Teaching data analysis method and device for Internet of Things courses of secondary vocational schools and medium

Abstract

The invention discloses a teaching data analysis method and device for Internet of Things courses of secondary vocational schools and a medium, and can be applied to the technical field of artificial intelligence. According to the method, after a teaching scene adaptive four-dimensional knowledge graph and a dual-channel generation architecture comprising a semantic generation channel and a graph logic channel are constructed, semantic adjustment is performed on dialogue question information input by students through the semantic generation channel to obtain a dialogue question semantic vector; inputting a four-dimensional knowledge graph through a graph logic channel, executing a knowledge graph neural algorithm to generate dialogue answer information corresponding to a dialogue question semantic vector, and optimizing teaching question and answer prompt information through a cognitive trajectory tracking algorithm according to dialogue question information and the dialogue answer information; therefore, when the knowledge graph is applied to the middle vocational internet-of-things course, the teaching effect of the middle vocational internet-of-things course can be effectively improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a teaching data analysis method, device and medium for Internet of Things (IoT) courses in secondary vocational schools. Background Technology

[0002] In related technologies, the teaching needs of IoT courses in secondary vocational schools are mainly reflected in the following aspects: Highly practical: The course delves into hands-on activities such as sensor deployment, device networking, and cloud platform development. In actual teaching, students frequently need to operate and debug hardware devices, requiring teaching tools to possess robust hardware fault diagnosis capabilities, providing timely and accurate diagnoses and solutions when students encounter equipment problems.

[0003] Students' cognitive characteristics: Vocational school students have certain difficulties in accepting abstract theories. Therefore, the teaching content needs to transform the complex Internet of Things principles into intuitive and easy-to-understand equipment operation instructions to match the cognitive level of vocational school students and improve learning outcomes.

[0004] Rapid knowledge iteration: The Internet of Things (IoT) field is developing rapidly, with frequent updates to protocols and hardware. If teaching resources cannot be dynamically synchronized with industry standards in a timely manner, the knowledge students acquire will quickly become disconnected from practical applications.

[0005] Based on the aforementioned needs, existing technologies apply large language models or knowledge graphs to vocational IoT courses. However, the application of existing large language models in vocational IoT courses suffers from issues such as confusion of knowledge concepts, insufficient explanation of theoretical basis, and low diagnostic rate for equipment malfunctions. Existing knowledge graphs, when applied to vocational IoT courses, suffer from slow knowledge graph updates and low accuracy of retrieval results when combined with large language models, leading to poor teaching effectiveness in vocational IoT courses.

[0006] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0007] The main objective of this application is to propose a teaching data analysis method, device, and medium for Internet of Things (IoT) courses in secondary vocational schools, which can effectively improve the teaching effectiveness of IoT courses in secondary vocational schools.

[0008] To achieve the above objectives, one aspect of this application proposes a teaching data analysis method for Internet of Things (IoT) courses in secondary vocational schools, the method comprising the following steps: Constructing a four-dimensional knowledge graph that adapts to teaching scenarios; A dual-channel generation architecture is constructed, which includes a semantic generation channel and a graph logic channel; The semantic generation channel is used to semantically adjust the dialogue question information input by the student to obtain a dialogue question semantic vector; After the four-dimensional knowledge graph is input through the graph logic channel, the knowledge graph neural algorithm is executed to generate the dialogue answer information corresponding to the semantic vector of the dialogue question. Based on the dialogue question information and the dialogue answer information, the teaching question and answer prompt information is optimized using a cognitive trajectory tracking algorithm.

[0009] In some embodiments, the four-dimensional knowledge graph includes: The concept layer is used to store the principles and conceptual relationships of the Internet of Things; The device layer is used to store the parameters and interface specifications of sensors or actuators. The operation layer is used to store the training operation instruction set and execution logic; The fault layer is used to store error cases and solutions extracted from historical training logs.

[0010] In some embodiments, the semantic generation channel includes a deep learning model based on an attention mechanism. The step of semantically adjusting the dialogue question information input by the student through the semantic generation channel to obtain a dialogue question semantic vector includes: The model parameters of the deep learning model are adjusted according to the learning stage information of secondary vocational school students; The deep learning model, after adjusting its model parameters, performs semantic adjustment on the dialogue question information input by the student to obtain the dialogue question semantic vector.

[0011] In some embodiments, after inputting the four-dimensional knowledge graph through the graph logic channel, the step of executing a knowledge graph neural algorithm to generate dialogue answer information corresponding to the semantic vector of the dialogue question includes: The semantic vector of the dialogue question is parsed using a teaching scenario classifier to obtain the user's question intent information; Match the target subgraph corresponding to the user's question intent information in the four-dimensional knowledge graph; The target subgraph is aligned with the dialogue question semantic vector to obtain an aligned feature vector. The output information generated after logical verification based on the aligned feature vector is used as the dialogue response information.

[0012] In some embodiments, the teaching scenario classifier includes: The fault diagnosis module is used to identify first query information, which includes sensor anomaly query information, communication protocol error query information, and power problem query information. The operation process guidance module is used to identify the second query information, which includes device wiring query information, code burning query information, and platform configuration query information. The principle concept analysis module is used to identify third query information, which includes protocol principle query information, data transmission mechanism query information, and system architecture query information.

