Dynamically adjusted power standardization course system generation method and system

Abstract

The invention discloses a dynamically adjusted power standardized course system generation method and system, and the method comprises the steps: obtaining power training data from a power teaching material database, and carrying out the character feature extraction of knowledge keywords from the power training data; dividing the electric power training data according to knowledge keywords to obtain knowledge point units; training keywords are extracted from the training demand expression, the training keywords are used as indexes for internet retrieval to obtain extension features, the training keywords are expanded according to the extension features to obtain training elements, and the matching degree of the training elements and the knowledge point units is calculated; screening knowledge point units as training nodes according to the matching degree, and extracting association features among the training nodes; constructing a connection vector of the two training nodes according to the association features, and constructing a power standardization course system through the training nodes and the connection vector; according to the invention, personalized adjustment can be carried out on the power standardization course system according to actual demands.

Description

Technical Field

[0001] The present invention belongs to the field of electric power training courses, and in particular relates to a method and system for generating a dynamically adjusted electric power standardization course system. Background Art

[0002] With the continuous development of the power industry and the deepening of standardization work, the power industry's requirements for the training and capacity building of professional personnel are increasing. Standardization work in the power industry involves a wide range of technical and management fields, covering multiple aspects such as technical specifications, operating standards, and equipment management. With the changes in the industry environment and technological advancements, the power industry's requirements for talent are also showing a dynamic trend. Therefore, cultivating high-quality talents that can meet industry needs has become a key task for the industry's development.

[0003] In this context, the traditional training model has gradually exposed some problems. The traditional curriculum system is usually based on static textbooks or teaching outlines, lacks precise adaptation to the learners' personalized needs, and cannot dynamically respond to changes in industry and enterprise needs. The fixed learning content and rigid course arrangement often lead to gaps in learners' actual ability improvement, making it difficult to effectively improve students' comprehensive abilities and actual work levels. To this end, we propose a method for generating a standardized power curriculum system based on a competency model. Summary of the Invention

[0004] The present invention provides a method and system for generating a dynamically adjusted electric power standardization curriculum system, which can perform personalized adjustments to the electric power standardization curriculum system according to actual needs.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] A first aspect of the present invention provides a method for generating a dynamically adjusted electric power standardized curriculum system, comprising:

[0007] Obtaining power training data from a power teaching material database, extracting knowledge keywords from text features of the power training data; dividing the power training data according to the knowledge keywords to obtain knowledge point units;

[0008] Extract training keywords from the training needs statement, use the training keywords as an index to search the Internet to obtain extended features, expand the training keywords based on the extended features to obtain training elements, and calculate the matching degree between the training elements and knowledge point units;

[0009] Knowledge point units are selected as training nodes according to the matching degree, and the correlation features between the training nodes are extracted; the connection vectors of the two training nodes are constructed according to the correlation features, and the power standardization curriculum system is constructed through the training nodes and the connection vectors.

[0010] Furthermore, the method further includes: receiving training evaluation data of the electric power standardization curriculum system, and extracting longitudinal features reflecting the difficulty of training and transverse features reflecting the scope of training from the training evaluation data;

[0011] Calculate the correlation between the vertical features and the knowledge point units, and between the horizontal features and the knowledge point units, and establish the mapping relationship between the vertical features and the knowledge point units, and between the horizontal features and the knowledge point units;

[0012] Convert the vertical features into difficulty indexes and label features; when the difficulty index is negative, delete the power training data in the mapped knowledge point unit according to the difficulty index; when the difficulty index is positive, filter the power training data in the knowledge point unit according to the label features to obtain content fragments, and expand the content fragments according to the difficulty index to obtain training extension features; add the training extension features to the corresponding knowledge point units in the power standardization curriculum system;

[0013] The horizontal features are used as indexes to retrieve the supplementary data of power training in the power teaching materials database, and new training nodes are constructed based on the supplementary data of power training. The new training nodes are added to the power standardization curriculum system.

