CN111476352A - Deep learning-based method for analyzing effectiveness of code scanning and evaluation data in hall - Google Patents
Deep learning-based method for analyzing effectiveness of code scanning and evaluation data in hall Download PDFInfo
- Publication number
- CN111476352A CN111476352A CN202010258633.8A CN202010258633A CN111476352A CN 111476352 A CN111476352 A CN 111476352A CN 202010258633 A CN202010258633 A CN 202010258633A CN 111476352 A CN111476352 A CN 111476352A
- Authority
- CN
- China
- Prior art keywords
- evaluation
- data
- student
- education
- 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
- 238000011156 evaluation Methods 0.000 title claims abstract description 151
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000013135 deep learning Methods 0.000 title claims abstract description 25
- 238000012549 training Methods 0.000 claims abstract description 41
- 238000003062 neural network model Methods 0.000 claims abstract description 31
- 238000012360 testing method Methods 0.000 claims abstract description 25
- 238000013528 artificial neural network Methods 0.000 claims abstract description 23
- 230000000694 effects Effects 0.000 claims abstract description 19
- 238000012216 screening Methods 0.000 claims abstract description 11
- 238000007781 pre-processing Methods 0.000 claims abstract description 8
- 238000012795 verification Methods 0.000 claims abstract description 8
- 238000004458 analytical method Methods 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 3
- 230000008439 repair process Effects 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 238000010200 validation analysis Methods 0.000 claims description 2
- 238000007405 data analysis Methods 0.000 abstract description 8
- 238000010801 machine learning Methods 0.000 description 9
- 230000006870 function Effects 0.000 description 4
- 238000012552 review Methods 0.000 description 3
- 238000011478 gradient descent method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 244000207740 Lemna minor Species 0.000 description 1
- 235000006439 Lemna minor Nutrition 0.000 description 1
- 235000001855 Portulaca oleracea Nutrition 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000011157 data evaluation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000008451 emotion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/20—Education
- G06Q50/205—Education administration or guidance
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Data Mining & Analysis (AREA)
- Educational Administration (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Strategic Management (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Tourism & Hospitality (AREA)
- Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Educational Technology (AREA)
- Development Economics (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- General Business, Economics & Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Primary Health Care (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Electrically Operated Instructional Devices (AREA)
Abstract
The invention provides a deep learning-based method for analyzing effectiveness of parade code scanning evaluation data, which comprises the following steps of: collecting the code scanning evaluation and education data along with the hall to obtain an evaluation and education data set; preprocessing the evaluation and education data set to obtain a marked evaluation and education data set, and dividing the marked evaluation and education data set into a training set, a verification set and a test set; training the deep neural network by using a training set to obtain a plurality of candidate deep neural network models; and testing the candidate deep neural network models by using the test set, and finally screening out the deep neural network model with the best effect. The invention innovatively provides a new idea of applying a deep learning algorithm to evaluation and education data analysis, the deep learning is applied to the judgment problem of the effectiveness of evaluation and education data for the first time, and deep characteristics related to the effectiveness of the evaluation and education data and self attributes of teachers, courses and students are learned by means of a deep neural network.
Description
Technical Field
The invention relates to the field of big data analysis and pattern recognition, in particular to an in-class code scanning evaluation and education data validity analysis method based on deep learning.
Background
The teaching quality is a life line cultivated by talents in colleges and universities, and in order to establish a perfect teaching quality supervision system, various teaching quality evaluation means are adopted to supervise the teaching quality of teachers in domestic and overseas colleges and universities. The most common method is that students evaluate the teaching quality of teachers, so that problems in teaching work can be found and solved in time, and the method plays an important role in improving the work of teachers, exciting the learning enthusiasm of students, improving the effect of teaching work and the like. However, in the course of student's assessment and education, there are cases of spreading the forms of the events or being influenced by subjective emotions such as personal complaints, so that the objectivity of the assessment and education results is difficult to be guaranteed, which leads to disputes in the industry whether the assessment and education results can truly reflect the teaching level of teachers. Especially, nowadays, more and more colleges and universities begin to award and punish teachers with reference to student evaluation results, the lack of fair and fair evaluation results can generate non-negligible negative influence on the working enthusiasm of teachers.
