CN112668476A - Data processing method and device, electronic equipment and storage medium - Google Patents
Data processing method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN112668476A CN112668476A CN202011585656.6A CN202011585656A CN112668476A CN 112668476 A CN112668476 A CN 112668476A CN 202011585656 A CN202011585656 A CN 202011585656A CN 112668476 A CN112668476 A CN 112668476A
- Authority
- CN
- China
- Prior art keywords
- matrix
- data
- representing
- membership
- teaching
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 13
- 239000011159 matrix material Substances 0.000 claims abstract description 305
- 230000006399 behavior Effects 0.000 claims abstract description 66
- 238000000034 method Methods 0.000 claims abstract description 31
- 230000036544 posture Effects 0.000 claims description 62
- 230000008451 emotion Effects 0.000 claims description 43
- 239000013598 vector Substances 0.000 claims description 40
- 238000012545 processing Methods 0.000 claims description 25
- 230000008921 facial expression Effects 0.000 claims description 22
- 230000014509 gene expression Effects 0.000 claims description 18
- 210000003141 lower extremity Anatomy 0.000 claims description 17
- 238000011156 evaluation Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 4
- 238000013441 quality evaluation Methods 0.000 abstract description 4
- 238000004364 calculation method Methods 0.000 description 10
- 238000005516 engineering process Methods 0.000 description 6
- 230000001815 facial effect Effects 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 230000007935 neutral effect Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000002996 emotional effect Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 108010076282 Factor IX Proteins 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000008909 emotion recognition Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000036651 mood Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
Abstract
The application provides a data processing method, a device, an electronic device and a storage medium, wherein the data processing method specifically comprises the following steps: firstly, acquiring original audio and video data for teaching of a teacher; then extracting characteristic data representing teaching behaviors from the original audio and video data, and generating a corresponding membership matrix according to the characteristic data, wherein each row in the membership matrix is used for representing the ratio score of the characteristic data corresponding to the row to a preset standard; and finally, obtaining a target matrix for representing the quality of the teaching behavior based on the membership matrix and a preset weight matrix corresponding to the characteristic data. The method and the device can objectively and accurately evaluate the quality of the teaching behavior, and solve the problem that the quality of the teaching behavior cannot be objectively and accurately evaluated by the conventional teaching behavior quality evaluation method.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a data processing method and apparatus, an electronic device, and a storage medium.
Background
Teacher's classroom teaching action is the important link of evaluation teacher's teaching quality, through fully knowing teacher's teaching action quality, can guarantee high-quality teaching level. At present, the teaching behavior quality of a teacher is obtained by adopting a mode of student record or teacher test simulation and teacher observation and supervision, the mode cannot objectively and comprehensively evaluate the teaching effect of the teacher, and the collection, analysis, record and evaluation of the classroom behavior of the teacher cannot be realized.
Disclosure of Invention
The application provides a data processing method, a data processing device, electronic equipment and a storage medium, which are used for solving the problem that the quality of a teaching behavior cannot be objectively and accurately evaluated by the conventional teaching behavior quality evaluation method.
In a first aspect, the present application provides a data processing method, including: acquiring original audio and video data for teaching of a teacher; extracting characteristic data representing teaching behaviors from the original audio and video data, and generating a corresponding membership matrix according to the characteristic data, wherein each row in the membership matrix is used for representing the ratio score of the characteristic data corresponding to the row to a preset standard; and obtaining a target matrix for representing the quality of the teaching behaviors based on the membership matrix and a preset weight matrix corresponding to the characteristic data.
In the embodiment of the application, the original audio and video data taught by a teacher are obtained, the characteristic data representing the teaching behavior is extracted from the obtained original audio and video data, the corresponding membership matrix is generated according to the characteristic data, and then the target matrix used for representing the quality of the teaching behavior can be obtained based on the membership matrix and the preset weight matrix corresponding to the characteristic data, so that the quality of the teaching behavior is objectively and accurately evaluated, the problem that the quality of the teaching behavior cannot be objectively and accurately evaluated by the conventional teaching behavior quality evaluation method is solved, and a feasible scheme is provided for objectively and accurately evaluating the quality of the teaching behavior.
With reference to the technical solution provided by the first aspect, in some possible implementations, extracting feature data representing a teaching behavior from the original audio/video data includes: extracting characteristic data respectively corresponding to facial expressions, voice emotions, head postures, hand postures and lower limb postures representing teaching behaviors and relative distances between a teacher and students from the original audio and video data; correspondingly, generating a corresponding membership matrix according to the characteristic data, comprising: generating a corresponding first membership matrix according to the feature data corresponding to the facial expression and the voice emotion respectively; generating a corresponding second membership matrix according to the characteristic data corresponding to the head posture, the hand posture and the lower limb posture respectively; generating a corresponding third membership matrix according to the feature data corresponding to the relative distance between the teacher and the student; correspondingly, obtaining a target matrix for representing the quality of the teaching behavior based on the membership matrix and the preset weight matrix corresponding to the characteristic data, including: obtaining a first matrix for representing emotion expression behavior quality based on the first membership matrix and preset weight matrixes corresponding to the facial expressions and the voice emotions; obtaining a second matrix for representing the behavior quality of the body posture based on the second membership matrix and preset weight matrixes corresponding to the head posture, the hand posture and the lower limb posture; obtaining a third matrix for representing the body distance behavior quality based on the third membership matrix and a preset weight matrix corresponding to the relative distance between the teacher and the students; and obtaining the target matrix based on the first matrix, the second matrix and the third matrix.