[0014] Furthermore, the content segments are expanded according to the difficulty index to obtain training extension features. The specific process is as follows:

[0015] Calculate the total number of words in the content segment based on the difficulty index. The expression formula is:

[0016]

[0017] In the formula, is the scaled total word count of the content snippet; is the total word count of the content snippet; To set the mapping coefficient; is an indicator of difficulty;

[0018] The scaled total word count and content snippet are input into the large language model to obtain training expansion features.

[0019] Furthermore, the power training data is subjected to text feature extraction to obtain knowledge keywords; the power training data is divided according to the knowledge keywords to obtain knowledge point units, specifically including:

[0020] Remove stop words, punctuation marks, pictures, numbers, and special symbols from the power training data to obtain the power training text; calculate the term frequency (TF) and inverse document frequency (IDF) of each word in the power training text, and calculate the weight of each word in the power training text based on the term frequency (TF) and inverse document frequency (IDF); sort the words in the power training text according to the weight and select knowledge keywords;

[0021] Use cosine similarity to group knowledge keywords to obtain the first sample cluster; if the Euclidean distance between the two first sample clusters is less than the preset minimum distance threshold , then by introducing the perturbation vector Push the Euclidean distance between the two first sample clusters to a greater distance than the preset maximum distance threshold. , then by introducing the contraction vector Shorten the Euclidean distance between the two first sample clusters;

[0022] Calculate the sharing probability between the first sample clusters, expressed as:

[0023]

[0024] In the formula, For the First sample subcluster With the First sample subcluster The sharing probability of For the First sample subcluster With the First sample subcluster The number of similar knowledge keywords; For the First sample subcluster Number of internal knowledge keywords; For the First sample subcluster Number of internal knowledge keywords;

[0025] The first sample clusters are merged and reorganized according to the sharing probability to obtain second sample subclusters; and the power training data is divided according to the second sample subclusters to obtain knowledge point units.

[0026] Furthermore, training keywords are extracted from the training demand statement, and the training keywords are used as indexes to search the Internet to obtain extended features. Based on the extended features, the training keywords are expanded to obtain training elements. The specific process is as follows:

[0027] Extract training keywords from the training needs statement, use the training keywords as an index to search the Internet to obtain the latest relevant documents; perform feature extraction on the latest relevant documents to obtain text paragraphs;

[0028] Calculate the correlation probability between the text paragraphs and the training keywords, and filter the text paragraphs according to the correlation probability to obtain extended features; input the extended features into the classifier to obtain the title of the extended features;

[0029] The title of the extended feature is input into the neural network model for information expansion to obtain training elements.

[0030] Furthermore, training keywords are extracted from the training needs statement, including:

[0031] The words in the training demand statement are regarded as demand nodes. If two demand nodes are adjacent in the training demand statement or co-occur within the set window range, a connecting edge is added between the two demand nodes to obtain a demand association graph;

[0032] Calculate the weight of each demand node in the demand association graph, and the expression formula is:

[0033]

[0034] In the formula, The first The weight of each demand node; The first The weight of the demand node; d is the damping coefficient; To point to A demand node set of demand nodes; For the demand nodes to The length of the connecting edge of each demand node; For the demand nodes to The length of the connecting edge of each demand node;

[0035] Sort the demand nodes according to their weights, and select the top N demand nodes with the largest weights as training keywords.

[0036] Furthermore, the title with extended features is input into the neural network model for information expansion to obtain training elements. The process includes:

[0037] The neural network model includes an input layer, a self-attention layer, a feedforward neural network layer and an output layer;

[0038] Inputting the title of the extended feature into the input layer of the neural network model to obtain a word vector;

[0039] The dependencies and semantic associations between word vectors are captured by multiple attention heads in the self-attention layer to obtain intermediate features. The expression formula is:

[0040]

[0041]

[0042]

[0043]

[0044] In the formula, is the query vector, is the key vector, is a value vector, is the linear matrix of word vectors; 、 and is the weight matrix; is an intermediate feature; is the dimension of the key vector; is the activation function; is the transpose of the matrix;

[0045] The intermediate features are concatenated and linearly transformed to obtain semantic features, which are then input into the feedforward neural network layer. The semantic features are nonlinearly transformed and then input into the output layer to obtain training elements.