[0013] In some embodiments, matching the target subgraph corresponding to the user's question intent information in the four-dimensional knowledge graph includes: Calculate the similarity between each candidate subgraph in the four-dimensional knowledge graph and the user's question intent information; The target subgraph is determined from the candidate subgraphs based on the similarity.

[0014] In some embodiments, optimizing the teaching question-and-answer prompts using a cognitive trajectory tracking algorithm based on the dialogue question information and the dialogue answer information includes: The dialogue question information and the dialogue answer information are mapped to a path sequence of the four-dimensional knowledge graph; The path sequence is input into a graph neural network to predict the probability distribution of knowledge blind spot nodes; The teaching question and answer prompts are optimized based on the probability distribution of the knowledge blind spots.

[0015] To achieve the above objectives, another aspect of this application proposes a teaching data analysis device for an Internet of Things (IoT) course in a secondary vocational school, the device comprising: The first building block is used to construct a four-dimensional knowledge graph that adapts to teaching scenarios; The second building unit is used to build a dual-channel generation architecture, which includes a semantic generation channel and a graph logic channel. The adjustment unit is used to semantically adjust the dialogue question information input by the student through the semantic generation channel to obtain the dialogue question semantic vector; An execution unit is used to execute a knowledge graph neural algorithm to generate dialogue response information corresponding to the semantic vector of the dialogue question after inputting the four-dimensional knowledge graph through the graph logic channel. The optimization unit is used to optimize the teaching question and answer prompts based on the dialogue question information and the dialogue answer information using a cognitive trajectory tracking algorithm.

[0016] To achieve the above objectives, another aspect of this application provides a computer device including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described method.

[0017] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0018] To achieve the above objectives, another aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method. The embodiments of this application include at least the following beneficial effects: This application provides a teaching data analysis method, device, and medium for IoT courses in secondary vocational schools. This solution constructs a teaching scenario-adaptive four-dimensional knowledge graph and a dual-channel generation architecture including a semantic generation channel and a graph logic channel. The semantic generation channel semantically adjusts the dialogue question information input by students to obtain a dialogue question semantic vector. Then, after inputting the four-dimensional knowledge graph through the graph logic channel, the knowledge graph neural algorithm is executed to generate dialogue answer information corresponding to the dialogue question semantic vector. Based on the dialogue question information and dialogue answer information, the cognitive trajectory tracking algorithm optimizes the teaching question and answer prompt information. Thus, when the knowledge graph is applied to IoT courses in secondary vocational schools, it can effectively improve the teaching effect of IoT courses. Attached Figure Description

[0019] Figure 1 This is a flowchart of the teaching data analysis method for the Internet of Things (IoT) course in secondary vocational schools provided in the embodiments of this application; Figure 2 This is a schematic diagram of the teaching data analysis device for the Internet of Things (IoT) course in secondary vocational schools provided in this application embodiment. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application.

[0021] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various concepts, but unless otherwise stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words “if,” “when,” or “in response to a determination” as used herein may be interpreted as “when…” or “when…” or “in response to a determination.”

[0022] As used in this application, the terms "at least one", "multiple", "each", "any", etc., "at least one" includes one, two or more, "multiple" includes two or more, "each" refers to each of the corresponding multiples, and "any" refers to any one of the multiples.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0024] In related technologies, the teaching needs of IoT courses in secondary vocational schools are mainly reflected in the following aspects: Highly practical: The course delves into hands-on activities such as sensor deployment, device networking, and cloud platform development. In actual teaching, students frequently need to operate and debug hardware devices, requiring teaching tools to possess robust hardware fault diagnosis capabilities, providing timely and accurate diagnoses and solutions when students encounter equipment problems.

[0025] Students' cognitive characteristics: Vocational school students have certain difficulties in accepting abstract theories. Therefore, the teaching content needs to transform the complex Internet of Things principles into intuitive and easy-to-understand equipment operation instructions to match the cognitive level of vocational school students and improve learning outcomes.

[0026] Rapid knowledge iteration: The Internet of Things (IoT) field is developing rapidly, with frequent updates to protocols and hardware. If teaching resources cannot be dynamically synchronized with industry standards in a timely manner, the knowledge students acquire will quickly become disconnected from practical applications.

[0027] Based on the aforementioned teaching needs, existing technologies apply large language models combined with static knowledge graphs to vocational secondary school teaching. However, large language models have the following problems in their application: Fragmented knowledge: General-purpose LLMs (such as the GPT series) lack a systematic knowledge structure in the field of IoT. When answering fundamental questions, conceptual confusion often occurs, such as confusing the differences between Zigbee and LoRa protocols in terms of application scenarios and transmission characteristics. There are also shortcomings in explaining the theoretical basis behind hardware operations, failing to clearly explain the connection between practical operation and theory, such as "why RS485 needs a terminating resistor".

[0028] Ineffective Practical Training Guidance: In actual teaching applications, traditional large language models have a low accuracy rate in diagnosing equipment faults. This is mainly because large language models lack a database of real-world training error cases and cannot effectively diagnose faults based on actual teaching scenarios. At the same time, the output content does not fully consider the cognitive level of vocational school students and uses a large number of professional terms, which makes it difficult for students to understand.

[0029] Static knowledge graphs have the following problems in application: Rigid Educational Maps: Existing educational maps (such as the CNKI Academic Map) suffer from long update cycles, averaging only once a year, which is far from keeping pace with the weekly iteration speed of IoT technology. Furthermore, these maps lack key teaching elements for practical training, such as equipment wiring diagrams and fault code tables, failing to meet the practical needs of IoT teaching in vocational schools.

