible, the current target activity feature is added to the secondary prediction queue according to a preset rule for waiting. The above preset rule can be a first-in, first-out rule. The secondary prediction queue will be started only after all target activity features are processed, and the target activity features in the queue and the historical activity feature set will be recalculated.

[0088] Optionally, in the embodiments of the present invention, both the target activity feature and the historical activity feature can be obtained by splicing the respective regional activity features according to a preset rule. For a campus area corresponding to a regional activity feature, a confidence network and a second data set can be constructed. The confidence network includes weight parameters and bias parameters corresponding to the regional activity feature, and the second data set includes second sample activity features and second evaluation attribute labels corresponding to the second sample activity features; the confidence network is trained through the second data set, and the weight parameters and bias parameters are continuously adjusted to make the confidence network converge, so as to obtain a trained confidence network; the trained confidence network is used as the preset confidence network. The second sample activity feature can be a historical activity feature.

[0089] Further, a confidence network can be constructed based on a linear logistic regression model. The constructed confidence network includes multiple weight parameters and multiple bias parameters. One weight parameter corresponds to one regional activity feature. The above target activity feature can be represented by (x1, x2, x3,...), and the above historical activity feature can be represented by (x1′, x2′, x3′,...). Among them, the target activity feature is obtained by splicing the regional activity features x1, x2, x3,.... For example, x1 represents the regional activity feature of the classroom, x2 represents the regional activity feature of the corridor, and x3 represents the regional activity feature of the stairs. The weight parameters can be set as w1, w2, w3,.... Among them, w1 can correspond to x1, w2 can correspond to x2, and w3 can correspond to x3. Specifically, the algorithm of the confidence network can be as shown in the following formula:

[0090] f(x) = w T x + b

[0091] Among them, f(x) is the calculation result of the confidence network, w is the weight parameter, and b is the bias parameter. It is trained with the goal of f(xi) ≈ yi, where y is the label value, which can be a manually labeled label. During the training process, the mean squared error can be minimized, as shown in the following formula:

[0092]

[0093] Among them, m is the total number of samples. The larger the mean squared error, the worse the effect of the confidence network; the smaller the mean squared error, the better the effect of the confidence network. The training process is to iteratively adjust the weight parameters and bias parameters in the confidence network with the goal of minimizing the mean squared error.

[0094] During the adjustment process of the above weight parameters, it can be carried out through the following formula:

[0095]

[0096] During the adjustment of the above bias parameters, the following formula can be used:

[0097]

[0098] In the embodiment of the present invention, the evaluation attributes and confidence levels of the target students can be predicted. For example, the performance of Chen Dan 2 in the first quarter of the next school year is: grade A, extroverted personality, confidence level 60%.

[0099] Optionally, in the embodiment of the present invention, the campus environment can be evaluated according to the weight parameters in the trained confidence network to obtain the campus environment evaluation result; the teaching quality analysis is carried out based on the campus environment evaluation result and the evaluation attributes of the target students.

[0100] Specifically, the weight w parameters in the trained confidence network can reflect the influence of each campus area on the evaluation attributes of students. Therefore, the campus environment can be evaluated according to the weight parameters in the trained confidence network. Specifically, for the campus areas with larger weight parameters, it can be considered that they have a positive impact on students and the teaching quality is better; for the campus areas with smaller weight parameters, it can be considered that they have no impact or a negative impact on students, and the teaching quality is average or poor. Thus, the teaching quality analysis can be carried out according to the evaluation result of the campus environment.

[0101] In a possible embodiment, the area or the number of areas of the campus areas with larger weight parameters can be appropriately increased, and the area or the number of areas of the campus areas with smaller weight parameters can be appropriately reduced, so as to improve the teaching quality.