[0046] Furthermore, the matching degree between training elements and knowledge point units is calculated, including:

[0047] ;

[0048] in, is the ability matching degree of the kth knowledge point, For the The weight of each training element, is the kth knowledge point and the The mapping value of the training element, is the total number of training elements.

[0049] A second aspect of the present invention provides a dynamically adjusted power standardization curriculum system generation system, comprising:

[0050] An acquisition module is used to acquire power training data from the power teaching material database, extract knowledge keywords from the power training data through text features, and divide the power training data into knowledge point units according to the knowledge keywords;

[0051] The extension module extracts training keywords from the training demand statement, uses the training keywords as an index to search the Internet to obtain extended features, and expands the training keywords based on the extended features to obtain training elements;

[0052] The output module calculates the matching degree between training elements and knowledge point units; selects knowledge point units as training nodes based on the matching degree, and extracts the correlation features between each training node; constructs the connection vector of two training nodes based on the correlation features, and constructs the power standardization curriculum system through the training nodes and connection vectors.

[0053] Furthermore, the system further includes an evaluation module, which receives training evaluation data of the electric power standardization curriculum system and extracts longitudinal features reflecting the difficulty of training and transverse features reflecting the scope of training from the training evaluation data;

[0054] Calculate the correlation between the vertical features and the knowledge point units, and between the horizontal features and the knowledge point units, and establish the mapping relationship between the vertical features and the knowledge point units, and between the horizontal features and the knowledge point units;

[0055] Convert the vertical features into difficulty indexes and label features; when the difficulty index is negative, delete the power training data in the mapped knowledge point unit according to the difficulty index; when the difficulty index is positive, filter the power training data in the knowledge point unit according to the label features to obtain content fragments, and expand the content fragments according to the difficulty index to obtain training extension features; add the training extension features to the corresponding knowledge point units in the power standardization curriculum system;

[0056] The horizontal features are used as indexes to retrieve the supplementary data of power training in the power teaching materials database, and new training nodes are constructed based on the supplementary data of power training. The new training nodes are added to the power standardization curriculum system.

[0057] The third aspect of the present invention provides an electronic device, comprising a storage medium and a processor; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the method for generating an electric power standardization curriculum system described in the first aspect.

[0058] A fourth aspect of the present invention provides a computer program product, comprising instructions, which, when executed by a processor, cause the processor to execute the method for generating a standardized electric power curriculum system according to the first aspect.

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] The present invention extracts training keywords from the training demand expression, uses the training keywords as indexes to conduct Internet search to obtain extended features, and expands the training keywords according to the extended features to obtain training elements. The present invention can obtain the latest extended features through Internet search in a timely manner, integrate the industry's cutting-edge knowledge and skills into the curriculum system, and ensure that the course content keeps pace with the times.

[0061] This method calculates the matching degree between training elements and knowledge point units; selects knowledge point units as training nodes based on the matching degree, and extracts the correlation features between each training node; constructs a connection vector between two training nodes based on the correlation features, constructs a standardized electric power curriculum system based on the training nodes and connection vectors, and selects training nodes based on the matching results to ensure that the selected knowledge points are highly consistent with the training requirements. This avoids the problems of insufficient knowledge point coverage or excessive redundancy in traditional curriculum systems, improving training quality and learning efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a flow chart of a method for generating a dynamically adjusted electric power standardization curriculum system provided in Example 1;

[0063] Figure 2 This is a flow chart of knowledge point unit generation provided in this embodiment 1;

[0064] Figure 3 This is a dynamic adjustment flow chart of the electric power standardization course system provided in this embodiment 1. DETAILED DESCRIPTION

[0065] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention.