[0030] Interactive fragmentation: Simply splicing the graph with the LLM has obvious drawbacks. In practical use, keyword searches often trigger incorrect subgraphs. For example, entering "cloud data anomaly" may mistakenly trigger the cloud computing module instead of specific fault diagnosis content related to IoT teaching, leading to biased search results and thus affecting teaching effectiveness.

[0031] In view of this, the embodiments of this application provide a teaching data analysis method, device and medium for Internet of Things (IoT) courses in secondary vocational schools, which can effectively improve the teaching effect of IoT courses in secondary vocational schools.

[0032] The teaching data analysis method for IoT courses in vocational schools provided in this application relates to the field of artificial intelligence technology. This teaching data analysis method for IoT courses in vocational schools can be applied to a terminal, a server, or software running on a terminal or server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, or in-vehicle terminal, but is not limited to these. The server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network. The software can be an application implementing the teaching data analysis method for IoT courses in vocational schools, but is not limited to the above forms.

[0033] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0034] It should be noted that in all specific embodiments of this application, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when embodiments of this application require access to sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirects to confirmation pages. Only after obtaining the user's separate permission or consent is the necessary user-related data required for the proper functioning of these embodiments acquired.

[0035] The embodiments of this application will be described in detail below with reference to the accompanying drawings: Figure 1 This is an optional flowchart of the teaching data analysis method for the Internet of Things (IoT) course in secondary vocational schools provided in the embodiments of this application. Figure 1 The method may include, but is not limited to, steps S110 to S150: Step S110: Construct a four-dimensional knowledge graph that adapts to teaching scenarios; Step S120: Construct a dual-channel generation architecture, which includes a semantic generation channel and a graph logic channel; Step S130: The semantic information of the dialogue question input by the student is semantically adjusted through the semantic generation channel to obtain the semantic vector of the dialogue question; Step S140: After inputting the four-dimensional knowledge graph through the graph logic channel, execute the knowledge graph neural algorithm to generate dialogue answer information corresponding to the semantic vector of the dialogue question; Step S150: Optimize the teaching question and answer prompts based on the dialogue question information and dialogue answer information using the cognitive trajectory tracking algorithm.

[0036] It is understood that the four-dimensional knowledge graph in this embodiment includes the following layers: The conceptual layer is used to store the principles and conceptual relationships of the Internet of Things (IoT). The device layer is used to store the parameters and interface specifications of sensors or actuators. The operation layer is used to store the training operation instruction set and execution logic; The fault layer stores error cases and solutions extracted from historical training logs.

[0037] Specifically, at the conceptual layer, this embodiment uses the Neo4j graph database for storage, modeling IoT concepts according to hierarchical relationships. For example, the "sensor" node connects to child nodes such as "temperature sensor" and "pressure sensor," and labels node attributes (such as definition and application scenario) and relationship attributes (such as inheritance and dependency). For instance, the "MQTT protocol" node establishes a "belongs to" relationship with the "IoT communication" node, and at the same time establishes an "adopt" relationship with the "publish / subscribe mode" node.

[0038] At the device layer, this embodiment integrates component database APIs such as Digi-Key and Mouser to automatically capture the latest sensor parameters (such as the temperature accuracy of DHT22 ±0.5°C and humidity range of 0-100%RH), and defines standardized interface description files (such as the GPIO pin definition JSON file of ESP32) through the OpenAPI specification.

[0039] In the operational layer, this embodiment uses a finite state machine model to represent the training process. Each state corresponds to an operation step (such as "Arduino IDE installation" → "ESP32 development board configuration" → "code upload"). The state transition condition is triggered by sensor data or user operation. For example, in the "temperature and humidity monitoring" training, when a student connects the DHT22 sensor, the system automatically activates the "read sensor data" state.

[0040] In the fault layer, this embodiment uses ELKStack (Elasticsearch + Logstash + Kibana) to process training logs, extracts error features (such as "Connection refused" and "Sensor timeout") through NLP technology, and builds a fault tree model (such as "communication failure" → "network configuration error" → "IP address conflict").

[0041] It is understandable that this embodiment, when constructing the four-dimensional knowledge graph, will also construct a dual-channel generation architecture including a semantic generation channel and a graph logic channel. The semantic generation channel can be implemented based on the HuggingFaceTransformers library, using the RoBERTa-large pre-trained model, and fine-tuned for domain adaptation using 100,000 teaching texts (including textbooks, lab manuals, and industry reports) for the Internet of Things (IoT) field. When the input question is "Why does RS485 need a terminating resistor?", the RoBERTa-large pre-trained model outputs a semantic vector with a dimension of 768. Specifically, in the semantic generation channel, this embodiment can process natural language input using a deep learning model (Transformer architecture) based on an attention mechanism. For example, suppose the input sequence is... Attention scores are calculated using a multi-head attention mechanism. The formula is as follows: ; In the formula, , These are the query vector and the key vector, respectively. Given the dimension of the key vector, attention weights are obtained through the softmax function, and then combined with the value vector. The weighted sum is used to obtain the output.

[0042] The graph logic channel is a custom-developed graph query engine that supports attribute-based conditional queries (such as MATCH(n:Sensor)-[r:CONNECTS_TO]-(m:Microcontroller)WHEREEn.type='Temperature'RETURNm) and path-based relational queries (such as querying the complete communication link "DHT22→ESP32→MQTT server"). Specifically, the graph logic channel is used to input the four-dimensional knowledge graph into the large language model in real time, and to impose logical constraints on the processing procedures in the large language model.

