English teaching quality evaluation method based on knowledge graph

Abstract

The invention relates to the field of English teaching, in particular to an English teaching quality evaluation method based on a knowledge graph. The technical problem to be solved by the invention is to provide the English teaching quality assessment method based on the knowledge graph, which can comprehensively and objectively assess the English teaching quality in combination with multi-dimensional information. The English teaching quality evaluation method based on the knowledge graph achieves the effect of comprehensively and objectively evaluating the English teaching quality in combination with multi-dimensional information of student learning data, teaching data and teaching resource information.

Description

Technical Field

[0001] The present invention relates to the field of English teaching, and in particular to an English teaching quality evaluation method based on knowledge graph. Background Art

[0002] In recent years, with the widespread application of information technology and digital technology in the field of education, our learning methods have undergone tremendous changes, breaking the time and space limitations of traditional teaching activities. Online learning has therefore gained widespread attention. Online learning mainly refers to learning through the Internet. Students can obtain texts, videos, pictures and other resources related to course knowledge from the Internet according to their actual needs. This open learning method lowers the threshold for knowledge acquisition, improves the ability to utilize online educational resources, and greatly promotes educational equity, which has extremely far-reaching significance.

[0003] With the development of Internet technology, various teaching management systems have gradually emerged, which greatly facilitated the teaching management work of schools. However, the teaching quality evaluation system in the existing teaching management system still uses the traditional evaluation method, which has problems such as large workload, small sampling range, untrue evaluation results, and high randomness, which reduces the actual value of the evaluation. Therefore, it is urgent to develop an English teaching quality evaluation method based on knowledge graph that can combine multi-dimensional information to achieve a comprehensive and objective evaluation of English teaching quality. Summary of the invention

[0004] In order to overcome the shortcomings of the teaching quality evaluation system in the existing teaching management system, such as large workload, small sampling range, untrue evaluation results, high randomness, etc., which reduce the actual value of the evaluation, the technical problem to be solved by the present invention is to provide an English teaching quality evaluation method based on knowledge graph that can combine multi-dimensional information to achieve a comprehensive and objective evaluation of English teaching quality.

[0005] The present invention is achieved by the following specific technical means: A method for evaluating the quality of English teaching based on knowledge graph, comprising the following steps: a. Multi-source data collection and processing: Collect structured data, including student terminal learning data, classroom attendance records, and teacher evaluation scores; Collect unstructured data, including classroom recordings, teaching videos, student assignment texts, and teacher lesson plans; Collect teaching environment data and obtain classroom acoustic parameters, light intensity, seat distribution density and teaching equipment usage status through IoT devices; Clean, deduplicate, and standardize multi-source data, and perform semantic annotation to generate a unified teaching data set; b. Construction of teaching knowledge graph: Construct a knowledge graph ontology framework including a teaching subject layer, a knowledge system layer, a teaching process layer and an effect evaluation layer; the teaching subject layer includes a teacher node with a teaching experience attribute and a student node with a learning style attribute, the knowledge system layer includes a grammar knowledge point network and a vocabulary semantic association matrix, the teaching process layer includes teaching activity nodes annotated with time and space characteristics, and the effect evaluation layer includes a mapping relationship between formative evaluation indicators and summative evaluation, wherein the formative evaluation indicators include the frequency of classroom interaction, and the summative evaluation includes test scores; Use the BERT-BiLSTM-CRF model to process unstructured text and combine it with the TransE algorithm to vectorize entity relationships; Complete missing triples through knowledge reasoning. When the analysis of the newly added teaching video shows that the average fixation time of a certain knowledge point exceeds the threshold, the weight of the node is automatically enhanced. c. Intelligent evaluation model construction: Extract quantitative evaluation indicators and qualitative evaluation indicators from the knowledge graph; the quantitative evaluation indicators include knowledge point coverage, teaching resource turnover rate and online learning activity, and the qualitative evaluation indicators include teaching goal achievement and classroom atmosphere index; Construct a hybrid evaluation model, use graph neural networks to analyze the relationship between teachers, students, and knowledge points, and dynamically adjust the weights of each indicator through the attention mechanism; Use time series models to track long-term changes in teaching effectiveness; An integrated expert rule base, which includes a teaching effect attenuation rule and a learning path optimization rule. The teaching effect attenuation rule triggers a course revision when the error rate of three consecutive evaluations exceeds a threshold, and the learning path optimization rule inserts intensive training when a weak association of "adjective usage → writing application" is detected; d. Visualization and optimization feedback: Generate a 3D visualization map where node size represents element importance, edge weights show association strength, and color coding indicates compliance status; Synchronously output teacher ability radar chart, knowledge point heat map and improvement path decision tree; Compare the evaluation results with the actual improvement effects through the reinforcement learning algorithm, and dynamically update the graph relationship weights; Push personalized strategies, including recommending teaching case libraries to teachers and prompting them to adjust the frequency of interaction, and generating adaptive paths for students that include remedial exercises and micro-class videos; Construct virtual teaching nodes, simulate the implementation effects of strategies such as flipped classrooms, and generate optimal solution prediction reports.