In the embodiment of the application, by acquiring feature data corresponding to facial expressions, voice emotions, head postures, hand postures, lower limb postures and relative distances between a teacher and students, classroom teaching behaviors of the teacher are comprehensively and objectively extracted, the facial expressions and the voice emotions which are used for reflecting emotion expressions are analyzed together, and a first membership matrix reflecting proportion scores of the emotion expressions and preset standards is obtained; analyzing the head posture, the hand posture and the lower limb posture which are all used for reflecting the body posture together to obtain a second membership matrix reflecting the proportion score of the body posture and a preset standard; analyzing the relative distance between the teacher and the student for reflecting the body distance to obtain a third membership matrix for reflecting the proportion score of the body distance and a preset standard; then, obtaining a first matrix for representing emotion expression behavior quality based on the first membership matrix and a preset weight matrix corresponding to feature data corresponding to facial expressions and voice emotions; obtaining a second matrix for representing the body posture behavior quality based on the second membership matrix and a preset weight matrix corresponding to the feature data corresponding to the head posture, the hand posture and the lower limb posture; obtaining a third matrix for representing the body distance behavior quality based on the third membership matrix and a preset weight matrix corresponding to the relative distance between the teacher and the students; and combining the first matrix, the second matrix and the third matrix to obtain a target matrix reflecting the overall teaching quality. Through the characteristic data of the categories such as facial expressions, voice emotions, head postures, hand postures, lower limb postures and relative distances between teachers and students, the teaching behaviors of the teachers can be comprehensively analyzed, and the quality of the teaching behaviors can be more accurately and comprehensively reflected by a finally obtained target matrix.
With reference to the technical solution provided by the first aspect, in some possible embodiments, obtaining an evaluation matrix price for characterizing teaching quality based on the membership matrix and a preset weight matrix corresponding to the feature data includes: and obtaining the target matrix for representing the teaching quality based on a preset fuzzy operator and the preset weight matrix corresponding to the membership degree matrix and the characteristic data.
In the embodiment of the application, a target matrix is obtained by using a preset fuzzy operator, a membership matrix and a preset weight matrix. According to the actual situation, a proper fuzzy operator can be selected, so that the calculated target matrix is more consistent with the actual situation, the application range of the scheme is expanded, and the reliability and the accuracy of the final result are improved.
With reference to the technical solution provided by the first aspect, in some possible implementations, a preset weight matrix corresponding to the feature data is obtained through the following steps: acquiring a judgment matrix for representing the importance of the characteristic data; and acquiring a feature vector corresponding to the judgment matrix, wherein the feature vector is transposed to obtain the preset weight matrix.
In the embodiment of the application, the characteristic vector of the judgment matrix is obtained, and then the characteristic vector is transposed to obtain the preset weight matrix corresponding to the characteristic data, namely the preset weight matrix is obtained through calculation instead of artificial definition, so that the final result obtained through calculation based on the weight matrix is more objective, and the subjectivity of the target matrix is reduced.
With reference to the technical solution provided by the first aspect, in some possible implementations, obtaining a judgment matrix for characterizing importance of the feature data includes: determining the importance degree of ith class of feature data to jth class of feature data in the feature data, wherein i and j are positive integers, and i and j are not more than the total number of classes of the feature data; and establishing the judgment matrix based on the importance degree of the ith class of characteristic data to the jth class of characteristic data.
In the embodiment of the application, the judgment matrix is generated by determining the relative importance among the various types of feature data, so that the relative importance among the various types of feature data can be effectively reflected by the weight value obtained by subsequent calculation according to the judgment matrix, and the reliability of the final result is ensured.
With reference to the technical solution provided by the first aspect, in some possible implementations, after the establishing the determination matrix, the method further includes: acquiring a characteristic vector of the judgment matrix, and obtaining a consistency ratio of the judgment matrix based on the judgment matrix and the characteristic vector of the judgment matrix; judging whether the consistency of the judgment matrix is qualified or not based on the consistency ratio; and when the judgment is unqualified, reestablishing the judgment matrix.
In the embodiment of the application, consistency check is carried out on the generated judgment matrix to judge whether the consistency of the judgment matrix is qualified or not, when the judgment matrix is unqualified, the judgment matrix is reestablished, and consistency check is carried out on the reestablished judgment matrix until the consistency check is passed. By the scheme, whether the elements of the judgment matrix have obvious difference can be detected, and contradiction can be prevented from occurring in the relative importance of characteristic data of each category in the judgment matrix, so that errors occur in the calculated weight value. Therefore, the reliability of the final evaluation result can be improved by the consistency check.
With reference to the technical solution provided by the first aspect, in some possible implementations, the obtaining the feature vector of the determination matrix includes: obtaining products of all elements in each row based on each row of the judgment matrix; aiming at the product of each row of elements, obtaining the square root of the product of each row of elements for n times, wherein n is the number of the elements in each row of the judgment matrix; and generating an initial vector according to the obtained n-th-order square root of each row, and performing normalization operation on the initial vector to obtain a feature vector of the judgment matrix.
In the embodiment of the application, the characteristic vector of the judgment matrix is solved by using the square root method, the calculation process is simpler, the calculation amount can be reduced, and the calculation efficiency is improved.
In a second aspect, an embodiment of the present application provides a data processing apparatus, including an obtaining module and a processing module, where the obtaining module is configured to obtain original audio and video data for teaching by a teacher; the processing module is used for extracting feature data representing teaching behaviors from the original audio and video data and generating a corresponding membership matrix according to the feature data, wherein each row in the membership matrix is used for representing the proportion score of the feature data corresponding to the row and a preset standard; the processing module is further used for obtaining a target matrix for representing the quality of the teaching behaviors based on the membership degree matrix and the preset weight matrix corresponding to the characteristic data.