[0043] It is understandable that the semantic generation channel incorporates a deep learning model based on an attention mechanism. In this embodiment, after constructing the semantic generation channel, the dialogue question information input by the student is semantically adjusted through the semantic generation channel to obtain a dialogue question semantic vector. Specifically, the semantic adjustment process can involve adjusting the model parameters of the deep learning model according to the learning stage information of vocational school students, and then using the adjusted deep learning model to semantically adjust the dialogue question information input by the student to obtain the dialogue question semantic vector. In particular, this embodiment introduces a course difficulty adaptive mechanism, setting the output parameter adjustment coefficients of the Transformer architecture according to the information of different learning stages of vocational school students. Output the adjusted semantic vector The formula is: ,in, The value range is (0.8, 1) during the basic learning stage and (1, 1.2) during the advanced learning stage.

[0044] Specifically, the learning phases are divided based on a course progress model, dividing the IoT course into four phases: basic understanding (1-2 weeks), skills mastery (3-8 weeks), comprehensive application (9-16 weeks), and innovative practice (17-20 weeks). Then, during the development phase, the evaluation and testing system regularly assesses students' learning levels through knowledge quizzes and skills assessments, automatically adjusting the learning phase accordingly. Student learning profiles are also established, recording learning progress, academic performance, and ability assessment results to support personalized learning path planning.

[0045] The parameter tuning strategy in the semantic generation channel is to achieve... The value is dynamically adjusted based on a combination of factors, including the learning stage, problem complexity, and student's historical performance. For example, in the basic stage, when a student asks "What is a sensor?", =0.85; In the advanced stage, when students ask "How to implement MQTT encrypted transmission of sensor data?", =1.15 Development domain terminology simplification module, when When the threshold is less than 1, technical jargon will be automatically replaced with plain and easy-to-understand expressions (e.g., "edge computing" → "local data processing"), and the description of technical details will be reduced. This implements a knowledge depth control mechanism. When the value is greater than 1, add theoretical derivation process and expand knowledge in related fields (such as introducing OSI model and TLS encryption mechanism when explaining the principle of MQTT).

[0046] Furthermore, an effectiveness evaluation mechanism is incorporated into the semantic generation process. Specifically, this involves developing an explanation quality assessment model that uses metrics such as BLEU and ROUGE to evaluate the similarity between the generated answer and the standard answer, and continuously optimizing it based on student feedback data. Value adjustment strategy. Establish a learning effect tracking system to compare student test scores, training completion rates, and other indicators before and after using the adaptive mechanism to evaluate the effectiveness of the mechanism. Implement student feedback collection channels, collecting students' subjective evaluations of the difficulty of the answers through questionnaires, comments, etc., as a basis for model optimization.

[0047] Understandably, after constructing the graph logic channel, this embodiment inputs the four-dimensional knowledge graph through the graph logic channel and then executes the knowledge graph neural algorithm to generate dialogue response information corresponding to the dialogue question semantic vector. Specifically, this embodiment uses a teaching scenario classifier to parse the dialogue question semantic vector to obtain user question intent information, matches the target subgraph corresponding to the user question intent information in the four-dimensional knowledge graph, aligns the target subgraph with the dialogue question semantic vector to obtain an alignment feature vector, and then generates logically verified output information as dialogue response information based on the alignment feature vector.

[0048] In this embodiment, the teaching scenario classifier includes, but is not limited to, the following modules: The fault diagnosis module is used to identify the first query information, which includes sensor anomaly query information, communication protocol error query information, and power problem query information. The operation process guidance module is used to identify the second query information, which includes device wiring query information, code burning query information, and platform configuration query information. The principle and concept analysis module is used to identify third-party query information, which includes protocol principle query information, data transmission mechanism query information, and system architecture query information.

[0049] Specifically, the fault diagnosis module's process for identifying sensor anomalies includes, but is not limited to, the following steps: Construct a sensor fault feature dictionary, such as keywords like "constant reading", "excessive fluctuation", and "out of range", and combine it with regular expression matching (e.g., / reading\s*(unchanged|constant|fixed) / ). Develop sensor signal analysis algorithms to detect signal spectrum anomalies through FFT transformation. For example, when high-frequency noise appears in the temperature sensor signal, it is determined to be an "electromagnetic interference fault". Implement a fault mode library and create a feature-cause-solution mapping table for common sensor faults (such as "CRC check failure" caused by a loose DHT22 data cable).

[0050] The fault diagnosis module's process for identifying communication protocol errors includes, but is not limited to, the following steps: Error templates are defined based on protocol specifications, such as the analysis of the CONNACK return code in the MQTT protocol (e.g., a return of 0x04 indicates "incorrect username or password"). Develop a network packet sniffing tool to capture communication data in real time, and use the Wireshark API to parse protocols and identify problems such as TCP connection timeouts and abnormal packet loss rates. Establish a protocol state machine model to monitor whether the communication process conforms to the protocol specifications. For example, if an HTTP request does not receive a 200 response, the "server response error" diagnostic process will be automatically triggered.

