CN-122022580-A - Intelligent teacher teaching practice quality evaluation method based on extreme learning machine algorithm

CN122022580ACN 122022580 ACN122022580 ACN 122022580ACN-122022580-A

Abstract

The invention is suitable for the technical field of teaching evaluation, and provides an intelligent teaching practice quality evaluation method for teachers based on an extreme learning machine algorithm, which comprises the steps of designing and issuing a questionnaire based on an evaluation index system, collecting samples and constructing a sample database containing a plurality of influence factors and comprehensive evaluation results; the method comprises the steps of checking the credibility and the effectiveness of questionnaire data, removing invalid and abnormal samples, adopting decision tree feature importance sorting to perform feature screening, combining mutual information analysis and correlation threshold constraint to achieve redundancy elimination, obtaining an optimal feature subset, dividing the data set into a training set and a testing set, determining model parameters in the training set through layered K-fold cross validation, establishing an extreme learning machine classification model, and evaluating the model performance through accuracy, precision, recall, confusion matrix and Kappa coefficient. The quantitative evaluation and decision feedback of the intelligent teaching practice quality of the college teacher can be realized, and a reusable method support is provided for the similar quality evaluation.

Inventors

XIA GUOPING
ZHAO LIPING
LIU MENGQI

Assignees

阜阳师范大学

Dates

Publication Date: 20260512
Application Date: 20260130

Claims (10)

1. The intelligent teaching practice quality evaluation method for the teacher based on the extreme learning machine algorithm is characterized by comprising the following steps of: designing and issuing a questionnaire based on an intelligent teaching practice quality evaluation system of a college teacher, collecting influence factor index data, generating a comprehensive score and a corresponding quality grade label for each sample according to the evaluation system, and constructing a sample database; Performing reliability and effectiveness test on the obtained questionnaire data, and removing or correcting invalid questions, abnormal questionnaires and abnormal samples to obtain a trusted data set; the method comprises the steps of respectively carrying out mutual information sorting, sorting based on feature importance degree of a decision tree and sorting based on correlation redundancy degree, synthesizing various sorting results to obtain feature comprehensive sorting, eliminating high redundancy features under the constraint of a preset correlation threshold, selecting K features with the front comprehensive sorting as model input feature subsets, wherein K is a positive integer or is determined by cross verification; Searching super parameters of the extreme learning machine model by adopting a cross verification and grid optimizing method on the training set, determining optimal parameters, and establishing an extreme learning machine classification model, wherein the super parameters at least comprise hidden layer node numbers and activation function types; And inputting actual teaching process data into the model, outputting an intelligent teaching practice quality evaluation result of a college teacher, and using the intelligent teaching practice quality evaluation result for teaching improvement decision feedback.
2. The method for evaluating the intelligent teaching practice quality of the teacher based on the extreme learning machine algorithm according to claim 1, wherein the system for evaluating the intelligent teaching practice quality of the teacher in the college comprises five influencing factors including a subject factor, a guest factor, a technical factor, an environmental factor and a policy factor, and the sample database comprises a plurality of index features.
3. The method for evaluating the intelligent teaching practice quality of the teacher based on the extreme learning machine algorithm according to claim 1, wherein basic information of an investigation object is acquired at the same time of acquiring evaluation index data for sample layering processing and data consistency verification.
4. The method for evaluating the intelligent teaching practice quality of the teacher based on the extreme learning machine algorithm according to claim 1 is characterized in that the reliability test adopts the clenbach alpha coefficient, and the investigation objects comprise college students, college teachers and related scientific researchers, are related with the evaluation content and are representative.
5. The method for evaluating the intelligent teaching practice quality of the teacher based on the extreme learning machine algorithm according to claim 1, wherein the correlation redundancy ranking is calculated based on correlation coefficients among features, and when the absolute value of the correlation coefficient between two features is not smaller than a preset threshold value, the high redundancy features are judged and eliminated.
6. The method for evaluating the quality of intelligent teaching practice of a teacher based on an extreme learning machine algorithm according to claim 1, wherein the quality grade label is obtained by dividing and encoding the composite score into discrete categories according to preset intervals, wherein the composite score intervals [60,70 ], [70,80 ], [80,90 ], [90,100] respectively correspond to the quality grade D, C, B, A.
7. The method for evaluating the intelligent teaching practice quality of the teacher based on the extreme learning machine algorithm according to claim 1, wherein the cross verification is 5-fold layered cross verification and is used for model performance evaluation and parameter selection.
8. The method for evaluating the intelligent teaching practice quality of the teacher based on the extreme learning machine algorithm according to claim 1, wherein in the model parameter optimizing process, the average error rate of quality level prediction under cross validation is used as an adaptability evaluation index of the extreme learning machine classification model.
9. The method for evaluating the intelligent teaching practice quality of the teacher based on the extreme learning machine algorithm according to claim 1, wherein an activation function of the extreme learning machine classification model is a Sigmoid function, a candidate value range of the hidden layer node number is 10-200, and an optimal value is determined through grid search.
10. The method for evaluating the intelligent teaching practice quality of the teacher based on the extreme learning machine algorithm according to claim 1, wherein the prediction performance of the model is evaluated by one or more indexes of accuracy, precision, recall, F1 score, confusion matrix and Kappa coefficient.

Description

Intelligent teacher teaching practice quality evaluation method based on extreme learning machine algorithm Technical Field The invention relates to the technical field of teaching evaluation, in particular to an intelligent teaching practice quality evaluation method for teachers based on an extreme learning machine algorithm. Background Along with the digital promotion of intelligent education and education, college classroom teaching fuses intelligent teaching platform, digital resource and interactive tool increasingly, and teacher's links such as teaching design, classroom organization, interactive feedback, study effect promotion and data application in intelligent teaching practice directly influence teaching quality and talent cultivation effect. The university management department and the teaching research institution are in urgent need to develop quantitative evaluation on the intelligent teaching practice quality of teachers so as to support closed loops of teaching diagnosis, teacher development, resource allocation and quality improvement. However, intelligent teaching practice has the characteristics of strong procedural, multiple dimensions, multiple data sources and the like, and an evaluation object not only comprises a teacher class implementation process, but also relates to information such as student feedback, platform use recording, learning results and the like, so that objective and stable evaluation conclusion is difficult to form only by relying on a single source or a small number of indexes. The existing college teaching quality evaluation usually adopts the modes of questionnaire evaluation, supervision and class, expert scoring, statistical analysis and the like, is easy to be influenced by subjective factors, sample fluctuation, index weight setting difference, in the application of a data driving method, the existing research attempts to predict or classify by adopting regression, a support vector machine, a tree model, a neural network and the like, but in the practical application, the method still faces the problems that the number of evaluation indexes is large, the correlation is strong, the feature redundancy is easy to reduce the model interpretability and generalization performance, the reliability of the model training result is difficult to ensure when the data quality control link is insufficient, and the model stability and the applicability still need to be improved under the condition of limited sample scale. Therefore, it is necessary to provide a teacher intelligent teaching practice quality evaluation method based on an extreme learning machine algorithm, and the purpose of the method is to solve the above problems. Disclosure of Invention Aiming at the defects existing in the prior art, the invention aims to provide an intelligent teaching practice quality evaluation method for teachers based on an extreme learning machine algorithm, so as to solve the problems existing in the background art. The invention is realized in such a way that a teacher intelligent teaching practice quality evaluation method based on an extreme learning machine algorithm comprises the following steps: designing and issuing a questionnaire based on an intelligent teaching practice quality evaluation system of a college teacher, collecting influence factor index data, generating a comprehensive score and a corresponding quality grade label for each sample according to the evaluation system, and constructing a sample database; Performing reliability and effectiveness test on the obtained questionnaire data, and removing or correcting invalid questions, abnormal questionnaires and abnormal samples to obtain a trusted data set; the method comprises the steps of respectively carrying out mutual information sorting, sorting based on feature importance degree of a decision tree and sorting based on correlation redundancy degree, synthesizing various sorting results to obtain feature comprehensive sorting, eliminating high redundancy features under the constraint of a preset correlation threshold, selecting K features with the front comprehensive sorting as model input feature subsets, wherein K is a positive integer or is determined by cross verification; Searching super parameters of the extreme learning machine model by adopting a cross verification and grid optimizing method on the training set, determining optimal parameters, and establishing an extreme learning machine classification model, wherein the super parameters at least comprise hidden layer node numbers and activation function types; And inputting actual teaching process data into the model, outputting an intelligent teaching practice quality evaluation result of a college teacher, and using the intelligent teaching practice quality evaluation result for teaching improvement decision feedback. As a further scheme of the invention, the college teacher intelligent teaching practice quality evaluation system comprises five infl