In a third aspect, an embodiment of the present application provides an electronic device, including: a memory and a processor, the memory and the processor connected; the memory is used for storing programs; the processor is configured to invoke a program stored in the memory to perform the method according to the embodiment of the first aspect and/or any possible implementation manner in combination with the embodiment of the first aspect.
In a fourth aspect, embodiments of the present application provide a storage medium, on which a computer program is stored, where the computer program, when executed by a computer, performs a method as described in the embodiments of the first aspect and/or any possible implementation manner in combination with the embodiments of the first aspect.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained from the drawings without inventive effort.
Fig. 1 is a schematic flow chart illustrating a data processing method according to an embodiment of the present application;
fig. 2 is a schematic structural diagram of a data processing apparatus according to an embodiment of the present application;
fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
The terms "first," "second," "third," and the like are used for descriptive purposes only and not for purposes of indicating or implying relative importance, and do not denote any order or order.
The technical solution of the present application will be clearly and completely described below with reference to the accompanying drawings.
At present, the quality of teaching behaviors of teachers is evaluated mainly by adopting experience description modes such as classroom observation, behavior comparison and the like, mainly by an expert and teacher field observation or experience evaluation methods such as classroom video analysis and the like, and a target matrix capable of objectively, accurately and fairly reflecting the quality of the teaching behaviors is lacked. Therefore, the scheme provides a data processing method to reduce the phenomenon that the evaluation of the teaching behavior depends on manual evaluation in the prior art.
Referring to fig. 1, fig. 1 is a data processing method according to an embodiment of the present application, and the steps included in the data processing method are described below with reference to fig. 1.
S100: the method comprises the steps of obtaining original audio and video data for teachers to teach, extracting characteristic data representing teaching behaviors from the original audio and video data, and generating a corresponding membership matrix according to the characteristic data.
Extracting characteristic data representing teaching behaviors from the original audio and video data, and generating a corresponding membership matrix according to the characteristic data, wherein each row in the membership matrix is used for representing the proportion score of the characteristic data corresponding to the row and a preset standard. The obtained membership matrix can be stored in a database and can be directly called when needed. The original audio and video data for the teacher to teach comprises video data and audio data for the teacher to teach.
In one embodiment, the characteristic data representing the teaching behavior is extracted from the original data by an artificial intelligence technology (intelligent recording and broadcasting technology, emotion recognition technology, intelligent recognition technology and the like). The characteristic data is analyzed, quantitatively calculated and fed back according to the actual classroom performance of the teacher, and each type of characteristic data is scored. In a class, even the same type of feature data changes with the change of time, so that the performance of the class can be scored based on preset standards; for example, the preset criterion may indicate that the feature data is judged to be four grades of "good", "better", "general" and "poor" in a certain time. For example, when the situation is obtainedThe characteristic data of the expression is R1＝[0.2 0.5 0.3 0]In this case, the meaning is that, in the video, compared with the total classroom duration, the time occupation ratio of the teacher with good emotion expression is 0.2, the better time occupation ratio is 0.5, the general occupation ratio is 0.3, and the worse time occupation ratio is 0, which is only an example provided by this application and should not be taken as a limitation to this application.
Since the evaluation value of each feature data is different, different grades are often formed, and therefore, the preset standard needs to be determined. In one embodiment, the predetermined criteria consisting of various decisions is denoted as V ═ V1,v2,v3,v4Good, better, generally, worse. According to a preset standard V, determining the characteristic data of each type of characteristic data for representing each judgment grade score of the characteristic data as ri＝(ri1,ri2,ri3,ri4) Wherein r isi1The ratio of the time for which the characteristic data is judged to be "good" to the total time, ri2The ratio of the time in which the characteristic data is judged to be "better" to the total time length, ri3The ratio of the time of which the characteristic data is judged to be 'normal' to the total time length, ri4Indicating the ratio of the time the signature data was judged to be "bad" relative to the total duration. And each judgment grade score of each type of feature data is a score which is obtained by analyzing and quantitatively calculating and feeding back the teacher classroom video and the teacher record according to the actual classroom performance of the teacher by using an artificial intelligence technology. And then generating a membership matrix based on the characteristic data. Taking facial expressions and speech emotions as examples, the corresponding membership matrix is
This is merely one example provided herein and should not be taken as a limitation on the present application. Wherein r is111The ratio of time to the total time length, r, of the feature data representing the facial expression judged as "good" is represented112The ratio of time to the total time length, r, of the feature data representing the facial expression judged as "better" is represented113The ratio of time to the total time length, r, representing the feature data representing the facial expression judged as "general114The ratio of the time of the feature data representing the facial expression judged to be "poor" to the total time length is represented, and the meanings of the rest row representations in the membership matrix are similar to the ratio, and are not repeated herein.
In one embodiment, for a category of feature data, the time length of each judgment factor for judging the category of feature data appearing in the classroom teaching video is counted, and then the result is obtained
Calculating the ratio of each grade of the feature data, wherein j represents that the preset standard has j grades, n represents the number of judgment factors of the kind of feature data, and betanDenotes the regression coefficient, αjDenotes a constant term, xnThe length of time that the nth judging factor representing the feature data of the category appears in the classroom teaching video. For different classes of feature data, their corresponding betan、αjAll are different in thatn、αjThe method may be performed by using a large amount of sample data obtained in advance, using SPSS (Statistical Product and Service Solutions) or MATLAB (matrix)&A matrix factory) tool, and calculating regression coefficients and constant terms of the corresponding class feature data.
The time length statistics method may be as follows: the method comprises the steps of dividing a classroom teaching video into a plurality of teaching segments with the same time length according to the same time length, aiming at characteristic data of one type, judging which judgment factor of the characteristic data the teacher's performance belongs to in each teaching segment, counting the number of each judgment factor after judging all the teaching segments, and obtaining the time length of each judgment factor based on the time length of each teaching segment.