In the prior most of teaching assessment methods, the focus is only put on collected assessment data of students, but teachers, courses and students as three main subjects in teaching activities all have influence on the effect of the teaching activities, so in order to obtain more objective teaching assessment results, relevant data of the teachers and the courses are also taken into the process of analysis of assessment data, yellow wave of the financial institute of the former Hitachi institute of China is in the document ' yellow wave, student assessment influence factor research is based on empirical analysis [ J ] of sequencing L logic/Probit regression, high-grade finance education research 2014(4) ' 1-8 ', comprehensive application descriptive statistics, correlation analysis, stepwise regression and sequencing L logic/Probit regression analysis methods are researched, influence of characteristics of the teachers and the courses on assessment results is researched, and evaluation basic information of the teachers, workload of the teachers in each period, performance factors of the courses, positive and negative evaluation relations of students and students on the assessment results, and positive and negative evaluation relations of the teachers on the students and students in any stage are respectively selected in an algorithm.
Some methods have been developed in recent years for analysis of assessment and education data by machine learning. Machine learning is a data-driven decision-making method, which is a form of artificial intelligence, and in a practical sense, machine learning is a method of training a model using data and then predicting using the model. The method is characterized in that an analysis method of evaluation and education data of students of independent colleges based on clustering is adopted in ' an innocent duckweed of Jia-G college of Xiamen university ' P ': CN108256102A,2018-07-06 ", adopts fuzzy K-models algorithm in machine learning algorithm to analyze data of evaluation and education data table. Deep learning is an algorithm for performing characterization learning on data in machine learning, other machine learning algorithms need to use manually designed features to train a model, and the deep learning algorithm can automatically extract low-level or high-level features which are helpful for solving problems from the data. Generally, under the condition of sufficient data quantity, the deep learning algorithm can perform better than other traditional machine learning algorithms.
In summary, the existing methods can be improved by:
1. most evaluation and education analysis methods do not consider the justice rationality of students in evaluating and teaching scores, and the use of unfair evaluation and education data can affect the objectivity of finally obtained teaching evaluation results.
2. The influence of the teacher, the course and the students on the evaluation and teaching result is not comprehensively considered when analyzing the evaluation and teaching data.
3. When the machine learning is applied to evaluation and education data analysis, traditional machine learning algorithms such as clustering or support vector machines are used, and no attempt is made to apply a deep learning algorithm with excellent performance in big data analysis to the relevant problems of evaluation and education data analysis.
Disclosure of Invention
The invention provides an assessment and teaching data effectiveness analysis method based on deep learning, aiming at overcoming the technical defect that the justice and reasonability of assessment and teaching activities are influenced by invalid assessment and teaching data generated by assessment and teaching students under the conditions of crowd psychology, subjective prejudice or intervention of teaching teachers in assessment and teaching and the like in the use process of the conventional assessment and teaching system.
In order to solve the technical problems, the technical scheme of the invention is as follows:
the method for analyzing the effectiveness of the code scanning and evaluation data in the hall based on deep learning comprises the following steps:
s1: collecting along with the code scanning evaluation and teaching data, collecting the evaluation and teaching data according to the along with the code scanning evaluation and teaching system, and carrying out quantitative operation and combination on each evaluation and teaching data and corresponding teacher, course and student information to obtain an unmarked evaluation and teaching data set;
s2: preprocessing an evaluation and education data set, namely firstly marking each piece of evaluation and education data in an unmarked evaluation and education data set, and then dividing the evaluation and education data set into a training set, a verification set and a test set;
s3: training parameters of the deep neural network model according to the obtained training data set, judging whether the model is over-fitted by using a verification set in the training process, and performing multiple training by using multiple groups of different hyper-parameters to obtain multiple candidate deep neural network models;
s4: and screening the deep neural network model, testing the effect of each candidate deep neural network model by using the test data set, screening the deep neural network model with the optimal effect, and carrying out effectiveness analysis on the evaluation and education data.