[0051] The fault diagnosis module's process for identifying power problems includes, but is not limited to, the following steps: To achieve voltage and current monitoring, the device's power supply parameters are acquired through ADC chips such as ADS1115, and a threshold range is set (e.g., a 5V power supply is allowed to fluctuate by ±0.5V). Develop a power timing analysis tool to check whether the power-on sequence meets the requirements (e.g., FPGA requires core voltage to be supplied before I / O voltage). Establish a power supply fault knowledge base, including detection methods and repair solutions for common power supply problems (such as "power adapter specifications are incompatible" and "damaged filter capacitors").

[0052] The operation process guidance module includes, but is not limited to, the following steps in the process of providing equipment wiring guidance: Develop a 3D wiring visualization tool based on Three.js to render device models and support drag-and-drop wiring operations. For example, when students connect an Arduino to an ultrasonic sensor in a virtual environment, the system checks in real time whether the wiring is correct (e.g., VCC connected to 5V, GND grounded, Trig connected to D2, Echo connected to D3). Establish a wiring standard knowledge base, including wiring color codes for different devices (e.g., red for the A line and blue for the B line in RS485), wiring sequence requirements (e.g., T568B standard for RJ45 crystal heads), and reverse connection prevention measures. Develop a wiring error detection algorithm that automatically identifies problems such as short circuits, open circuits, and reverse connections through methods such as current loop detection and voltage divider testing.

[0053] The operation process guidance module includes, but is not limited to, the following steps during the code flashing process: Build an IDE integration plugin that supports mainstream development environments such as ArduinoIDE and PlatformIO, automatically detect compilation errors and provide repair suggestions (such as "undefined variable 'SensorPin'" → "check variable name spelling or add definition"). Develop an automatic configuration tool for programming parameters, which can automatically set parameters such as baud rate, COM port, and partition table according to the target device type (such as ESP8266, STM32); Establish a code template library containing code frameworks for common IoT application scenarios (such as temperature and humidity monitoring, remote control, and data uploading), supporting intelligent completion and error highlighting.

[0054] The operation process guidance module includes, but is not limited to, the following steps in the process of providing platform configuration guidance: The development of a cloud platform configuration wizard supports mainstream platforms such as Alibaba Cloud IoT and Tencent Cloud IoT Explorer. It guides students through operations such as product creation, device registration, and rule engine configuration via a graphical interface. Establish a configuration error detection mechanism to verify configuration parameters (such as MQTT connection parameters and Topic permission settings) through API calls and provide real-time feedback on the configuration status. Develop a configuration file generator that supports formats such as JSON and YAML, automatically fills in necessary parameters, and provides sample code (such as MQTT connection configuration file generation).

[0055] The process of performing protocol principle analysis in the principle concept analysis module includes, but is not limited to, the following steps: Develop a protocol visualization simulator to display the protocol execution process in real time based on network packet capture data. For example, use WebSocket to dynamically visualize the MQTT message publish / subscribe process, showing the interaction of control messages such as CONN, PUBLISH, and SUB.

[0056] Establish a protocol comparison knowledge base, and use feature matrix form to compare the key characteristics (such as transport layer protocol, message format, QoS level, power consumption) of different protocols (such as HTTP, CoAP, MQTT), supporting multi-dimensional query and comparative analysis.

[0057] Develop a protocol principle animation generation tool that generates protocol working principle animations based on SVG technology (such as the formation process of Zigbee mesh network and the spread spectrum modulation principle of LoRa), and supports interactive parameter adjustment and result preview.

[0058] The process of analyzing the data transmission mechanism in the principle concept analysis module includes, but is not limited to, the following steps: Build a data transmission path tracing tool to monitor data flow through network probe technology and generate a complete data transmission path map from sensor → edge device → gateway → cloud platform.

[0059] Develop transmission efficiency analysis tools to calculate metrics such as bandwidth utilization, latency, and packet loss rate for different transmission methods (e.g., TCP / IP, UDP, serial communication) and provide optimization suggestions.

[0060] Establish a data format conversion knowledge base to support mutual conversion between common data formats such as JSON, XML, and binary, and provide conversion code examples and error handling methods.

[0061] The system architecture analysis process in the principle and concept analysis module includes, but is not limited to, the following steps: Develop a visual architecture editor based on Graphviz to create drag-and-drop IoT system architecture diagrams, supporting automatic layout and layered display (such as perception layer → network layer → platform layer → application layer).

[0062] Establish a library of typical architecture patterns, including application scenarios, advantages and disadvantages analysis, and design guidance for star, mesh, and bus topologies.

[0063] Develop architecture evaluation tools to quantitatively evaluate candidate architecture solutions and recommend the optimal solution based on performance metrics (such as scalability, reliability, and security) and constraints (such as cost, power consumption, and deployment environment).

[0064] It is understood that, after obtaining the user's question intent information, this embodiment matches the target subgraph corresponding to the user's question intent information in the four-dimensional knowledge graph. Specifically, this embodiment can determine the target subgraph from the candidate subgraphs based on the similarity between each candidate subgraph in the four-dimensional knowledge graph and the user's question intent information. This embodiment aligns the target subgraph with the dialogue question semantic vector to obtain an aligned feature vector. For example, assuming the knowledge graph is G=(V,E) and the user intent vector is I, the similarity calculation function... Determine the target subgraph The formula is as follows: ; In the formula, For subgraph The set of nodes, for The weight, For node v and intention vector The feature matching function.