For example, taking a 45-minute classroom teaching video as an example, selecting 3 seconds as the time period for dividing the classroom teaching video, the classroom teaching video is divided into 900 teaching segmentsWhen the type of the judged feature data is facial emotion, judging the facial emotion of a teacher in each teaching segment, wherein the judgment factor for judging the facial emotion of the teacher can be positive emotion, neutral emotion and negative emotion, and the picture of a specific frame of each teaching segment can be used as a judgment basis, for example, the first frame pictures of all teaching segments are judged; or randomly selecting a frame of picture for judging each teaching segment. The number of positive emotions measured was recorded as u1111And the number of measured neutral emotions is recorded as u1112And the number of negative emotions measured is recorded as u1113And u is1111+u1112+u1113900, the time x occupied by the positive emotion in the teaching video can be obtained1Is 3 x u1111Second, time occupied by neutral mood x2Is 3 x u1112Second, time occupied by negative emotion x3Is 3 x u1113And second. Then
The ratio of the characteristic "good" is
The ratio characterizing "better" is
The ratio characterized by "general" is
Characterised by a "poor" ratio of p4＝1-(p1+p2+p3) Then the feature data characterizing facial emotion is (p)1 p2 p3 p4). This embodiment is only one embodiment of the present application and should not be taken as limiting the present application.
Different teaching behaviors and different types of corresponding characteristic data. In one embodiment, the types of characteristic data that can comprehensively evaluate the teaching quality are shown in table 1.
TABLE 1
S200: and obtaining a target matrix for representing the quality of the teaching behavior based on the membership matrix and the preset weight matrix corresponding to the characteristic data.
After obtaining the membership matrix and the preset weight matrix corresponding to the feature data, in one embodiment, a target matrix for characterizing the quality of the teaching behavior is obtained by calculating a cartesian product of the membership matrix and the preset weight matrix. In another embodiment, a target matrix for representing teaching quality is obtained based on a preset fuzzy operator and a preset weight matrix corresponding to the membership matrix and the characteristic data. The operation rule of the fuzzy operator can be selected according to the actual situation, so that the calculated target matrix is more consistent with the actual situation, and the reliability and the accuracy of the final evaluation result are improved.
Wherein, R is used for representing the membership degree matrix, D is used for representing the preset weight matrix, when the total number of the characteristic data categories is n, and the preset standard V is { V ═ V }1,v2,v3,v4When { good, and generally poor }, the membership matrix R is an n × 4 matrix, and the predetermined weight matrix D is a 1 × n matrix. And calculating a Cartesian product of the preset weight matrix D and the membership degree matrix R to obtain a 1 x n target matrix. When calculating the target matrix by using fuzzy operator, adopting
Representing a fuzzy operator, then
When the fuzzy operators adopt different operation rules, the calculation results are also different, and here, the operation rule of the fuzzy operator which is common in 4 is taken as an example:
1. the fuzzy operator is an M (,) operator (Zadeh operator, also called "size-taking operator"), and the operation rule can be expressed as:
wherein n is the total number of the characteristic data categories, m is the total number of elements in each row in the membership degree matrix, diIs the ith element of the weight matrix, rjkIs the kth element of the ith row of the membership matrix.
2. The fuzzy operator is an M (-) operator, "·" represents multiplication, then a maximum operator is taken, and the operation rule can be expressed as:
wherein n and M have the same meanings as in the M (,) operator.
3. Fuzzy operator is
And an operator, which indicates that the minimum value is solved first and then added, wherein the operation rule can be expressed as:
wherein n and M have the same meanings as in the M (,) operator.
4. Fuzzy operator is
And an operator which represents multiplication and then addition, wherein the operation rule can be expressed as:
wherein n and M have the same meanings as in the M (,) operator.
The four fuzzy operators are only an embodiment, and should not be taken as a limitation of the present application, and each of the four fuzzy operators has characteristics, and the characteristics in the comprehensive evaluation are shown in table 2:
TABLE 2
The preset weight matrix may be obtained in advance, stored in a database, and called directly when needed, or the preset weight matrix may be calculated when needed. In one embodiment, the preset weight matrix may be obtained by: acquiring a judgment matrix for representing the importance of the feature data; and acquiring the eigenvector of the judgment matrix, and transposing the eigenvector to obtain a weight matrix corresponding to the characteristic data. For example, with A ═ aij)n×nRepresenting a judgment matrix, wherein n represents the total number of the characteristic data categories, i is greater than or equal to 1 and less than or equal to n, and j is greater than or equal to 1 and less than or equal to n, the above calculation process can be represented as: firstly, calculating the product of each row of elements of the judgment matrix, wherein the expression of the product is as follows:
then calculate MiN times of root:
will vector
Normalized to obtain
The feature vector is then: w ═ W1 W2 … Wn]TTransposing the feature vectors to obtain a weight matrix as follows: d ═ D1 D2 … Dn]Wherein D is1＝W1；D2＝W2；D3＝W3；D4＝W4。
The judgment matrix can be obtained in advance, stored in a database and called directly when needed, or the judgment matrix can be established according to the importance of the feature data when needed. In one embodiment, the step of obtaining a decision matrix for characterizing the importance of the feature data includes: firstly, determining the importance degree of ith type characteristic data to jth type characteristic data in the characteristic data, wherein i and j are positive integers, and i and j are not more than the total number of the characteristic data categories; and then establishing a judgment matrix based on the importance degree of the ith class of characteristic data to the jth class of characteristic data. When the total number of the feature data categories is n, i and j are sequentially selected from 1, and only after the value of j is taken from 1 to n, the value of i is added with 1, and the value of j is taken from 1 again until the values of i and j are both n.