Wherein, the step S1 specifically includes the following steps:
s11: in the class, a student scans an assessment and education two-dimensional code by using intelligent mobile equipment to open an assessment and education Web page, and logs in by using an account of a teaching system of the student;
s12: the filled teaching evaluation questionnaire comprises 10 selection questions, and options of each selection question comprise four grades of poor, general, good and excellent;
s13: after the students finish filling and submit a questionnaire for evaluation and education, the server carries out quantitative operation on teacher information, course information and student information which are inquired from the educational administration system and are inquired into the questionnaire, and results of each question in the questionnaire for evaluation and education are spliced to obtain evaluation and education data; and (4) processing all the evaluation and education questionnaires to obtain all the evaluation and education data, namely the unmarked evaluation and education data set.
Each evaluation and education data format of the evaluation and education data set is as follows:
{ UID, course ID, course category, course attendance, teacher's teaching age, teacher's gender, student's age, student's gender, student's department, student's specialty, average student's performance in mandatory lessons, average student's performance in optional lessons, average number of courses taken by students per school period, number of courses taken by students in current school period, student's attendance, student's psychological assessment performance, assessment 1 result, assessment 2 result, … …, assessment 10 result, validity } contain 27 pieces of information in total, wherein: the value of the "validity" field defaults to 0.
Wherein, the step S2 specifically includes the following steps:
s21: for each course, the specific operation is as follows:
assuming that the course ID is 1, randomly selecting n more than or equal to 5 teachers and m more than or equal to 5 students from the teachers and students in the whole school, and requiring each person to listen to the course for at least 4 classes, wherein the m students should be extracted from the students who select and repair the course; filling out a teaching evaluation questionnaire of the course by the m + n individuals to obtain m + n evaluation samples, and recording the evaluation samples as Set _ Standard, wherein the format of each evaluation sample is as follows:
{ course ID, score 1 results, score 2 results, … …, score 10 results }
Screening all the evaluation data with the course ID of 1 in the evaluation data Set, recording the evaluation data as Set _ unabled, assuming that the Set _ unabled has N pieces of evaluation data, and recording each piece of evaluation data as Ui(1<＝i<N), and each of Set _ Standard evaluates sample Sj(1<＝j<M + n) for comparison; if S is presentjSo that U isiAnd SjIf at least 4 of the 10 assessment questions are the same, the assessment data are concentrated into UiUpdating the value of the 'validity' field of the corresponding record to be 1;
s22: after step S21 is performed for each course, the labeling of each piece of assessment data in the assessment data set is completed, and then 60%, 20%, and 20% are randomly extracted from the assessment data set, which are training set, validation set, and test set, respectively.
In step S3, the deep neural network is composed of an upper part and a lower part, each of which includes 8 layers, wherein 5 layers are convolutional layers and 3 layers are fully connected layers.
Wherein, the step S3 specifically includes the following steps:
s31: before each piece of data in the evaluation and education data set is input into the deep neural network, the data is divided into data and label, wherein:
the data is a two-dimensional matrix of 5 × 5 converted from { course ID, course category, average course attendance rate, teacher teaching age, teacher gender, student age, student gender, student department, student specialty, average student grade, average number of courses taken by students in each period, number of courses taken by students in the current period, student attendance rate, student psychological assessment grade, assessment and education question 1 result, assessment and education question 2 result, … … and assessment and education question 10 result }; label is the value of the "validity" field;
s32: training on a train _ set data set, training parameters of a deep neural network by means of a BP back propagation algorithm and a gradient descent algorithm, verifying a set valid _ set to judge whether the model is over-fitted or not in the training process, and obtaining 1 deep neural network model after training is finished;
s33: step S32 is repeated 10 times using 10 different sets of hyper-parameters to obtain 10 candidate deep neural network models.
Wherein the hyper-parameters comprise a training iteration number epoch and a learning rate learn _ rate.