[0102] In an embodiment of the present invention, according to the campus activity file of the target student, the target activity characteristics of the target student within a first preset time period are extracted; in the historical campus activity file, the historical activity characteristics and historical evaluation attributes corresponding to each student within a second preset time period are extracted to obtain a historical activity characteristic set and a historical evaluation attribute set; according to the target activity characteristics, the historical activity characteristic set and the historical evaluation attribute set, the evaluation attribute of the target student within a third preset time is predicted; and teaching quality analysis is performed based on the evaluation attribute of the target student. By establishing files for students' campus activities, extracting the target activity characteristics of students within a first preset time period from the students' campus activity files, and extracting the historical activity characteristics and historical evaluation attributes of each student within a second preset time period from the historical campus activity files, the evaluation attributes of students are predicted using the target activity characteristics, historical activity characteristics, and historical evaluation attributes, and teaching quality analysis is performed based on the predicted evaluation attributes. Without waiting for a too long assessment cycle, the potential common connection between students' campus activities and evaluation attributes is used to predict evaluation attributes, improving the prediction accuracy of evaluation attributes. The evaluation attributes of students at any current or future time period can be predicted, so that the teaching quality at the current or future time can be accurately evaluated, providing continuous and real-time reference for improving teaching quality.

[0103] It should be noted that the teaching quality analysis method provided in the embodiment of the present invention can be applied to devices such as smartphones, computers, and servers that can perform teaching quality analysis.

[0104] Optionally, please refer to Figure 2 , Figure 2 which is a schematic structural diagram of a teaching quality analysis device provided in an embodiment of the present invention. As Figure 2 shown, the device includes:

[0105] A first extraction module 201, configured to extract the target activity characteristics of the target student within a first preset time period according to the campus activity file of the target student;

[0106] A second extraction module 202, configured to extract the historical activity characteristics and historical evaluation attributes corresponding to each student within a second preset time period from the historical campus activity file to obtain a historical activity characteristic set and a historical evaluation attribute set;

[0107] A prediction module 203, configured to predict the evaluation attribute of the target student within a third preset time according to the target activity characteristics, the historical activity characteristic set, and the historical evaluation attribute set;

[0108] An analysis module 204, configured to perform teaching quality analysis on the school where the target student is located within a third preset time based on the evaluation attribute of the target student.

[0109] Optionally, the prediction module 203 includes:

[0110] A calculation sub-module, configured to calculate the distance between the target activity feature and each of the historical activity features in the historical activity feature set, and determine K historical activity features that are closest to the target activity feature;

[0111] A first determination sub-module, configured to determine, based on the K historical activity features, K historical evaluation attributes corresponding to the K historical activity features from the historical evaluation attribute set;

[0112] A second determination sub-module, configured to determine the evaluation attribute of the target student within a third preset time according to the K historical evaluation attributes.

[0113] Optionally, the second determination sub-module includes:

[0114] A calculation unit, configured to calculate the occurrence frequency of each type of historical evaluation attribute among the K historical evaluation attributes;

[0115] A determination unit, configured to determine the evaluation attribute of the target student within a third preset time according to the historical evaluation attribute corresponding to the type with the highest occurrence frequency.

[0116] Optionally, the prediction module 203 further includes:

[0117] A construction sub-module, configured to construct a first data set, where the first data set includes first sample activity features and first evaluation attribute labels corresponding to the first sample activity features;

[0118] A division sub-module, configured to divide the first data set into a training data set and a validation data set;

[0119] A training sub-module, configured to initialize a K value, and perform training and adjustment on the K value through the training data set to obtain an adjusted K value;

[0120] A third determination sub-module, configured to perform a validation process on the adjusted K value through the validation data set to determine the final K value.

[0121] Optionally, the device further includes:

[0122] A calculation module, configured to input the current target activity feature and the current evaluation attribute into a preset confidence network to calculate the confidence of the current evaluation attribute, and the secondary prediction queue is used to recalculate the target activity features in the secondary prediction queue and the historical activity feature set after all target activity features are processed;

[0123] The first addition module is used to add the current target activity feature to the historical activity feature set and add the current evaluation attribute to the historical evaluation attribute set if the confidence level meets a preset condition;

[0124] The second addition module is used to add the current target activity feature to the secondary prediction queue for waiting according to a preset rule if the confidence level does not meet the preset condition.