[0006] Furthermore, the student terminal learning data in step a includes the number of times a knowledge point is repeatedly learned and the standard deviation of test score fluctuations; the teaching environment data includes the regional speech clarity differences obtained through distributed acoustic sensors.

[0007] Furthermore, the knowledge system layer construction in step b includes: Explicit knowledge network: construct the predecessor and successor relationship diagram of grammatical rules based on the curriculum standards; Implicit knowledge network: Building a pragmatic strategy correlation matrix for cross-cultural communication scenarios; Dynamic update mechanism: When the accuracy of the test associated with a certain knowledge point continues to decline, the weight attenuation of the teaching resources connected to the node is automatically triggered.

[0008] Furthermore, the expert rule base in step c also includes: Teaching strategy verification rules: If the improvement in goal achievement after using the task-driven method does not meet expectations, the method suitability score will be reduced; Environmental adaptation rules: When the noise decibel value is continuously higher than 55dB, a suggestion to adjust the teaching venue is triggered.

[0009] Furthermore, the three-dimensional visualization map in step d supports: Multi-level drill-down analysis: Drill down from the entire school course network to the knowledge mastery path of a single student; Causal path tracing: linking classroom interaction records with post-test performance to generate a report on the effectiveness of teaching interventions; Virtual scenario comparison: side-by-side display of the differences in indicators between actual teaching and virtual strategies.

[0010] Furthermore, it also includes building a teaching effect prediction model to generate the optimal strategy through the following steps: Inject virtual teaching behavior nodes to build parallel teaching scenarios; Use the time series graph convolutional network to predict the indicator change trend in each scenario; Calculate the risk-benefit ratio of strategy implementation and output improvement plans with priority ranking.

[0011] Furthermore, an English teaching quality assessment system includes: Multi-source data acquisition module: integrated with IoT sensors, learning platform API, and audio and video analysis unit; Knowledge graph engine: a distributed graph database that supports dynamic ontology modeling and relational reasoning; Hybrid computing core: includes the GNN-TCN joint model and the Drools rule engine; Adaptive Feedback Module: A policy optimizer with reinforcement learning capabilities; Multi-terminal interactive platform: a visual system that supports access by teachers, students, and administrators based on their permissions.

[0012] Furthermore, a computer-readable storage medium stores a computer program, which implements the steps of any of the above methods when executed by a processor.

[0013] Compared with the prior art, the present invention has the following beneficial effects: The present invention combines multi-dimensional information of student learning data, teaching data and teaching resource information to achieve the effect of comprehensive and objective evaluation of English teaching quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a schematic diagram of the steps of the present invention. DETAILED DESCRIPTION

[0015] The present invention is further described below in conjunction with the accompanying drawings: Example

[0016] A method for evaluating the quality of English teaching based on knowledge graph, comprising the following steps: a. Multi-source data collection and processing: Collect structured data, including student terminal learning data, classroom attendance records, and teacher evaluation scores; Collect unstructured data, including classroom recordings, teaching videos, student assignment texts, and teacher lesson plans; Collect teaching environment data and obtain classroom acoustic parameters, light intensity, seat distribution density and teaching equipment usage status through IoT devices; Clean, deduplicate, and standardize multi-source data, and perform semantic annotation to generate a unified teaching data set; b. Construction of teaching knowledge graph: Construct a knowledge graph ontology framework including a teaching subject layer, a knowledge system layer, a teaching process layer and an effect evaluation layer; the teaching subject layer includes a teacher node with a teaching experience attribute and a student node with a learning style attribute, the knowledge system layer includes a grammar knowledge point network and a vocabulary semantic association matrix, the teaching process layer includes teaching activity nodes annotated with time and space characteristics, and the effect evaluation layer includes a mapping relationship between formative evaluation indicators and summative evaluation, wherein the formative evaluation indicators include the frequency of classroom interaction, and the summative evaluation includes test scores; Use the BERT-BiLSTM-CRF model to process unstructured text and combine it with the TransE algorithm to vectorize entity relationships; Complete missing triples through knowledge reasoning. When the analysis of the newly added teaching video shows that the average fixation time of a certain knowledge point exceeds the threshold, the weight of the node is automatically enhanced. c. Intelligent evaluation model construction: Extract quantitative evaluation indicators and qualitative evaluation indicators from the knowledge graph; the quantitative evaluation indicators include knowledge point coverage, teaching resource turnover rate and online learning activity, and the qualitative evaluation indicators include teaching goal achievement and classroom atmosphere index; Construct a hybrid evaluation model, use graph neural networks to analyze the relationship between teachers, students, and knowledge points, and dynamically adjust the weights of each indicator through the attention mechanism; Use time series models to track long-term changes in teaching effectiveness; An integrated expert rule base, which includes a teaching effect attenuation rule and a learning path optimization rule. The teaching effect attenuation rule triggers a course revision when the error rate of three consecutive evaluations exceeds a threshold, and the learning path optimization rule inserts intensive training when a weak association of "adjective usage → writing application" is detected; d. Visualization and optimization feedback: Generate a 3D visualization map where node size represents element importance, edge weights show association strength, and color coding indicates compliance status; Synchronously output teacher ability radar chart, knowledge point heat map and improvement path decision tree; Compare the evaluation results with the actual improvement effects through the reinforcement learning algorithm, and dynamically update the graph relationship weights; Push personalized strategies, including recommending teaching case libraries to teachers and prompting them to adjust the frequency of interaction, and generating adaptive paths for students that include remedial exercises and micro-class videos; Construct virtual teaching nodes, simulate the implementation effects of strategies such as flipped classrooms, and generate optimal solution prediction reports.