[0066] Example 1

[0067] like Figure 1 As shown, this implementation provides a method for generating a dynamically adjusted electric power standardization curriculum system, including

[0068] Obtaining power training data from the power teaching material database, extracting knowledge keywords from the power training data through text features; and dividing the power training data according to the knowledge keywords to obtain knowledge point units, specifically including:

[0069] like Figure 2As shown, the stop words, punctuation marks, pictures, numbers and special symbols in the power training data are removed to obtain the power training text; the word frequency TF and inverse document frequency IDF of each word in the power training text are calculated, and the weight of each word in the power training text is calculated according to the word frequency TF and inverse document frequency IDF; the words in the power training text are sorted according to the weight and knowledge keywords are selected; the knowledge keywords include but are not limited to analyzing knowledge, skills, qualities and other ability elements; knowledge keywords are used to effectively extract the key ability elements in the field of power standardization, so that each ability element can be accurately understood and quantified, and provide a clear learning goal and ability improvement direction for each knowledge point.

[0070] Use cosine similarity to group knowledge keywords to obtain the first sample cluster; if the Euclidean distance between the two first sample clusters is less than the preset minimum distance threshold , then by introducing the perturbation vector Push the Euclidean distance between the two first sample clusters to a greater distance than the preset maximum distance threshold. , then by introducing the contraction vector Shorten the Euclidean distance between the two first sample clusters;

[0071] Calculate the sharing probability between the first sample clusters, expressed as:

[0072]

[0073] In the formula, For the First sample subcluster With the First sample subcluster The sharing probability of For the First sample subcluster With the First sample subcluster The number of similar knowledge keywords; For the First sample subcluster Number of internal knowledge keywords; For the First sample subcluster Number of internal knowledge keywords;

[0074] The first sample clusters are merged and reorganized according to the sharing probability to obtain second sample subclusters; and the power training data is divided according to the second sample subclusters to obtain knowledge point units.

[0075] This embodiment modularizes power training data, avoiding redundant or missing knowledge points and ensuring the practicality of the extracted knowledge points, allowing them to be directly applied to training and curriculum development. Knowledge point units are clearly mapped to key knowledge keywords, ensuring that course design is precisely aligned with industry needs.

[0076] Extract training keywords from the training needs statement, use the training keywords as an index to search the Internet to obtain extended features, and expand the training keywords based on the extended features to obtain training elements. The specific process is as follows:

[0077] The words in the training demand statement are regarded as demand nodes. If two demand nodes are adjacent in the training demand statement or co-occur within the set window range, a connecting edge is added between the two demand nodes to obtain a demand association graph;

[0078] Calculate the weight of each demand node in the demand association graph, and the expression formula is:

[0079]

[0080] In the formula, The first The weight of each demand node; The first The weight of the demand node; d is the damping coefficient; To point to A demand node set of demand nodes; For the demand nodes to The length of the connecting edge of each demand node; For the demand nodes to The length of the connecting edge of each demand node;

[0081] Sort the demand nodes according to their weights, and select the top N demand nodes with the largest weights as training keywords.

[0082] This implementation uses demand association graph modeling (converting words in demand descriptions into nodes and constructing connecting edges based on co-occurrence relationships) and TR weight calculation (assessing importance based on the strength of node associations) to accurately select high-frequency, highly relevant core words from complex training demand descriptions as training keywords. This approach avoids the subjectivity and omissions inherent in traditional manual keyword extraction, ensuring that training content closely aligns with actual needs and improving the alignment of courses with student competency development goals.