Wherein the judgment matrix may be a ═ aij)n×nWherein n represents the total number of the characteristic data categories, i is more than or equal to 1 and less than or equal to n, and j is more than or equal to 1 and less than or equal to n. a isijRepresenting the importance degree of the ith class characteristic data to the jth class characteristic data, wherein i and j are sequentially valued from 1, and aijAssignment of values typically uses a scale of 1-9. Wherein the significance of the scale of 1-9 is shown in Table 3.
TABLE 3
| Proportional scaling | Significance of the significance |
| 1 | Indicates that the two factors are compared and have equal importance |
| 3 | Indicating that one factor is slightly more important than the other factor when compared to the other factor |
| 5 | Indicating a comparison of two factors, one factor over the otherThe factors being important |
| 7 | Indicating that one factor is more important than the other factor |
| 9 | Indicating that one factor is more important than the other factor |
Wherein 2, 4, 6, 8 are intermediate values of adjacent judgment scales, and 1/aijHaving a andijthe opposite meaning.
After the determination matrix is established, the determination matrix needs to be checked, and in one embodiment, whether the established determination matrix is qualified is determined by checking the consistency of the determination matrix. The step of checking the consistency of the judgment matrix comprises the following steps: firstly, acquiring a characteristic vector of a judgment matrix, and obtaining a consistency ratio of the judgment matrix based on the judgment matrix and the characteristic vector of the judgment matrix; then, judging whether the consistency of the judgment matrix is qualified or not based on the consistency ratio; and when the judgment is unqualified, reestablishing the judgment matrix, and carrying out consistency check on the reestablished judgment matrix until the established judgment matrix passes the consistency check. For example, the feature vector of the decision matrix is represented as W ═ W1 W2 … Wn]TCalculating the product AW of the judgment matrix A and the eigenvector W, and then classifying the judgment matrix A in the ithiAnd the feature vector WiThe product of (A) can be expressed As (AW)iThen its maximum feature root is
And (3) solving a judgment matrix consistency index CI according to the maximum characteristic root:
then, the table (root) can be looked up according to the average random consistency index RI, the value of RIAnd obtaining a random consistency ratio CR according to the value of n):
when CR is reached<At 0.10, the judgment matrix is considered to have acceptable consistency; and when CR is more than or equal to 0.10, the judgment matrix is not acceptable in consistency, the initial judgment matrix needs to be adjusted and corrected, and the judgment matrix is reestablished.
There are various methods for calculating the eigenvectors of the decision matrix, such as the square root method, the sum-product method, and the power method. The step of solving the feature vector by the square root method may be: firstly, calculating the product of each row of elements of the judgment matrix; then, respectively calculating n times of roots of the products, wherein n is the number of elements in each row of the judgment matrix; and finally, generating an initial vector according to the n-time root obtained by calculation, and carrying out normalization operation on the initial vector to obtain a characteristic vector corresponding to the judgment matrix. When A is ═ aij)n×nRepresenting a judgment matrix, wherein n represents the total number of categories of the feature data, i is greater than or equal to 1 and less than or equal to n, and j is greater than or equal to 1 and less than or equal to n, the above calculation process can be represented as: firstly, calculating the product of each row of elements of the judgment matrix, wherein the expression of the product is as follows:
then calculate MiN times of root:
will vector
Normalized to obtain
The feature vector is then found to be: w ═ W1 W2… n]T。
And carrying out comprehensive judgment analysis by utilizing a maximum membership principle according to the obtained target matrix for representing the quality of the teaching behavior, and determining the evaluation result of the teaching quality.
For the convenience of understanding, it is exemplified that the feature data corresponding to the facial expression, the speech emotion, the head posture, the hand posture, the lower limb posture, and the relative distance between the teacher and the student in table 1 are acquired.
And extracting characteristic data respectively corresponding to facial expressions, voice emotions, head postures, hand postures and lower limb postures representing teaching behaviors and relative distances between a teacher and students from the original audio and video data. Correspondingly, the step of generating the corresponding membership matrix according to the characteristic data comprises the following steps: generating a corresponding first membership matrix according to the feature data corresponding to the facial expression and the voice emotion respectively; generating a corresponding second membership matrix according to the characteristic data corresponding to the head posture, the hand posture and the lower limb posture respectively; and generating a corresponding third membership matrix according to the characteristic data corresponding to the relative distance between the teacher and the student. Obtaining a first matrix for representing emotion expression behavior quality based on the first membership matrix and preset weight matrixes corresponding to the facial expressions and the voice emotions; obtaining a second matrix for representing the behavior quality of the body posture based on the second membership matrix and preset weight matrixes corresponding to the head posture, the hand posture and the lower limb posture; then, obtaining a third matrix for representing the body distance behavior quality based on the third membership matrix and a preset weight matrix corresponding to the relative distance between the teacher and the students; and finally, obtaining a target matrix based on the first matrix, the second matrix and the third matrix. In one embodiment, a fourth matrix including all elements in the first matrix, the second matrix and the third matrix is generated based on the first matrix, the second matrix and the third matrix, a corresponding judgment matrix is obtained according to relative importance of emotion expression, body posture and body distance, and a corresponding weight matrix is obtained. And obtaining a target matrix based on the weight matrix and the fourth matrix.
For ease of understanding, a first matrix R is obtained1A second matrix R2A third matrix R3Wherein R is1＝[0.2 0.5 0.3 0]；R2＝[0.3 0.4 0.3 0]；R3＝[0 0.3 0.5 0.2]. Then the fourth matrix formed based on the three matrices is:
the fourth matrix can be represented by table 4, where the predetermined criteria corresponding to emotional expression, body posture and body distance is V ═ V1,v2,v3,v4Good, better, generally, worse.