[0065] Using a graph attention network, align the graph features corresponding to the target subgraph with the semantic features corresponding to the dialogue question semantic vector. Let the graph feature vector be... The semantic feature vector is Aligned feature vectors The calculation formula is as follows: ; In the formula, These are the weighting coefficients, obtained through training and optimization.

[0066] It is understood that in this embodiment, after obtaining the alignment feature vector, the logically verified output information is generated based on the alignment feature vector as the dialogue response information. Specifically, the logical verification process in this embodiment can adopt a combination of a rule engine and a probabilistic model. Assuming the rule matching degree is... The confidence level of the probability model is Final verification score The calculation formula is as follows: ; In the formula, ,and It was obtained through training and optimization using historical validation data.

[0067] The implementation of the rule engine can utilize the Drools rule engine to define a logical rule base for the Internet of Things (IoT) domain. This process includes, but is not limited to: Device connection rules (e.g., "RS485 bus must be connected to a terminating resistor"); Protocol constraints (e.g., "MQTTQoS2 must include four messages: CONNACK, PUBREC, PUBREL, and PUBCOMP"). Data processing rules (e.g., "Sensor data must be filtered before it can be uploaded"); Develop a rule matching algorithm to convert the generated answers into fact objects that the rule engine can process, and calculate the rule matching degree R (value range 0-1).

[0068] The process of constructing a probabilistic model includes, but is not limited to: A probabilistic verification model is built based on Bayesian networks, with nodes including question type, knowledge point, reasoning steps, and conclusion reliability. The model is trained using historical validation data to learn the probability distribution under different conditions (such as the conditional probability of "abnormal sensor reading" → "loose wiring" in the "hardware fault diagnosis" scenario). Implement a confidence score calculation algorithm to calculate the confidence score C (range 0-1) of the answer through probabilistic reasoning.

[0069] In this embodiment, a weight optimization mechanism is also provided. The process of the weight optimization mechanism includes, but is not limited to, the following: Develop a historical validation dataset management system to record the questions, answers, rule matching results, probability model prediction results, and final validation results for each validation session. accomplish The parameter optimization algorithm uses gradient descent to minimize the mean square error between the predicted verification score and the actual verification result. Establish a verification result feedback mechanism so that when the system verification result is inconsistent with the expert review result, the rule base and probability model update process is automatically triggered.

[0070] It is understood that, in this embodiment, when performing intelligent question-and-answer operations for IoT courses, the teaching question-and-answer prompts can be optimized using a cognitive trajectory tracking algorithm based on the dialogue question information and dialogue answer information. Specifically, this embodiment can map the dialogue question information and dialogue answer information into a path sequence P of a four-dimensional knowledge graph, and then input the path sequence into a graph neural network M to predict the probability distribution Y of knowledge blind spot nodes. The prediction process for the probability distribution Y of knowledge blind spot nodes is as follows: Y = M(P); This embodiment, after predicting the probability distribution of knowledge blind spots, optimizes the teaching question-and-answer prompts based on this probability distribution. Specifically, this embodiment can optimize the prompting process using a reinforcement learning algorithm based on the predicted knowledge blind spots, and sets a reward function. State transition probability Update the hint strategy through iterative formulas ,in As a discount factor, For state The value function of .

[0071] The process of defining the state space is as follows: The nodes in the knowledge graph are used as a set of states S, and each state represents a knowledge point (such as "MQTT protocol" or "temperature and humidity sensor"). Define a state feature vector, which includes dimensions such as node difficulty coefficient, learning frequency, mastery level, and number of related knowledge points; A state transition detection algorithm was developed to determine the student's current knowledge state in real time by analyzing the student's dialogue content and operation behavior.

[0072] The design process of the action space includes defining a set of cue actions A, which includes several elements a. Action set A includes, but is not limited to: Knowledge review prompts (such as "Review the sensor signal conditioning circuits learned in the last lesson"); Analogical explanation suggestion (e.g., "MQTT's QoS mechanism can be compared to regular mail, registered mail, and express delivery"). Step-by-step guidance prompts (such as "First, we check the power supply voltage of the sensor, and then measure the signal output"). Case study prompts (such as "Let's look at a solution to a similar problem..."); Implement action feature encoding, converting each prompt action into a multi-dimensional vector containing features such as prompt type, expected effect, and resource consumption.

[0073] The design process of the reward function includes, but is not limited to: Design a multi-dimensional reward function : Learning outcome rewards (e.g., +10 points for successfully solving a problem, -5 points for repeatedly asking the same question); Learning efficiency rewards (e.g., +5 points for students who master knowledge points in a shorter time); Learning experience rewards (e.g., +3 points for high student satisfaction with prompts); Develop a reward calculation engine to comprehensively analyze student behavior data (such as answer accuracy, operation time, and feedback evaluation) to calculate reward values.

[0074] The process of strategy optimization includes, but is not limited to: A reinforcement learning model is implemented using a deep Q-network (DQN), and the neural network is constructed using the PyTorch framework. Implement an experience replay mechanism and store state transition samples. Random sampling is used for training to break sample correlation; Develop a strategy evaluation system to compare student learning outcomes under different prompting strategies through A / B testing, and continuously optimize model parameters.