[0083] Use the training keywords as indexes to search the Internet to obtain the latest relevant documents; perform feature extraction on the latest relevant documents to obtain text paragraphs;

[0084] Calculate the correlation probability between the text paragraphs and the training keywords, and filter the text paragraphs according to the correlation probability to obtain extended features; input the extended features into the classifier to obtain the title of the extended features;

[0085] The title of the extended feature is input into the neural network model to expand the information and obtain the training elements. The process includes:

[0086] The neural network model includes an input layer, a self-attention layer, a feedforward neural network layer and an output layer;

[0087] Inputting the title of the extended feature into the input layer of the neural network model to obtain a word vector;

[0088] The dependencies and semantic associations between word vectors are captured by multiple attention heads in the self-attention layer to obtain intermediate features. The expression formula is:

[0089]

[0090]

[0091]

[0092]

[0093] In the formula, is the query vector, is the key vector, is a value vector, is the linear matrix of word vectors; 、 and is the weight matrix; is an intermediate feature; is the dimension of the key vector; is the activation function; is the transpose of the matrix;

[0094] The intermediate features are concatenated and linearly transformed to obtain semantic features, which are then input into the feedforward neural network layer. The semantic features are nonlinearly transformed and then input into the output layer to obtain training elements.

[0095] By calculating the correlation probability between text paragraphs and training keywords to screen extended features, and using neural network models to capture the dependencies between word vectors, the degree of correlation between different information and core needs can be quantitatively evaluated; by using classifiers and neural network models (such as the Transformer architecture with self-attention mechanism) to perform semantic analysis and information expansion on extended features, fragmented retrieval results can be converted into structured training elements.

[0096] Calculate the matching degree between training elements and knowledge point units, including:

[0097] ;

[0098] in, is the ability matching degree of the kth knowledge point, For the The weight of each training element, is the kth knowledge point and the The mapping value of the training element, is the total number of training elements.

[0099] Knowledge point units are selected as training nodes according to the matching degree, and the correlation features between the training nodes are extracted; the connection vectors of the two training nodes are constructed according to the correlation features, and the power standardization curriculum system is constructed through the training nodes and the connection vectors.

[0100] By using formulas to quantify the degree of match between each knowledge point and required competency, the curriculum system can be dynamically adjusted based on the diverse needs of students, ensuring that each student can maximize their competency during the learning process.

[0101] By classifying and combining the smallest knowledge point units according to course objectives and learner needs, course modules suitable for different learners are designed. The modular design not only enables flexible content arrangement, but also enables personalized learning path planning based on the learning objectives and ability levels of different learners; the connection vector enables learners to gradually master knowledge in sequence and level, thereby achieving the predetermined training objectives. The sequential design of course modules can help learners gradually learn from basic knowledge to complex knowledge, avoiding discontinuity in learning progress. The reasonable connection and combination between modules makes the learner's learning path more systematic and improves the overall training efficiency.

[0102] like Figure 3 As shown, the training evaluation data of the electric power standardization course system is received, and the vertical characteristics reflecting the difficulty of the training and the horizontal characteristics reflecting the scope of the training are extracted from the training evaluation data;

[0103] Calculate the correlation between the vertical features and the knowledge point units, and between the horizontal features and the knowledge point units, and establish the mapping relationship between the vertical features and the knowledge point units, and between the horizontal features and the knowledge point units;

[0104] The vertical features are converted into difficulty indexes and label features. When the difficulty index is negative, the power training data in the mapped knowledge point unit is deleted according to the difficulty index. When the difficulty index is positive, the power training data in the knowledge point unit is filtered according to the label features to obtain content fragments, and the content fragments are expanded according to the difficulty index to obtain training extension features.

[0105] Expand the content segments according to the difficulty index to obtain training extension features. The specific process is as follows:

[0106] Calculate the total number of words in the content segment based on the difficulty index. The expression formula is:

[0107]

[0108] In the formula, is the scaled total word count of the content snippet; is the total word count of the content snippet; To set the mapping coefficient; is an indicator of difficulty;

[0109] The scaled total word count and content snippet are input into the large language model to obtain training expansion features.

[0110] The horizontal features are used as indexes to retrieve the supplementary data of power training in the power teaching materials database, and new training nodes are constructed based on the supplementary data of power training. The new training nodes are added to the power standardization curriculum system.

[0111] This embodiment can dynamically adjust the course according to the learner's needs and learning progress. The real-time feedback and flexible adjustment mechanism during the learning process ensure that students can improve their abilities at a suitable pace.