Wherein, the step S4 specifically includes:
and (3) testing and recording the effects of the 10 deep neural networks obtained in the step (S3) on a test _ set data set, selecting and storing a deep neural network model with the best effect, wherein the deep neural network model can be used for identifying the effectiveness of the evaluation and teaching data, and after a piece of evaluation and teaching data is input into the model, the model outputs a prediction result whether the evaluation and teaching data is effective.
Compared with the prior art, the technical scheme of the invention has the beneficial effects that:
the method for analyzing the effectiveness of the parade code scanning evaluation data based on deep learning provided by the invention focuses on judging the effectiveness of the evaluation data, so that the finally obtained teaching evaluation result on the judged effective evaluation data is more fair.
The invention fully utilizes the Web technology to realize the data collected by the in-house code scanning evaluation and teaching system, provides a method for preprocessing the collected evaluation and teaching data, and obtains an evaluation and teaching data set which can be used for training a deep neural network after preprocessing. And screening out attributes (such as teacher teaching age, course type, student attendance and the like) which have great influence on the effectiveness of the teaching data evaluation from the teacher attributes, the courses and the students attributes according to practical experience.
The deep learning algorithm is used for solving the two classification problems of judging the effectiveness of the evaluation and teaching data for the first time, a new idea of applying the deep learning algorithm to the evaluation and teaching data analysis is innovatively provided, a double-path deep convolution neural network architecture is provided, and a deep neural network model capable of judging the effectiveness of the evaluation and teaching data is finally obtained after training is carried out by means of an evaluation and teaching data set.
Drawings
FIG. 1 is a flow chart of the method steps of the present invention;
FIG. 2 is a flowchart of the method for collecting the code scanning and evaluation data in the embodiment of the present invention;
FIG. 3 is a flow chart of pre-processing of an assessment dataset according to an embodiment of the present invention;
FIG. 4 is a diagram of a deep neural network architecture used in an embodiment of the present invention.
Detailed Description
The drawings are for illustrative purposes only and are not to be construed as limiting the patent;
for the purpose of better illustrating the embodiments, certain features of the drawings may be omitted, enlarged or reduced, and do not represent the size of an actual product;
it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
The technical solution of the present invention is further described below with reference to the accompanying drawings and examples.
Example 1
As shown in fig. 1, the method for analyzing the effectiveness of the code scanning evaluation data based on deep learning comprises the following steps:
s1: collecting along with the code scanning evaluation and teaching data, collecting the evaluation and teaching data according to the along with the code scanning evaluation and teaching system, and carrying out quantitative operation and combination on each evaluation and teaching data and corresponding teacher, course and student information to obtain an unmarked evaluation and teaching data set;
s2: preprocessing an evaluation and education data set, namely firstly marking each piece of evaluation and education data in an unmarked evaluation and education data set, and then dividing the evaluation and education data set into a training set, a verification set and a test set;
s3: training parameters of the deep neural network model according to the obtained training data set, judging whether the model is over-fitted by using a verification set in the training process, and performing multiple training by using multiple groups of different hyper-parameters to obtain multiple candidate deep neural network models;
s4: and screening the deep neural network model, testing the effect of each candidate neural network by using the test data set, screening the deep neural network model with the optimal effect, and carrying out effectiveness analysis on the evaluation and education data.
In a specific implementation process, the method for analyzing the effectiveness of the code scanning and evaluating data based on deep learning fully utilizes the Web technology to realize the data collected by the code scanning and evaluating system, and comprehensively considers the influence of the attributes carried by a teacher, a course and a student on the effectiveness of the evaluating and teaching data in teaching activities; deep features are extracted from attributes of teachers, courses and students by utilizing a deep neural network, and effectiveness analysis of evaluation and education data is carried out through the features, so that objectivity of evaluation and education results is guaranteed.
Example 2
More specifically, as shown in fig. 2, the results of each review and education questionnaire uploaded by the student are converted and combined with the teacher information, the course information and the student information to which the student belongs on the review and education system Web server to obtain a review and education data set.