[0075] In this embodiment, the device layer of the four-dimensional knowledge graph also stores 3D model data of IoT devices and virtual training operation step data to support virtual simulation teaching scenarios. The virtual training operation step data includes operation duration thresholds and operation sequence constraints. The management process of the 3D model data includes, but is not limited to, the following: Establish a 3D model library for devices, supporting mainstream 3D model formats such as OBJ, STL, and FBX, covering common IoT devices (such as Arduino development boards, sensor modules, and gateway devices). Develop a model parameter configuration tool that allows adjustment of model size, color, transparency and other attributes according to different teaching needs, and supports model combination (such as assembling multiple sensors into a monitoring system). Implement model interaction functions, supporting drag, rotate, scale and other operations, and annotate key interfaces and operation points (such as pin names and wiring order) on the model.

[0076] The design process of virtual training workflow includes, but is not limited to: The virtual training process is represented by a finite state machine model, defining a set of states (such as "preparation stage", "wiring stage", "programming stage" and "testing stage") and state transition conditions. Develop an operation step editor that supports visual creation of training steps. Each training step includes, but is not limited to, the following: Operation description text; Expected operation time (e.g., "Connecting DHT22 sensor: 3 minutes"); Operation sequence constraints (e.g., "Power must be connected first, then signal lines"). Verification conditions (e.g., "temperature sensor readings should be within the range of 20-30°C").

[0077] Then, set the operation verification mechanism in the operation step editor. The setting process includes: Develop an operation monitoring engine to capture students' operational behavior in the virtual environment in real time and compare it with preset operation steps; Implement a time threshold detection algorithm to automatically trigger a prompt mechanism (such as "The current step has taken 4 minutes, it is recommended to check whether the wiring is correct") when a student's operation exceeds the expected time. Develop a sequence constraint verifier that uses graph algorithms to detect whether student operations violate sequence constraints. For example, when a student tries to upload code without installing the Arduino IDE, the system immediately prompts "Please complete the development environment configuration first".

[0078] After completing the 3D model construction, simulation effects are implemented. The implementation process includes, but is not limited to, the following: Develop a physical simulation engine to simulate sensor data generation (such as calculating temperature and humidity sensor readings based on environmental parameters), signal transmission (such as simulating RS485 bus signal attenuation), and device response (such as motor start-up time and delay characteristics). Implement an error injection mechanism to support the simulation of hardware failures (such as sensor short circuits and communication interruptions) in virtual training, thereby cultivating students' fault diagnosis capabilities. Develop a virtual oscilloscope function to support real-time monitoring and analysis of electrical signal waveforms, helping students understand the principles of signal processing.

[0079] In this embodiment of the application, an incremental update strategy is also adopted in constructing a teaching scenario-adaptive four-dimensional knowledge graph. When new IoT technology standards, device parameters, or practical training cases are generated, the knowledge graph nodes are updated using the following formula. attribute values ,in For nodes The original attribute value, These are newly added or updated attribute values. This embodiment uses an update strategy to ensure that the knowledge graph used in the teaching scenario is a dynamically updated knowledge graph, thereby effectively improving the teaching effect of vocational Internet of Things courses.

[0080] As can be seen from the above, the embodiments of this application, by constructing a teaching scenario-adaptive four-dimensional knowledge graph, designing a dual-channel generation architecture, executing a knowledge graph neural triggering algorithm, and deploying a cognitive trajectory tracking algorithm, achieve the following beneficial effects in application: First, the knowledge system is systematic and precise: the four-dimensional knowledge graph stores IoT knowledge in a structured way, covering four dimensions: concepts, devices, operations, and faults. The concept layer outlines principles and conceptual relationships; the device layer clarifies parameters and interfaces; the operation layer standardizes training processes; and the fault layer accumulates real-world cases. Compared to the fragmented knowledge of traditional large language models, this invention can accurately answer principle-based questions based on a complete knowledge system, avoiding conceptual confusion; when explaining the theoretical basis of hardware operation, it can also clearly connect knowledge nodes, providing students with accurate and systematic knowledge content.

[0081] Secondly, the practical training is made more efficient: the fault layer, combined with a knowledge graph neural triggering algorithm, significantly improves the equipment fault diagnosis capability. Through historical training error cases and solutions, this embodiment overcomes the bottleneck of low diagnostic accuracy in traditional large language models, providing accurate diagnosis and effective solutions for problems encountered by students during training, such as abnormal sensor data and equipment communication interruptions. Simultaneously, the operational layer's training instruction set and execution logic provide clear and standardized guidance for students' practical operations, significantly improving the efficiency of students' mastery of practical skills and making the teaching more aligned with the highly practical needs of vocational IoT courses.

[0082] Third, flexible knowledge updates and adaptation: The knowledge graph, which adapts to teaching scenarios, has dynamic update capabilities, solving the problems of long update cycles and inability to keep up with the rapid iteration of IoT technology in static knowledge graphs. Whether it's a new IoT protocol or hardware device update, the knowledge graph can keep pace with industry standards in a timely manner. Furthermore, its key teaching elements, such as device wiring diagrams and fault code tables, fully consider the cognitive characteristics of vocational school students and the actual needs of teaching, ensuring that teaching resources are always in sync with industry development and adapted to vocational school teaching scenarios.

[0083] Fourth, precise interaction and intelligent learning: The dual-channel generation architecture and the knowledge graph neural triggering algorithm work together to accurately interpret students' question intentions, avoid triggering incorrect subgraphs by keyword searches, and achieve precise interaction. The cognitive trajectory tracking algorithm records students' learning paths in real time, predicts knowledge blind spots through graph neural networks, and optimizes the prompting process based on blind spot analysis to provide students with personalized learning guidance. This intelligent interaction and personalized learning support effectively improves students' learning outcomes and experience, and promotes students' knowledge absorption and ability enhancement.