[0112] A threshold-based module jump mechanism is introduced, and a dynamic threshold is set. When the learner's ability matching degree in a certain module is lower than the preset evaluation threshold, it will automatically trigger the jump to the intensive training state. The evaluation threshold calculation formula is:

[0113]

[0114] In the formula, For the The weight of each training element, is the dynamic threshold coefficient; is the evaluation threshold; n is the total number of training elements.

[0115] It can automatically adjust to a more suitable module for intensive training. This mechanism improves the flexibility and adaptability of the learning process and avoids the occurrence of learning obstacles. By setting thresholds, it ensures that learners can jump courses according to their actual ability level and avoid the frustration caused by overly complex modules. The dynamic threshold setting ensures that the course content is adjusted in real time according to the learner's ability level, enhancing the personalization and accuracy of the course.

[0116] Example 2

[0117] This embodiment discloses a dynamically adjusted electric power standardized course system generation system, which is used to execute the electric power standardized course system generation method described in Example 1. The electric power standardized course system generation system includes:

[0118] An acquisition module is used to acquire power training data from the power teaching material database, extract knowledge keywords from the power training data through text features, and divide the power training data into knowledge point units according to the knowledge keywords;

[0119] The extension module extracts training keywords from the training demand statement, uses the training keywords as an index to search the Internet to obtain extended features, and expands the training keywords based on the extended features to obtain training elements;

[0120] The output module calculates the matching degree between training elements and knowledge point units; selects knowledge point units as training nodes based on the matching degree, and extracts the correlation features between each training node; constructs the connection vector of two training nodes based on the correlation features, and constructs the power standardization curriculum system through the training nodes and connection vectors.

[0121] The evaluation module receives the training evaluation data of the electric power standardization course system, and extracts the vertical characteristics reflecting the difficulty of the training and the horizontal characteristics reflecting the scope of the training from the training evaluation data;

[0122] Calculate the correlation between the vertical features and the knowledge point units, and between the horizontal features and the knowledge point units, and establish the mapping relationship between the vertical features and the knowledge point units, and between the horizontal features and the knowledge point units;

[0123] Convert the vertical features into difficulty indexes and label features; when the difficulty index is negative, delete the power training data in the mapped knowledge point unit according to the difficulty index; when the difficulty index is positive, filter the power training data in the knowledge point unit according to the label features to obtain content fragments, and expand the content fragments according to the difficulty index to obtain training extension features; add the training extension features to the corresponding knowledge point units in the power standardization curriculum system;

[0124] The horizontal features are used as indexes to retrieve the supplementary data of power training in the power teaching materials database, and new training nodes are constructed based on the supplementary data of power training. The new training nodes are added to the power standardization curriculum system.

[0125] The evaluation module can effectively track learners' learning progress and knowledge mastery, and adjust course content in a timely manner based on learners' feedback to ensure learners' learning effects. Through real-time feedback, course content can be dynamically adjusted to avoid the situation where fixed course content cannot adapt to the needs of different learners. Learners' knowledge mastery progress can be tracked in real time to ensure maximum learning effects.

[0126] Example 3

[0127] This embodiment provides an electronic device, including a storage medium and a processor; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the method for generating an electric power standardization curriculum system described in Example 1.

[0128] Example 4

[0129] This embodiment provides a computer program product, including instructions, which, when executed by a processor, enable the processor to execute the method for generating a standardized electric power curriculum system described in Example 1.

[0130] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0131] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0132] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0133] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0134] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for generating a dynamically adjusted electric power standardization curriculum system, characterized in that: include: Obtain power training data from the power teaching material database and extract knowledge keywords from the power training data using text features; Dividing the power training data according to knowledge keywords to obtain knowledge point units; Extract training keywords from the training needs statement, use the training keywords as an index to search the Internet to obtain extended features, expand the training keywords based on the extended features to obtain training elements, and calculate the matching degree between the training elements and knowledge point units; Filter knowledge point units as training nodes based on matching degree, and extract correlation features between training nodes; The connection vector of two training nodes is constructed according to the correlation characteristics, and the power standardization curriculum system is constructed through the training nodes and the connection vector.