The method comprises the following steps of filling out an evaluation and education questionnaire and uploading evaluation and education results for students:
i. in the class room, the students scan the assessment and education two-dimensional codes by using the intelligent mobile device to open the assessment and education Web pages and log in by using the teaching system account of the students.
The filled-in teaching evaluation questionnaire contains 10 choice questions, and each choice of the choice questions is designed into four grades of { poor, general, good, and excellent }.
And iii, after the students fill in the teaching evaluation questionnaire and confirm that the teaching evaluation questionnaire is correct, clicking a confirmation button, and uploading the teaching evaluation result to a Web server of the teaching evaluation system.
The data requested by the Web server of the assessment and education system to the database of the school education system comprises the following data: course ID, course category, course attendance rate, teacher's teaching age, teacher's gender, student's age, student's gender, student's department, student's specialty, average student's performance in selecting course, average number of courses taken by students per period, number of courses taken by students at current period, student's attendance rate, and student's psychological assessment performance. The detailed description is shown in the following table.
Detailed data introduction table
In the specific implementation process, the evaluation and teaching system Web server converts teacher information, course information and student information into digital texts as shown in the table above, and after splicing, an unlabeled evaluation and teaching data set is obtained, wherein each evaluation and teaching data format is as follows:
{ UID, course ID, course category, course attendance, teacher's teaching age, teacher's gender, student's age, student's gender, student's department, student's specialty, average student's performance in required courses, average student's performance in selected courses, average number of courses taken by students per session, number of courses taken by students at current session, student's attendance, student's psychological assessment performance, assessment 1 result, assessment 2 result, … …, assessment 10 result, validity }
The UID is a self-increment field with uniqueness and is used as a main key of the database; the value of the "validity" field defaults to 0; and finally, storing the evaluation and education data set into a database of the data analysis server.
In a specific implementation process, a flow of pre-processing the assessment and education data is shown in fig. 3, and in the data analysis server, the validity field of each record in the assessment and education data set obtained in the step (1) is updated, and after the update, the whole assessment and education data set is divided into a training set, a verification set and a test set.
The steps of updating the validity field are as follows:
i. for each course, the following operations are carried out:
assuming that the course ID is 1, randomly selecting n more than or equal to 5 teachers and m more than or equal to 5 students from the teachers and students in the whole school, and requiring each person to listen to the course for at least 4 classes, wherein the m students should be extracted from the students who select and repair the course; filling out a teaching evaluation questionnaire of the course by the m + n individuals to obtain m + n evaluation samples, and recording the evaluation samples as Set _ Standard, wherein the format of each evaluation sample is as follows:
{ course ID, score 1 results, score 2 results, … …, score 10 results }
Screening all the evaluation data with the course ID of 1 in the evaluation data Set, marking as Set _ Unlabled, and comparing each piece of evaluation data as Ui (1< ═ i < ═ N) and each evaluation sample Sj (1< ═ j < ═ m + N) in Set _ Standard, assuming that the Set _ Unlabled has N pieces of evaluation data; if Sj exists, at least 4 results of 10 evaluation questions of Ui and Sj are correspondingly the same, updating the value of the 'validity' field of the corresponding evaluation data of Ui in the evaluation data set to 1;
after step i is performed on each course, marking of each piece of assessment data in the assessment data set is completed, and then 60%, 20% and 20% of the assessment data set are randomly extracted, namely training set train _ set, verification set valid _ set and test set test _ set.
(3) And (3) training the deep neural network, constructing the deep neural network under a Pythrch framework, training parameters of the deep neural network by using the training set train _ set obtained in the step (2) through a BP back propagation algorithm and a gradient descent method, and judging whether the model is over-fitted or not by using valid _ set in the training process. And repeating the training for 10 times by using different hyper-parameters to obtain 10 candidate deep neural networks.
i. The deep neural network is composed of an upper part and a lower part, each part comprises 8 layers, wherein 5 layers are convolution layers, and 3 layers are full-connection layers. As shown in fig. 4.