[0125] Optionally, both the target activity feature and the historical activity feature are obtained by splicing various regional activity features according to a preset rule. One campus area corresponds to one regional activity feature. The device further includes:

[0126] The construction module is used to construct a confidence network and a second data set. The confidence network includes weight parameters and bias parameters corresponding to the regional activity features, and the second data set includes second sample activity features and second evaluation attribute labels corresponding to the second sample activity features;

[0127] The training module is used to train the confidence network through the second data set, and make the confidence network converge by continuously adjusting the weight parameters and the bias parameters to obtain a trained confidence network, and use the trained confidence network as the preset confidence network.

[0128] Optionally, the analysis module 204 includes:

[0129] The evaluation sub-module is used to evaluate the campus environment according to the weight parameters in the trained confidence network to obtain a campus environment evaluation result;

[0130] The analysis sub-module is used to perform teaching quality analysis based on the campus environment evaluation result and the evaluation attributes of the target student.

[0131] It should be noted that the teaching quality analysis device provided by the embodiments of the present invention can be applied to devices such as smart phones, computers, and servers that can perform teaching quality analysis.

[0132] The teaching quality analysis device provided by the embodiments of the present invention can implement each process implemented by the teaching quality analysis method in the above method embodiments and can achieve the same beneficial effects. To avoid repetition, it will not be elaborated here.

[0133] See Figure 3 , Figure 3 is a schematic structural diagram of an electronic device provided by the embodiments of the present invention. As Figure 3As shown in the figure, it includes: a memory 302, a processor 301, and a computer program of a teaching quality analysis method stored on the memory 302 and operable on the processor 301, where:

[0134] The processor 301 is configured to call the computer program stored in the memory 302 and execute the following steps:

[0135] Extract the target activity characteristics of the target student within a first preset time period according to the campus activity file of the target student;

[0136] Extract the historical activity characteristics and historical evaluation attributes corresponding to each student within a second preset time period from the historical campus activity files to obtain a historical activity characteristic set and a historical evaluation attribute set;

[0137] Predict the evaluation attribute of the target student within a third preset time according to the target activity characteristics, the historical activity characteristic set, and the historical evaluation attribute set;

[0138] Analyze the teaching quality of the school where the target student is located within a third preset time based on the evaluation attribute of the target student.

[0139] Optionally, the predicting, by the processor 301, the evaluation attribute of the target student within a third preset time according to the target activity characteristics, the historical activity characteristic set, and the historical evaluation attribute set includes:

[0140] Calculate the distances between the target activity characteristics and each of the historical activity characteristics in the historical activity characteristic set, and determine the K historical activity characteristics that are closest to the target activity characteristics;

[0141] Based on the K historical activity characteristics, determine the K historical evaluation attributes corresponding to the K historical activity characteristics from the historical evaluation attribute set;

[0142] Determine the evaluation attribute of the target student within a third preset time according to the K historical evaluation attributes.

[0143] Optionally, the determining, by the processor 301, the evaluation attribute of the target student within a third preset time according to the K historical evaluation attributes includes:

[0144] Calculate the occurrence frequencies of each type of historical evaluation attribute among the K historical evaluation attributes;

[0145] Determine the evaluation attribute of the target student within a third preset time according to the historical evaluation attribute corresponding to the type with the highest occurrence frequency.

[0146] Optionally, before the step of calculating the distance between the target activity feature and the historical activity feature and determining the K historical activity features closest to the target activity feature, the method executed by the processor 301 further includes:

[0147] Construct a first data set, where the first data set includes first sample activity features and first evaluation attribute labels corresponding to the first sample activity features;

[0148] Divide the first data set into a training data set and a validation data set;

[0149] Initialize a K value, and perform training and adjustment on the K value through the training data set to obtain an adjusted K value;

[0150] Perform verification processing on the adjusted K value through the validation data set to determine the final K value.

[0151] Optionally, before the step of analyzing teaching quality based on the evaluation attributes of the target student, the method executed by the processor 301 further includes:

[0152] Input the current target activity feature and the current evaluation attribute into a preset confidence network to calculate the confidence of the current evaluation attribute;

[0153] If the confidence meets the preset condition, add the current target activity feature to the historical activity feature set and add the current evaluation attribute to the historical evaluation attribute set;

[0154] If the confidence does not meet the preset condition, add the current target activity feature to a secondary prediction queue according to a preset rule for waiting. The secondary prediction queue is used to recalculate the target activity features in the secondary prediction queue and the historical activity feature set after all target activity features are processed.

[0155] Optionally, both the target activity feature and the historical activity feature are obtained by splicing various regional activity features according to a preset rule. One campus area corresponds to one regional activity feature. Before the step of inputting the current target activity feature and the current evaluation attribute into a preset confidence network to calculate the confidence of the current evaluation attribute, the method executed by the processor 301 further includes:

[0156] Construct a confidence network and a second data set. The confidence network includes weight parameters and bias parameters corresponding to the regional activity features. The second data set includes second sample activity features and second evaluation attribute labels corresponding to the second sample activity features;

[0157] The confidence network is trained with the second data set, and the confidence network is converged by continuously adjusting the weight parameters and the bias parameters to obtain a trained confidence network, and the trained confidence network is used as the preset confidence network.

[0158] Optionally, the teaching quality analysis performed by the processor 301 based on the evaluation attributes of the target student includes:

[0159] The campus environment is evaluated according to the weight parameters in the trained confidence network to obtain a campus environment evaluation result;

[0160] Teaching quality analysis is performed based on the campus environment evaluation result and the evaluation attributes of the target student.

[0161] The electronic device provided in the embodiment of the present invention can implement each process implemented by the teaching quality analysis method in the above method embodiment, and can achieve the same beneficial effects. To avoid repetition, it will not be described in detail here.

[0162] The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements each process of the teaching quality analysis method or the teaching quality analysis method of the application end provided by the embodiment of the present invention, and can achieve the same technical effects. To avoid repetition, it will not be described in detail here.

[0163] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0164] The above-disclosed are only the preferred embodiments of the present invention. Of course, the scope of the rights of the present invention cannot be limited thereby. Therefore, equivalent changes made according to the claims of the present invention still fall within the scope covered by the present invention.

Claims

1. A method for analyzing teaching quality, characterized in that, it includes the following steps: According to the campus activity file of the target student, extract the target activity characteristics of the target student within the first preset time period; Extract the historical activity characteristics and historical evaluation attributes corresponding to each student within the second preset time period from the historical campus activity file to obtain a historical activity characteristic set and a historical evaluation attribute set; Predict the evaluation attribute of the target student within the third preset time according to the target activity characteristics, the historical activity characteristic set and the historical evaluation attribute set; Input the current target activity characteristics and the current evaluation attribute into a preset confidence network to calculate the confidence of the current evaluation attribute; If the confidence meets the preset conditions, add the current target activity characteristics to the historical activity characteristic set, and add the current evaluation attribute to the historical evaluation attribute set to expand the historical activity characteristic set and the historical evaluation attribute set, and calculate the target activity characteristics in the secondary prediction queue with more historical activity characteristics; If the confidence does not meet the preset conditions, add the current target activity characteristics to the secondary prediction queue according to the preset rules for waiting. The secondary prediction queue is used to recalculate the target activity characteristics in the secondary prediction queue and the historical activity characteristic set after all target activity characteristics are processed; Analyze the teaching quality of the school where the target student is located within the third preset time based on the evaluation attribute of the target student.