2. The method for generating a standardized electric power curriculum system according to claim 1, characterized in that: Also includes: Receive training evaluation data of the electric power standardization curriculum system, and extract vertical characteristics reflecting training difficulty and horizontal characteristics reflecting training scope from the training evaluation data; Calculate the correlation between the vertical features and the knowledge point units, and between the horizontal features and the knowledge point units, and establish a mapping relationship between the vertical features and the knowledge point units, and between the horizontal features and the knowledge point units; Convert the vertical features into difficulty indexes and label features; when the difficulty index is negative, delete the power training data in the mapped knowledge point unit according to the difficulty index; when the difficulty index is positive, filter the power training data in the knowledge point unit according to the label features to obtain content fragments, and expand the content fragments according to the difficulty index to obtain training extension features; add the training extension features to the corresponding knowledge point units in the power standardization curriculum system; The horizontal features are used as indexes to retrieve the supplementary data of power training in the power teaching materials database, and new training nodes are constructed based on the supplementary data of power training. The new training nodes are added to the power standardization curriculum system.

3. The method for generating a standardized electric power curriculum system according to claim 2, characterized in that: Expand the content segments according to the difficulty index to obtain training expansion features. The specific process is as follows: Calculate the total number of words in the content segment based on the difficulty index. The expression formula is: ； In the formula, is the scaled total word count of the content snippet; is the total word count of the content snippet; To set the mapping coefficient; is an indicator of difficulty; The scaled total word count and content snippet are input into the large language model to obtain training expansion features.

4. The method for generating a standardized electric power curriculum system according to claim 1, characterized in that: Extract knowledge keywords from text features of power training data; The power training data is divided according to knowledge keywords to obtain knowledge point units, specifically including: Remove stop words, punctuation marks, pictures, numbers, and special symbols from the power training data to obtain the power training text; calculate the term frequency (TF) and inverse document frequency (IDF) of each word in the power training text, and calculate the weight of each word in the power training text based on the term frequency (TF) and inverse document frequency (IDF); sort the words in the power training text according to the weight and select knowledge keywords; Use cosine similarity to group knowledge keywords to obtain the first sample cluster; if the Euclidean distance between the two first sample clusters is less than the preset minimum distance threshold , then by introducing the perturbation vector Push the Euclidean distance between the two first sample clusters to a greater distance than the preset maximum distance threshold. , then by introducing the contraction vector Shorten the Euclidean distance between the two first sample clusters; Calculate the sharing probability between the first sample clusters, expressed as: ； In the formula, For the First sample subcluster With the First sample subcluster The sharing probability of For the First sample subcluster With the First sample subcluster The number of similar knowledge keywords; For the First sample subcluster Number of internal knowledge keywords; For the First sample subcluster Number of internal knowledge keywords; The first sample clusters are merged and reorganized according to the sharing probability to obtain second sample subclusters; and the power training data is divided according to the second sample subclusters to obtain knowledge point units.

5. The method for generating a standardized electric power curriculum system according to claim 1, characterized in that: Extract training keywords from the training needs statement, use the training keywords as an index to search the Internet to obtain extended features, and expand the training keywords based on the extended features to obtain training elements. The specific process is as follows: Extract training keywords from the training needs statement, use the training keywords as an index to search the Internet to obtain the latest relevant documents; perform feature extraction on the latest relevant documents to obtain text paragraphs; Calculate the correlation probability between the text paragraphs and the training keywords, and filter the text paragraphs according to the correlation probability to obtain extended features; input the extended features into the classifier to obtain the title of the extended features; The title of the extended feature is input into the neural network model for information expansion to obtain training elements.