[0017] Working principle: This method is based on knowledge graph technology and realizes intelligent evaluation and optimization of English teaching quality through multi-source data fusion and dynamic modeling. Its core principles are as follows: First, through the integration of multi-source heterogeneous data, the system collects student terminal learning behavior data (such as the frequency of repeated learning of knowledge points, test score fluctuations), teacher teaching process data (classroom recordings, lesson plan texts) and teaching environment parameters (acoustics, lighting, seat density), and standardizes and cleans them and semantically annotates them to build a unified teaching data set. On this basis, a multi-layer teaching knowledge graph is constructed: the teaching subject layer depicts the relationship between teacher and student attributes, the knowledge system layer integrates explicit grammatical rules and implicit pragmatic strategies, the teaching process layer records teaching activities in the time and space dimensions, and the effect evaluation layer establishes a dynamic mapping relationship between formative evaluation (such as classroom interaction frequency) and summative evaluation (such as test scores). The BERT-BiLSTM-CRF model is used to extract the semantic features of unstructured text, combined with the TransE algorithm to vectorize entity relationships, and automatically complete missing associations through knowledge reasoning (such as dynamically adjusting the weight of knowledge points according to the length of time students watch videos).

[0018] Secondly, a hybrid intelligent evaluation model is constructed: based on the graph neural network (GNN), the potential patterns in the teacher-student-knowledge point ternary relationship are mined, and the temporal convolutional network (TCN) is used to track the long-term evolution trend of teaching effectiveness, and the weights of evaluation indicators are dynamically allocated through the attention mechanism (such as the synergistic influence of the classroom atmosphere index and the knowledge point coverage rate). The expert rule library is integrated to achieve closed-loop optimization: when it is detected that the association between "adjective usage → writing application" is weak, a special training module is automatically inserted; if the environmental noise continues to exceed the standard, the adjustment suggestion of the teaching venue is triggered. At the same time, the feedback system driven by reinforcement learning dynamically updates the weight of the knowledge graph relationship by comparing the simulation effect of virtual teaching strategies (such as flipped classrooms) with the actual improvement data, generates a three-dimensional visualization map to reveal the strength of the association between teaching elements, and pushes personalized case libraries to teachers and recommends adaptive learning paths to students, forming a continuous improvement mechanism of "evaluation-feedback-optimization".

[0019] Through the deep integration of data-driven and knowledge reasoning, the system realizes a full-chain evaluation from the mastery of micro knowledge points to the optimization of macro teaching strategies, significantly improving the accuracy of English teaching quality analysis and the timeliness of intervention.

[0020] Although the present disclosure has been described in detail with reference to the exemplary embodiments, the present disclosure is not limited thereto, and it will be apparent to those skilled in the art that various modifications and changes may be made thereto without departing from the scope of the present disclosure.