[0084] Please see Figure 2 This application also provides a teaching data analysis device for IoT courses in secondary vocational schools, the device comprising: The first building unit 210 is used to build a four-dimensional knowledge graph that is adaptive to teaching scenarios; The second building unit 220 is used to build a dual-channel generation architecture, which includes a semantic generation channel and a graph logic channel. The adjustment unit 230 is used to semantically adjust the dialogue question information input by the student through the semantic generation channel to obtain the dialogue question semantic vector; The execution unit 240 is used to execute the knowledge graph neural algorithm to generate dialogue answer information corresponding to the semantic vector of the dialogue question after inputting a four-dimensional knowledge graph through the graph logic channel. The optimization unit 250 is used to optimize the teaching question and answer prompts based on the dialogue question information and dialogue answer information using a cognitive trajectory tracking algorithm.

[0085] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0086] This application also provides a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method. This computer device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0087] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0088] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0089] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0090] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0091] It is understood that the content of the above method embodiments is applicable to the embodiments of this program product. The specific functions implemented by the embodiments of this program product are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0092] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0093] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0094] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0095] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0096] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0097] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0098] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0099] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0100] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A method for analyzing teaching data in an Internet of Things (IoT) course at a secondary vocational school, characterized in that, The method includes the following steps: Constructing a four-dimensional knowledge graph that adapts to teaching scenarios; A dual-channel generation architecture is constructed, which includes a semantic generation channel and a graph logic channel; The semantic generation channel is used to semantically adjust the dialogue question information input by the student to obtain a dialogue question semantic vector; After the four-dimensional knowledge graph is input through the graph logic channel, the knowledge graph neural algorithm is executed to generate the dialogue answer information corresponding to the semantic vector of the dialogue question. Based on the dialogue question information and the dialogue answer information, the teaching question and answer prompt information is optimized using a cognitive trajectory tracking algorithm.

2. The method according to claim 1, characterized in that, The four-dimensional knowledge graph includes: The concept layer is used to store the principles and conceptual relationships of the Internet of Things; The device layer is used to store the parameters and interface specifications of sensors or actuators. The operation layer is used to store the training operation instruction set and execution logic; The fault layer is used to store error cases and solutions extracted from historical training logs.

3. The method according to claim 1, characterized in that, The semantic generation channel includes a deep learning model based on an attention mechanism. The semantic generation channel performs semantic adjustment on the student's input dialogue question information to obtain a dialogue question semantic vector, including: The model parameters of the deep learning model are adjusted according to the learning stage information of secondary vocational school students; The deep learning model, after adjusting its model parameters, performs semantic adjustment on the dialogue question information input by the student to obtain the dialogue question semantic vector.

4. The method according to claim 1, characterized in that, After inputting the four-dimensional knowledge graph through the graph logic channel, the knowledge graph neural algorithm is executed to generate dialogue response information corresponding to the semantic vector of the dialogue question, including: The semantic vector of the dialogue question is parsed using a teaching scenario classifier to obtain the user's question intent information; Match the target subgraph corresponding to the user's question intent information in the four-dimensional knowledge graph; The target subgraph is aligned with the dialogue question semantic vector to obtain an aligned feature vector. The output information generated after logical verification based on the aligned feature vector is used as the dialogue response information.

5. The method according to claim 4, characterized in that, The teaching scenario classifier includes: The fault diagnosis module is used to identify first query information, which includes sensor anomaly query information, communication protocol error query information, and power problem query information. The operation process guidance module is used to identify the second query information, which includes device wiring query information, code burning query information, and platform configuration query information. The principle concept analysis module is used to identify third query information, which includes protocol principle query information, data transmission mechanism query information, and system architecture query information.

6. The method according to claim 4, characterized in that, The step of matching the target subgraph corresponding to the user's question intent information in the four-dimensional knowledge graph includes: Calculate the similarity between each candidate subgraph in the four-dimensional knowledge graph and the user's question intent information; The target subgraph is determined from the candidate subgraphs based on the similarity.

7. The method according to claim 1, characterized in that, The step of optimizing the teaching question-and-answer prompts based on the dialogue question information and the dialogue answer information using a cognitive trajectory tracking algorithm includes: The dialogue question information and the dialogue answer information are mapped to a path sequence of the four-dimensional knowledge graph; The path sequence is input into a graph neural network to predict the probability distribution of knowledge blind spot nodes; The teaching question and answer prompts are optimized based on the probability distribution of the knowledge blind spots.

8. A teaching data analysis device for an Internet of Things (IoT) course in a secondary vocational school, characterized in that, The device includes: The first building block is used to construct a four-dimensional knowledge graph that adapts to teaching scenarios; The second building unit is used to build a dual-channel generation architecture, which includes a semantic generation channel and a graph logic channel. The adjustment unit is used to semantically adjust the dialogue question information input by the student through the semantic generation channel to obtain the dialogue question semantic vector; An execution unit is used to execute a knowledge graph neural algorithm to generate dialogue response information corresponding to the semantic vector of the dialogue question after inputting the four-dimensional knowledge graph through the graph logic channel. The optimization unit is used to optimize the teaching question and answer prompts based on the dialogue question information and the dialogue answer information using a cognitive trajectory tracking algorithm.

9. A computer device, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1 to 7.

Images