6. The method for generating a standardized electric power curriculum system according to claim 5, characterized in that: Extract training keywords from the training needs statement, including: The words in the training demand statement are regarded as demand nodes. If two demand nodes are adjacent in the training demand statement or co-occur within the set window range, a connecting edge is added between the two demand nodes to obtain a demand association graph; Calculate the weight of each demand node in the demand association graph, and the expression formula is: ； In the formula, The first The weight of each demand node; The first The weight of the demand node; d is the damping coefficient; To point to A demand node set of demand nodes; For the demand nodes to The length of the connecting edge of each demand node; For the demand nodes to The length of the connecting edge of each demand node; Sort the demand nodes according to their weights, and select the top N demand nodes with the largest weights as training keywords.

7. The method for generating a standardized electric power curriculum system according to claim 5, characterized in that: The title of the extended feature is input into the neural network model to expand the information and obtain the training elements. The process includes: The neural network model includes an input layer, a self-attention layer, a feedforward neural network layer and an output layer; Inputting the title of the extended feature into the input layer of the neural network model to obtain a word vector; The dependencies and semantic associations between word vectors are captured by multiple attention heads in the self-attention layer to obtain intermediate features. The expression formula is: ； ； ； ； In the formula, is the query vector, is the key vector, is a value vector, is the linear matrix of word vectors; 、 and is the weight matrix; is an intermediate feature; is the dimension of the key vector; is the activation function; is the transpose of the matrix; The intermediate features are concatenated and linearly transformed to obtain semantic features, which are input into the feedforward neural network layer; the semantic features are nonlinearly transformed and then input into the output layer to obtain training elements.

8. The method for generating a standardized electric power curriculum system according to claim 1, characterized in that: Calculate the matching degree between training elements and knowledge point units, including: ； in, is the ability matching degree of the kth knowledge point, For the The weight of each training element, is the kth knowledge point and the The mapping value of the training element, is the total number of training elements.

9. A dynamically adjusted power standardization curriculum system generation system, characterized by: include: An acquisition module is used to obtain power training data from the power teaching material database and extract knowledge keywords from the power training data through text features; Dividing the power training data according to knowledge keywords to obtain knowledge point units; The extension module extracts training keywords from the training demand statement, uses the training keywords as an index to search the Internet to obtain extended features, and expands the training keywords based on the extended features to obtain training elements; Output module, calculates the matching degree between training elements and knowledge point units; Filter knowledge point units as training nodes based on matching degree, and extract correlation features between training nodes; The connection vector of two training nodes is constructed according to the correlation characteristics, and the power standardization curriculum system is constructed through the training nodes and the connection vector.

10. The power standardization course system generation system according to claim 9, characterized in that: The system further includes an evaluation module, which receives training evaluation data of the electric power standardization curriculum system and extracts longitudinal features reflecting training difficulty and transverse features reflecting training scope from the training evaluation data; Calculate the correlation between the vertical features and the knowledge point units, and between the horizontal features and the knowledge point units, and establish a mapping relationship between the vertical features and the knowledge point units, and between the horizontal features and the knowledge point units; Convert the vertical features into difficulty indexes and label features; when the difficulty index is negative, delete the power training data in the mapped knowledge point unit according to the difficulty index; when the difficulty index is positive, filter the power training data in the knowledge point unit according to the label features to obtain content fragments, and expand the content fragments according to the difficulty index to obtain training extension features; add the training extension features to the corresponding knowledge point units in the power standardization curriculum system; The horizontal features are used as indexes to retrieve the supplementary data of power training in the power teaching materials database, and new training nodes are constructed based on the supplementary data of power training. The new training nodes are added to the power standardization curriculum system.

11. An electronic device comprising a storage medium and a processor; the storage medium is used to store instructions; characterized in that, The processor is configured to operate according to the instructions to execute the method for generating an electric power standardized curriculum system according to any one of claims 1 to 8.

12. A computer program product comprising instructions, characterized in that When the instructions are executed by a processor, the processor executes the method for generating a standardized electric power curriculum system according to any one of claims 1 to 8.

Images