TABLE 4
| Judgment set of factor sets | Good taste | Is preferably used | In general | Is poor |
| Expression of emotions | 0.2 | 0.5 | 0.3 | 0 |
| Body posture | 0.3 | 0.4 | 0.3 | 0 |
| Body distance | 0 | 0.3 | 0.5 | 0.2 |
The preset weight matrix corresponding to the emotion expression, the body posture and the body distance is D ═ 0.50.30.2]Namely, the weight value of emotion expression is 0.5, the weight value of body posture is 0.3, and the weight value of body distance is 0.2. Calculating a target matrix by using a preset fuzzy operator
When the blurring operator is the M (,) operator:
when the fuzzy operator is an M (·,) operator:
when the fuzzy operator is
When in operator time:
when the fuzzy operator is
When in operator time:
the results of the four fuzzy operators are analyzed according to the maximum membership principle, and it can be seen that the numerical value of the second data representing 'better' is the largest in the four results, and the emotional expression, the body posture and the body distance are all the characteristic data used for representing the quality of the teaching behavior, so that the quality evaluation result of the teaching behavior can be 'better' based on the three groups of characteristic data.
Optionally, before obtaining the preset weight matrix, a judgment matrix a of emotion expression, body posture and body distance is constructed,
then, the product M of each row of elements of the judgment matrix is calculatediThen M is1＝35；M2＝0.6；M10.05. Recalculating MiOf three roots, i.e.
To obtain
(Vector)
To pair
Normalizing to obtain a characteristic vector W ═ 0.730.190.08]T. The product AW of the decision matrix a and the eigenvector W is calculated,
then its maximum feature root is
And (3) solving a judgment matrix consistency index CI according to the maximum characteristic root:
then, the random consistency ratio CR is calculated:
where RI takes on a value of 0.58, then
Therefore, the decision matrix A is recognized to have acceptable consistency.
Referring to fig. 2, fig. 2 is a data processing apparatus 100 according to an embodiment of the present disclosure, which includes an obtaining module 110 and a processing module 120.
The obtaining module 110 is configured to obtain original audio and video data for teaching by a teacher.
The processing module 120 is configured to extract feature data representing teaching behaviors from the original audio/video data, and generate a corresponding membership matrix according to the feature data, where each row in the membership matrix is used to represent a percentage score between the feature data corresponding to the row and a preset standard.
The processing module 120 is further configured to obtain a target matrix for characterizing quality of the teaching behavior based on the membership matrix and a preset weight matrix corresponding to the feature data.
Optionally, the obtaining module 110 is further configured to extract feature data representing facial expressions, voice emotions, head gestures, hand gestures, lower limb gestures of the teaching behaviors, and relative distances between the teacher and the students from the original audio/video data.
Optionally, the processing module 120 is further configured to generate a corresponding first membership matrix according to the feature data corresponding to the facial expression and the speech emotion respectively; generating a corresponding second membership matrix according to the characteristic data corresponding to the head posture, the hand posture and the lower limb posture respectively; and generating a corresponding third membership matrix according to the characteristic data corresponding to the relative distance between the teacher and the student. Meanwhile, the processing module 120 is further configured to obtain a first matrix for representing emotion expression behavior quality based on the first membership matrix and a preset weight matrix corresponding to the facial expression and the speech emotion; obtaining a second matrix for representing the behavior quality of the body posture based on the second membership matrix and preset weight matrixes corresponding to the head posture, the hand posture and the lower limb posture; and obtaining a third matrix for representing the body distance behavior quality based on the third membership matrix and a preset weight matrix corresponding to the relative distance between the teacher and the student. And obtaining the target matrix based on the first matrix, the second matrix and the third matrix.
Optionally, the processing module 120 is further configured to obtain the target matrix for characterizing the teaching quality based on a preset fuzzy operator and a preset weight matrix corresponding to the membership matrix and the feature data.
Optionally, the processing module 120 is further configured to obtain a judgment matrix for characterizing the importance of the feature data; and acquiring a feature vector corresponding to the judgment matrix, wherein the feature vector is transposed to obtain the preset weight matrix.
Optionally, the processing module 120 is further configured to determine an importance degree of an ith class of feature data to a jth class of feature data in the feature data, where i and j are positive integers, and i and j are not greater than a total number of classes of the feature data; and establishing the judgment matrix based on the importance degree of the ith class of characteristic data to the jth class of characteristic data.
Optionally, the processing module 120 is further configured to obtain a feature vector of the determination matrix, and obtain a consistency ratio of the determination matrix based on the determination matrix and the feature vector of the determination matrix; judging whether the consistency of the judgment matrix is qualified or not based on the consistency ratio; and when the judgment is unqualified, reestablishing the judgment matrix.
Optionally, the processing module 120 is further configured to obtain a product of all elements in each row based on each row of the determination matrix; aiming at the product of each row of elements, obtaining the square root of the product of each row of elements for n times, wherein n is the number of the elements in each row of the judgment matrix; and generating an initial vector according to the obtained n-th-order square root of each row, and performing normalization operation on the initial vector to obtain a feature vector of the judgment matrix.
Referring to fig. 3, fig. 3 is an electronic device 200 according to an embodiment of the disclosure. The electronic device 200 includes: a transceiver 210, a memory 220, a communication bus 230, and a processor 240.