Claims

1. A method for evaluating English teaching quality based on knowledge graph, characterized in that: The method comprises the following steps: a. Multi-source data collection and processing: Collect structured data, including student terminal learning data, classroom attendance records, and teacher evaluation scores; Collect unstructured data, including classroom recordings, teaching videos, student assignment texts, and teacher lesson plans; Collect teaching environment data and obtain classroom acoustic parameters, light intensity, seat distribution density and teaching equipment usage status through IoT devices; Clean, deduplicate, and standardize multi-source data, and perform semantic annotation to generate a unified teaching data set; b. Construction of teaching knowledge graph: Construct a knowledge graph ontology framework including a teaching subject layer, a knowledge system layer, a teaching process layer and an effect evaluation layer; the teaching subject layer includes a teacher node with a teaching experience attribute and a student node with a learning style attribute, the knowledge system layer includes a grammar knowledge point network and a vocabulary semantic association matrix, the teaching process layer includes teaching activity nodes annotated with time and space characteristics, and the effect evaluation layer includes a mapping relationship between formative evaluation indicators and summative evaluation, wherein the formative evaluation indicators include the frequency of classroom interaction, and the summative evaluation includes test scores; Use the BERT-BiLSTM-CRF model to process unstructured text and combine it with the TransE algorithm to vectorize entity relationships; Complete missing triples through knowledge reasoning. When the analysis of the newly added teaching video shows that the average fixation time of a certain knowledge point exceeds the threshold, the weight of the node is automatically enhanced. c. Intelligent evaluation model construction: Extract quantitative evaluation indicators and qualitative evaluation indicators from the knowledge graph; the quantitative evaluation indicators include knowledge point coverage, teaching resource turnover rate and online learning activity, and the qualitative evaluation indicators include teaching goal achievement and classroom atmosphere index; Construct a hybrid evaluation model, use graph neural networks to analyze the relationship between teachers, students, and knowledge points, and dynamically adjust the weights of each indicator through the attention mechanism; Use time series models to track long-term changes in teaching effectiveness; An integrated expert rule base, which includes teaching effect attenuation rules and learning path optimization rules. The teaching effect attenuation rule triggers course revision when the error rate of three consecutive evaluations exceeds a threshold, and the learning path optimization rule inserts intensive training when a weak association of "adjective usage → writing application" is detected; d. Visualization and optimization feedback: Generate a 3D visualization map where node size represents element importance, edge weights show association strength, and color coding indicates compliance status; Synchronously output teacher ability radar chart, knowledge point heat map and improvement path decision tree; Compare the evaluation results with the actual improvement effects through the reinforcement learning algorithm, and dynamically update the graph relationship weights; Push personalized strategies, including recommending teaching case libraries to teachers and prompting them to adjust the frequency of interaction, and generating adaptive paths for students that include remedial exercises and micro-class videos; Construct virtual teaching nodes, simulate the implementation effects of strategies such as flipped classrooms, and generate optimal solution prediction reports.

2. According to claim 1, a method for evaluating English teaching quality based on knowledge graph is characterized in that: The student terminal learning data in step a includes the number of times a knowledge point is repeatedly learned and the standard deviation of test score fluctuations; the teaching environment data includes the regional speech clarity differences obtained through distributed acoustic sensors.

3. According to claim 1, a method for evaluating English teaching quality based on knowledge graph, characterized in that: The knowledge system layer construction in step b includes: Explicit knowledge network, constructing the predecessor and successor relationship diagram of grammatical rules based on curriculum standards; Implicit knowledge network, building a pragmatic strategy correlation matrix for cross-cultural communication scenarios; Dynamic update mechanism: when the accuracy of the test associated with a certain knowledge point continues to decline, it automatically triggers the weight attenuation of the teaching resources connected to the node.

4. According to claim 1, a method for evaluating English teaching quality based on knowledge graph is characterized in that: The expert rule base in step c also includes: Teaching strategy verification rules: If the improvement in goal achievement after using the task-driven method does not meet expectations, the method suitability score will be reduced; Environmental adaptation rules: When the noise decibel value is continuously higher than 55dB, a suggestion to adjust the teaching venue is triggered.

5. The method for evaluating English teaching quality based on knowledge graph according to claim 1, characterized in that: The 3D visualization map in step d supports: Multi-level drill-down analysis: Drill down from the entire school course network to the knowledge mastery path of a single student; Causal path tracing: linking classroom interaction records with post-test performance to generate a report on the effectiveness of teaching interventions; Virtual scenario comparison: side-by-side display of the differences in indicators between actual teaching and virtual strategies.

6. The method for evaluating English teaching quality based on knowledge graph according to claim 1, characterized in that: It also includes building a teaching effect prediction model and generating strategies through the following steps: Inject virtual teaching behavior nodes to build parallel teaching scenarios; Use the time series graph convolutional network to predict the indicator change trend in each scenario; Calculate the risk-benefit ratio of strategy implementation and output improvement plans with priority ranking.

7. An English teaching quality evaluation system, characterized in that: include: Multi-source data acquisition module: integrated with IoT sensors, learning platform API, and audio and video analysis unit; Knowledge graph engine: a distributed graph database that supports dynamic ontology modeling and relational reasoning; Hybrid computing core: includes the GNN-TCN joint model and the Drools rule engine; Adaptive Feedback Module: A policy optimizer with reinforcement learning capabilities; Multi-terminal interactive platform: a visual system that supports access by teachers, students, and administrators based on their permissions.

8. A computer-readable storage medium storing a computer program, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images