Before inputting the data in the train _ set and valid _ set data sets into the deep neural network, dividing the data into two parts of data and label, wherein the data is a two-dimensional matrix of 5 × 5 processed by { course ID, course category, course average attendance rate, teacher teaching age, teacher gender, student age, student gender, student department, student specialty, student mandatory lesson average score, student optional lesson average score, student average number of lessons repaired per period, student current period number of lessons repaired, student attendance rate, student psychological assessment score, assessment 1 result, assessment 2 result, … …, assessment 10 result }; label is the value of the "validity" field.
The loss function used when training the neural network on its parameters by means of the BP back-propagation algorithm and the gradient descent method is a "0-1 loss function", which is defined as follows:
of these, L (x)1,x2) Representing a loss function, parameter x1、x2Is the input of a function; (x) is the result (0: invalid; 1: valid) of predicting the validity of the input evaluation data by the neural network; y represents the actual validity of the evaluation data X (0: invalid; 1: valid).
(4) Testing the deep neural network model, testing the effect of the 10 candidate deep neural network models obtained in the step (3) on the test data set test _ set obtained in the step (2), and selecting the deep neural network model with the best effect according to Precision (Precision), wherein the Precision is defined as follows:
the finally obtained deep neural network model can be used for judging the effectiveness of the evaluation data.
It should be understood that the above-described embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the claims of the present invention.
Claims (8)
1. Deep learning-based method for analyzing effectiveness of code scanning and evaluation data in the hall is characterized by comprising the following steps: the method comprises the following steps:
s1: collecting along with the code scanning evaluation and teaching data, collecting the evaluation and teaching data according to the along with the code scanning evaluation and teaching system, and carrying out quantitative operation and combination on each evaluation and teaching data and corresponding teacher, course and student information to obtain an unmarked evaluation and teaching data set;
s2: preprocessing a teaching data set, firstly marking each piece of evaluation and teaching data in an unmarked evaluation and teaching data set, and then dividing the evaluation and teaching data set into a training set, a verification set and a test set;
s3: training a deep neural network model, training parameters of the deep neural network by using data in a training set, and performing multiple training by using multiple groups of different hyper-parameters to obtain multiple candidate deep neural network models;
s4: and screening the deep neural network model, testing the effect of each candidate deep neural network model by using the test data set, screening the deep neural network model with the optimal effect, and carrying out effectiveness analysis on the evaluation and education data.
2. The deep learning-based method for analyzing effectiveness of code scanning and evaluation data in an auditorium according to claim 1, wherein: the step S1 specifically includes the following steps:
s11: in the class, a student scans an assessment and education two-dimensional code by using intelligent mobile equipment to open an assessment and education Web page, and logs in by using an account of a teaching system of the student;
s12: the filled teaching evaluation questionnaire comprises 10 selection questions, and options of each selection question comprise four grades of poor, general, good and excellent;
s13: after the students finish filling and submit a questionnaire for evaluation and education, the server carries out quantitative operation on teacher information, course information and student information which are inquired from the educational administration system and are inquired into the questionnaire, and results of each question in the questionnaire for evaluation and education are spliced to obtain evaluation and education data; and (4) processing all the evaluation and education questionnaires to obtain all the evaluation and education data, namely the unmarked evaluation and education data set.
3. The deep learning-based method for analyzing effectiveness of the code scanning and evaluation data in the hall according to claim 2, wherein: the evaluation and education data set comprises the following evaluation and education data formats:
{ UID, course ID, course category, course attendance, teacher's teaching age, teacher's gender, student's age, student's gender, student's department, student's specialty, average student's performance in mandatory lessons, average student's performance in optional lessons, average number of courses taken by students per school period, number of courses taken by students in current school period, student's attendance, student's psychological assessment performance, assessment 1 result, assessment 2 result, … …, assessment 10 result, validity } contain 27 pieces of information in total, wherein: the value of the "validity" field defaults to 0.
4. The deep learning-based method for analyzing effectiveness of the code scanning and evaluation data in the hall according to claim 2, wherein: the step S2 specifically includes the following steps:
s21: for each course, the specific operation is as follows:
assuming that the course ID is 1, randomly selecting n more than or equal to 5 teachers and m more than or equal to 5 students from the teachers and students in the whole school, and requiring each person to listen to the course for at least 4 classes, wherein the m students should be extracted from the students who select and repair the course; filling out a teaching evaluation questionnaire of the course by the m + n individuals to obtain m + n evaluation samples, and recording the evaluation samples as Set _ Standard, wherein the format of each evaluation sample is as follows:
{ course ID, score 1 results, score 2 results, … …, score 10 results }
Screening all the evaluation data with the course ID of 1 in the evaluation data Set, recording the evaluation data as Set _ unabled, and recording each piece of evaluation data as U assuming that the Set _ unabled has N pieces of evaluation datai(1<＝i<N), and each of Set _ Standard evaluates sample Sj(1<＝j<M + n) for comparison; if S is presentjSo that U isiAnd Sj10 scores ofIf at least 4 of the results of the education questions are corresponding to the same, the evaluation data is concentrated into UiThe value of the 'validity' field corresponding to the evaluation and education data is updated to be 1;
s22: after step S21 is performed for each course, the labeling of each piece of assessment data in the assessment data set is completed, and then 60%, 20%, and 20% are randomly extracted from the assessment data set, which are training set, validation set, and test set, respectively.
5. The deep learning-based method for analyzing effectiveness of code scanning and evaluation data in an auditorium according to claim 4, wherein: in step S3, the deep neural network is composed of an upper part and a lower part, each of which comprises 8 layers, wherein 5 layers are convolutional layers and 3 layers are fully connected layers.
6. The deep learning-based method for analyzing effectiveness of code scanning and evaluation data in an auditorium according to claim 5, wherein: the step S3 specifically includes the following steps:
s31: before each piece of data in the evaluation and education data set is input into the deep neural network, the data is divided into data and label, wherein:
the data is a two-dimensional matrix of 5 × 5 converted from { course ID, course category, average course attendance rate, teacher teaching age, teacher gender, student age, student gender, student department, student specialty, average student grade, average number of courses taken by students in each period, number of courses taken by students in the current period, student attendance rate, student psychological assessment grade, assessment and education question 1 result, assessment and education question 2 result, … … and assessment and education question 10 result }; label is the value of the "validity" field;
s32: training on a train _ set data set, training parameters of a deep neural network by means of a BP back propagation algorithm and a gradient descent algorithm, verifying a set valid _ set to judge whether the model is over-fitted or not in the training process, and obtaining 1 deep neural network model after training is finished;
s33: step S32 is repeated 10 times using 10 different sets of hyper-parameters to obtain 10 candidate deep neural network models.
7. The deep learning-based method for analyzing effectiveness of the code scanning and evaluation data in the assembly hall according to claim 6, wherein: the hyper-parameters include the number of training iterations epoch and the learning rate learn _ rate.
8. The deep learning-based method for analyzing effectiveness of the code scanning and evaluation data in the hall according to claim 7, wherein: the step S4 specifically includes:
and (3) testing and recording the effects of the 10 deep neural network models obtained in the step (S3) on a test _ set data set, selecting and storing the deep neural network model with the best effect, wherein the deep neural network model can be used for identifying the effectiveness of the evaluation and teaching data, and after a piece of evaluation and teaching data is input into the model, the model outputs a prediction result whether the evaluation and teaching data is effective.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010258633.8A CN111476352B (en) | 2020-04-03 | 2020-04-03 | Method for analyzing effectiveness of along-with-hall code scanning evaluation and teaching data based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010258633.8A CN111476352B (en) | 2020-04-03 | 2020-04-03 | Method for analyzing effectiveness of along-with-hall code scanning evaluation and teaching data based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111476352A true CN111476352A (en) | 2020-07-31 |
| CN111476352B CN111476352B (en) | 2023-04-28 |
Family
ID=71749809
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010258633.8A Active CN111476352B (en) | 2020-04-03 | 2020-04-03 | Method for analyzing effectiveness of along-with-hall code scanning evaluation and teaching data based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111476352B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113822558A (en) * | 2021-09-09 | 2021-12-21 | 电子科技大学 | Student evaluation method and system based on AI identification data |
| CN117350902A (en) * | 2023-09-28 | 2024-01-05 | 广州铭智教育科技有限公司 | On-line training time supervision system, method and storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015128700A1 (en) * | 2014-02-25 | 2015-09-03 | Al Midfa Khalid Omar | Electronic educational content mapping and learning system and a method of ensuring quality teaching and learning effectiveness |
-
2020
- 2020-04-03 CN CN202010258633.8A patent/CN111476352B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015128700A1 (en) * | 2014-02-25 | 2015-09-03 | Al Midfa Khalid Omar | Electronic educational content mapping and learning system and a method of ensuring quality teaching and learning effectiveness |
Non-Patent Citations (1)
| Title |
|---|
| 武彤;王秀坤;: "基于BP神经网络的教学质量分析评估模型研究" * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113822558A (en) * | 2021-09-09 | 2021-12-21 | 电子科技大学 | Student evaluation method and system based on AI identification data |
| CN113822558B (en) * | 2021-09-09 | 2023-04-07 | 电子科技大学 | Student evaluation method and system based on AI identification data |
| CN117350902A (en) * | 2023-09-28 | 2024-01-05 | 广州铭智教育科技有限公司 | On-line training time supervision system, method and storage medium |
| CN117350902B (en) * | 2023-09-28 | 2024-04-09 | 广州铭智教育科技有限公司 | On-line training time supervision system, method and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111476352B (en) | 2023-04-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112069970B (en) | Classroom teaching event analysis method and device | |
| CN111143750B (en) | University computer class course management platform based on fragmented learning | |
| KR20110079252A (en) | Study management system for online lecture, and method for the same | |
| CN114820248A (en) | Work interpolation method based on student ability portrait | |
| CN111476352B (en) | Method for analyzing effectiveness of along-with-hall code scanning evaluation and teaching data based on deep learning | |
| CN114037569A (en) | Artificial intelligence-based multi-scene two-way simulation internet medical customer service personnel training method | |
| Bussaman et al. | Prediction models of learning strategies and learning achievement for lifelong learning | |
| Wang et al. | Results and insights from diagnostic questions: The neurips 2020 education challenge | |
| Williamson | Digitizing Education Governance: Pearson, Real-Time Data Analytics, Visualization and Machine Intelligence | |
| Ramadani et al. | Digital entrepreneurship intentions: evidence from Kosovo | |
| Srivastava et al. | An analysis of various training evaluation models | |
| Anzer et al. | Predicting academic performance of students in UAE using data mining techniques | |
| Stefanova et al. | Educational data mining perspectives within university big data environment | |
| Hu et al. | A multivariate learning evaluation model for programming course in online learning environment | |
| CN116883208A (en) | Classroom quality assessment system based on images | |
| Mostoli et al. | Using the Malmquist Index in evaluation process to enhance mathematical literacy in high school students | |
| Angra et al. | Analysis of Student's Data Using Rapid Miner | |
| Smith | Integrative literature review: Methods in identifying exemplary performance: A review of the literature and implications for HRD | |
| Olaguer et al. | Total quality management of public elementary school principals in relation to school performance | |
| Hrich et al. | ARTIFICIAL INTELLIGENCE FOR EDUCATIONAL ASSESSMENT | |
| Ogunkunle et al. | A data mining based optimization of selecting learning material in an intelligent tutoring system for advancing STEM education | |
| KR100847020B1 (en) | Method for managing learning and business and Apparatus thereof | |
| Petersen et al. | Preliminary Evaluation of a Survey Checklist in the Context of Evidence-based Software Engineering Education. | |
| BEKTENOVA et al. | FORMATION OF INDIVIDUAL TRAJECTORIES OF GIFTEDNESS OF STUDENTS BASED ON THE ANALYSIS OF LARGE DATA ARRAYS | |
| Ranjan et al. | Data analysis to uncover critical parameters for performance appraisal of students in higher education |
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 |