The elements of the transceiver 210, the memory 220, and the processor 240 are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, the components may be electrically coupled to each other via one or more communication buses 230 or signal lines. The transceiver 210 is used for transceiving data. The memory 220 is used for storing a computer program, such as the software functional module shown in fig. 2, i.e., the data processing apparatus 100. The data processing apparatus 100 includes at least one software functional module, which may be stored in the memory 220 in the form of software or firmware (firmware) or solidified in an Operating System (OS) of the electronic device 200. The processor 240 is configured to execute executable modules stored in the memory 220, such as software functional modules or computer programs included in the data processing apparatus 100. For example, original audio and video data for teaching by a teacher are obtained; extracting characteristic data representing teaching behaviors from the original audio and video data, and generating a corresponding membership matrix according to the characteristic data, wherein each row in the membership matrix is used for representing the ratio score of the characteristic data corresponding to the row to a preset standard; and obtaining a target matrix for representing the quality of the teaching behaviors based on the membership matrix and a preset weight matrix corresponding to the characteristic data.
The Memory 220 may be, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Read-Only Memory (EPROM), an electrically Erasable Read-Only Memory (EEPROM), and the like.
The processor 240 may be an integrated circuit chip having signal processing capabilities. The Processor may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but also Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components. The various methods, steps, and logic blocks disclosed in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor 240 may be any conventional processor or the like.
The electronic device 200 includes, but is not limited to, a personal computer, a server, and the like.
The present embodiment also provides a non-volatile computer-readable storage medium (hereinafter, referred to as a storage medium), where the storage medium stores a computer program, and the computer program is executed by the computer, such as the electronic device 200, to execute the data processing method described above.
The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (10)
1. A data processing method is applied to an electronic device, and the method comprises the following steps:
acquiring original audio and video data for teaching of a teacher;
extracting characteristic data representing teaching behaviors from the original audio and video data, and generating a corresponding membership matrix according to the characteristic data, wherein each row in the membership matrix is used for representing the ratio score of the characteristic data corresponding to the row to a preset standard;
and obtaining a target matrix for representing the quality of the teaching behaviors based on the membership matrix and a preset weight matrix corresponding to the characteristic data.
2. The method according to claim 1, wherein extracting feature data representing teaching behaviors from the raw audio-video data comprises:
extracting characteristic data respectively corresponding to facial expressions, voice emotions, head postures, hand postures and lower limb postures representing teaching behaviors and relative distances between a teacher and students from the original audio and video data;
correspondingly, generating a corresponding membership matrix according to the characteristic data, comprising:
generating a corresponding first membership matrix according to the feature data corresponding to the facial expression and the voice emotion respectively; generating a corresponding second membership matrix according to the characteristic data corresponding to the head posture, the hand posture and the lower limb posture respectively; generating a corresponding third membership matrix according to the feature data corresponding to the relative distance between the teacher and the student;
correspondingly, obtaining a target matrix for representing the quality of the teaching behavior based on the membership matrix and the preset weight matrix corresponding to the characteristic data, including:
obtaining a first matrix for representing emotion expression behavior quality based on the first membership matrix and preset weight matrixes corresponding to the facial expressions and the voice emotions; obtaining a second matrix for representing the behavior quality of the body posture based on the second membership matrix and preset weight matrixes corresponding to the head posture, the hand posture and the lower limb posture; obtaining a third matrix for representing the body distance behavior quality based on the third membership matrix and a preset weight matrix corresponding to the relative distance between the teacher and the students;
and obtaining the target matrix based on the first matrix, the second matrix and the third matrix.
3. The method of claim 1, wherein obtaining an evaluation matrix price for characterizing teaching quality based on the membership matrix and a preset weight matrix corresponding to the feature data comprises:
and obtaining the target matrix for representing the teaching quality based on a preset fuzzy operator and the preset weight matrix corresponding to the membership degree matrix and the characteristic data.
4. The method according to claim 2, wherein the preset weight matrix corresponding to the feature data is obtained by:
acquiring a judgment matrix for representing the importance of the characteristic data;
and acquiring a feature vector corresponding to the judgment matrix, wherein the feature vector is transposed to obtain the preset weight matrix.
5. The method of claim 4, wherein obtaining a decision matrix for characterizing the importance of the feature data comprises:
determining the importance degree of ith class of feature data to jth class of feature data in the feature data, wherein i and j are positive integers, and i and j are not more than the total number of classes of the feature data;
and establishing the judgment matrix based on the importance degree of the ith class of characteristic data to the jth class of characteristic data.
6. The method of claim 5, wherein after establishing the decision matrix, the method further comprises:
acquiring a characteristic vector of the judgment matrix, and obtaining a consistency ratio of the judgment matrix based on the judgment matrix and the characteristic vector of the judgment matrix;
judging whether the consistency of the judgment matrix is qualified or not based on the consistency ratio;
and when the judgment is unqualified, reestablishing the judgment matrix.
7. The method of claim 4, wherein obtaining the eigenvector of the decision matrix comprises:
obtaining products of all elements in each row based on each row of the judgment matrix;
aiming at the product of each row of elements, obtaining the square root of the product of each row of elements for n times, wherein n is the number of the elements in each row of the judgment matrix;
and generating an initial vector according to the obtained n-th-order square root of each row, and performing normalization operation on the initial vector to obtain a feature vector of the judgment matrix.
8. A data processing apparatus, comprising:
the acquisition module is used for acquiring original audio and video data for teaching of a teacher;
the processing module is used for extracting feature data representing teaching behaviors from the original audio and video data and generating a corresponding membership matrix according to the feature data, wherein each row in the membership matrix is used for representing the proportion score of the feature data corresponding to the row and a preset standard;
the processing module is further used for obtaining a target matrix for representing the quality of the teaching behaviors based on the membership degree matrix and the preset weight matrix corresponding to the characteristic data.
9. An electronic device, comprising: a memory and a processor, the memory and the processor connected;
the memory is used for storing programs;
the processor to invoke a program stored in the memory to perform the method of any of claims 1-7.