2. The method according to claim 1, characterized in that, the predicting the evaluation attribute of the target student within the third preset time according to the target activity characteristics, the historical activity characteristic set and the historical evaluation attribute set includes: Calculate the distance between the target activity characteristics and each of the historical activity characteristics in the historical activity characteristic set, and determine the K historical activity characteristics closest to the target activity characteristics; Based on the K historical activity characteristics, determine the K historical evaluation attributes corresponding to the K historical activity characteristics from the historical evaluation attribute set; Determine the evaluation attribute of the target student within the third preset time according to the K historical evaluation attributes.

3. The method according to claim 2, characterized in that, the determining the evaluation attribute of the target student within the third preset time according to the K historical evaluation attributes includes: Calculate the occurrence frequency of each type of historical evaluation attribute among the K historical evaluation attributes; Determine the evaluation attribute of the target student within the third preset time according to the historical evaluation attribute corresponding to the type with the highest occurrence frequency.

4. The method according to claim 2, characterized in that, before the step of calculating the distance between the target activity characteristics and the historical activity characteristics and determining the K historical activity characteristics closest to the target activity characteristics, the method further includes: Construct a first data set, which includes first sample activity characteristics and first evaluation attribute labels corresponding to the first sample activity characteristics; Divide the first data set into a training data set and a validation data set; Initialize a K value, and train and adjust the K value through the training data set to obtain an adjusted K value; Perform a verification process on the adjusted K value through the verification data set to determine the final K value.

5. The method according to claim 1, characterized in that, both the target activity feature and the historical activity feature are obtained by splicing each regional activity feature according to a preset rule. A campus area corresponds to one regional activity feature. Before the step of inputting the current target activity feature and the current evaluation attribute into a preset confidence network to calculate the confidence of the current evaluation attribute, the method further includes: Construct a confidence network and a second data set. The confidence network includes weight parameters and bias parameters corresponding to the regional activity features, and the second data set includes second sample activity features and second evaluation attribute labels corresponding to the second sample activity features; Train the confidence network through the second data set, and continuously adjust the weight parameters and the bias parameters to make the confidence network converge, so as to obtain a trained confidence network, and use the trained confidence network as the preset confidence network.

6. The method according to claim 5, characterized in that, the teaching quality analysis based on the evaluation attributes of the target student includes: Evaluate the campus environment according to the weight parameters in the trained confidence network to obtain a campus environment evaluation result; Conduct teaching quality analysis based on the campus environment evaluation result and the evaluation attributes of the target student.

7. A teaching quality analysis device, characterized in that, the device includes: A first extraction module, configured to extract the target activity features of the target student within a first preset time period according to the campus activity file of the target student; A second extraction module, configured to extract the historical activity features and historical evaluation attributes corresponding to each student within a second preset time period from the historical campus activity file to obtain a historical activity feature set and a historical evaluation attribute set; A prediction module, configured to predict the evaluation attributes of the target student within a third preset time according to the target activity features, the historical activity feature set and the historical evaluation attribute set; A calculation module, configured to input the current target activity feature and the current evaluation attribute into a preset confidence network to calculate the confidence of the current evaluation attribute; A first addition module, configured to, if the confidence meets a preset condition, add the current target activity feature to the historical activity feature set, and add the current evaluation attribute to the historical evaluation attribute set, so as to expand the historical activity feature set and the historical evaluation attribute set, and calculate the target activity features in the secondary prediction queue with more historical activity features; A second addition module, configured to, if the confidence does not meet a preset condition, add the current target activity feature to the secondary prediction queue according to a preset rule for waiting. The secondary prediction queue is used to recalculate the target activity features in the secondary prediction queue and the historical activity feature set after all the target activity features are processed; An analysis module, configured to analyze the teaching quality of the school where the target student is located within a third preset time based on the evaluation attributes of the target student.

8. An electronic device, characterized in that it includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, where when the processor executes the computer program, the steps in the teaching quality analysis method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps in the teaching quality analysis method according to any one of claims 1 to 6 are implemented.

Images