10. A storage medium having stored thereon a computer program which, when executed by a computer, performs the method of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011585656.6A CN112668476B (en) | 2020-12-28 | 2020-12-28 | Data processing method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011585656.6A CN112668476B (en) | 2020-12-28 | 2020-12-28 | Data processing method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112668476A true CN112668476A (en) | 2021-04-16 |
| CN112668476B CN112668476B (en) | 2024-04-16 |
Family
ID=75411464
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011585656.6A Active CN112668476B (en) | 2020-12-28 | 2020-12-28 | Data processing method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112668476B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112990105A (en) * | 2021-04-19 | 2021-06-18 | 北京优幕科技有限责任公司 | Method and device for evaluating user, electronic equipment and storage medium |
| CN114298515A (en) * | 2021-12-22 | 2022-04-08 | 海信集团控股股份有限公司 | Method, device and storage medium for generating student quality portrait |
| CN116665294A (en) * | 2023-03-31 | 2023-08-29 | 华南师范大学 | Teaching gesture and expression application skill analysis method based on computer vision |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3070625A1 (en) * | 2015-03-19 | 2016-09-21 | Accenture Global Services Limited | Image-recognition-based guidance for network device configuration and other enviroments |
| CN108764710A (en) * | 2018-05-25 | 2018-11-06 | 中山大学 | Evaluation Method of Teaching Quality and device |
| WO2019223056A1 (en) * | 2018-05-22 | 2019-11-28 | 深圳市鹰硕技术有限公司 | Gesture recognition-based teaching and learning method and apparatus |
| CN111325486A (en) * | 2020-03-31 | 2020-06-23 | 佛山科学技术学院 | Teacher quality intelligent evaluation method based on analytic hierarchy process and storage medium |
| CN111461089A (en) * | 2020-06-17 | 2020-07-28 | 腾讯科技（深圳）有限公司 | Face detection method, and training method and device of face detection model |
| CN111480348A (en) * | 2017-12-21 | 2020-07-31 | 脸谱公司 | System and method for audio-based augmented reality |
-
2020
- 2020-12-28 CN CN202011585656.6A patent/CN112668476B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3070625A1 (en) * | 2015-03-19 | 2016-09-21 | Accenture Global Services Limited | Image-recognition-based guidance for network device configuration and other enviroments |
| CN111480348A (en) * | 2017-12-21 | 2020-07-31 | 脸谱公司 | System and method for audio-based augmented reality |
| WO2019223056A1 (en) * | 2018-05-22 | 2019-11-28 | 深圳市鹰硕技术有限公司 | Gesture recognition-based teaching and learning method and apparatus |
| CN108764710A (en) * | 2018-05-25 | 2018-11-06 | 中山大学 | Evaluation Method of Teaching Quality and device |
| CN111325486A (en) * | 2020-03-31 | 2020-06-23 | 佛山科学技术学院 | Teacher quality intelligent evaluation method based on analytic hierarchy process and storage medium |
| CN111461089A (en) * | 2020-06-17 | 2020-07-28 | 腾讯科技（深圳）有限公司 | Face detection method, and training method and device of face detection model |
Non-Patent Citations (1)
| Title |
|---|
| 於实;: "模糊层次分析法在教学质量评价中的应用研究", 计算机仿真, no. 06 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112990105A (en) * | 2021-04-19 | 2021-06-18 | 北京优幕科技有限责任公司 | Method and device for evaluating user, electronic equipment and storage medium |
| CN114298515A (en) * | 2021-12-22 | 2022-04-08 | 海信集团控股股份有限公司 | Method, device and storage medium for generating student quality portrait |
| CN116665294A (en) * | 2023-03-31 | 2023-08-29 | 华南师范大学 | Teaching gesture and expression application skill analysis method based on computer vision |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112668476B (en) | 2024-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112668476B (en) | Data processing method and device, electronic equipment and storage medium | |
| CN110222791B (en) | Sample labeling information auditing method and device | |
| Romao et al. | An empirical power comparison of univariate goodness-of-fit tests for normality | |
| CN110009210B (en) | Comprehensive assessment method for student class listening level based on attention degree and concentration degree | |
| CN114091443B (en) | Network information propagation index system construction and evaluation method based on deep learning | |
| CN109003266B (en) | Method for counting image quality subjective evaluation results based on fuzzy clustering | |
| WO2021180249A1 (en) | Occupation recommendation method and apparatus, and device and medium | |
| Schiff et al. | Practical engineering statistics | |
| CN112685324B (en) | Method and system for generating test scheme | |
| CN110457469B (en) | Information classification method, device and computer equipment based on long-term and short-term memory network | |
| CN112836937A (en) | Flood disaster loss evaluation method based on entropy weight and BP neural network technology | |
| Scott | Quantifying certainty in design decisions: examining AHP | |
| CN117079504B (en) | Wrong question data management method of big data accurate teaching and reading system | |
| CN116350203B (en) | Physical testing data processing method and system | |
| Popov et al. | Comparative analysis of various methods treatment expert assessments | |
| CN111369140A (en) | Teaching evaluation system and method | |
| CN109272262B (en) | Method for analyzing natural language features | |
| CN113392086B (en) | Medical database construction method, device and equipment based on Internet of things | |
| CN112598202B (en) | Test question difficulty evaluation method and device, storage medium and computing equipment | |
| CN115129834A (en) | Question answering method and system | |
| Abayomi-Alli et al. | Adaptive regression splines models for predicting facial image verification and quality assessment scores | |
| CN112116264A (en) | Activity evaluation method and apparatus | |
| CN113537759A (en) | User experience measurement model based on weight self-adaptation | |
| CN110069783A (en) | A kind of answer content evaluating method and device | |
| CN117952796B (en) | Reading teaching quality assessment method and system based on data analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |