Knowledge graph construction method based on grammar teaching content

Abstract

The invention discloses a knowledge graph construction method based on grammar teaching content, relates to the technical field of knowledge graph construction, and solves the problems that similar grammar boundaries are fuzzy, association relationships are difficult to sort and a structured knowledge system is difficult to form. The method comprises the following steps: firstly, constructing a tree-shaped branch graph in each similar partition according to assignment column sorting, so that the same grammar content forms a unified branch, the different content forms a hierarchical sub-branch, and the vein of grammar knowledge in a single partition is ensured to be clear; then organic integration of a multi-partition graph is achieved through partition feature mean value sorting, the finally formed complete knowledge graph presents independence of all grammar units and highlights relevance among different grammar systems, a panoramic cognitive framework from a micro grammar point to a macro grammar system is provided for a user, and the knowledge graph has the advantages that the knowledge graph is more interesting and more interesting. And the systematicness and the understandability of grammatical knowledge are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of knowledge graph construction technology, specifically a method for constructing knowledge graphs based on grammar teaching content. Background Technology

[0002] In the field of grammar teaching, with the continuous enrichment of teaching resources, characteristic grammar rules and accompanying examples are characterized by their large quantity and diverse types. Their effective organization and efficient application have become the key to improving teaching quality.

[0003] However, existing technologies have significant limitations in grammatical knowledge management:

[0004] On the one hand, the classification of characteristic grammars often relies on human experience or simple keyword division, lacking a unified quantitative standard. This leads to blurred boundaries between similar grammars, difficulty in sorting out relationships, and difficulty in forming a structured knowledge system. This is not only unfavorable for teachers to explain systematically, but also creates obstacles for learners to understand grammatical logic.

[0005] On the other hand, the methods for retrieving grammatical knowledge are relatively traditional, often employing item-by-item traversal or basic keyword matching. When faced with massive amounts of content, this not only results in a lengthy retrieval process but also makes it difficult to quickly locate target grammatical points and their deeper connections, severely restricting teaching efficiency and the speed of knowledge acquisition. Furthermore, existing knowledge graph construction methods are not well-suited for grammar teaching scenarios, failing to fully consider the unique attributes of grammatical rules such as hierarchy and correlation, leading to often loosely structured graphs that cannot accurately reflect the inherent logic between different grammatical features.

[0006] These problems make it difficult to fully realize the value of grammar teaching resources. There is an urgent need for a technical solution that can accurately classify characteristic grammars, construct systematic maps, and retrieve information efficiently, so as to meet the actual needs of grammar teaching for structured knowledge management and rapid application. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides a method for constructing knowledge graphs based on grammar teaching content, which solves the problems of unclear boundaries between similar grammars, difficulty in sorting out relationships, and difficulty in forming a structured knowledge system.

[0008] To achieve the above objectives, this invention provides the following technical solution: a knowledge graph construction method based on grammar teaching content, comprising the following steps:

[0009] Step 1: Assign values ​​to the characteristic grammar of different types of grammar teaching content, identify the different assignment columns associated with different characteristic grammars, then classify the identified assignment columns by feature, and simultaneously classify the associated characteristic grammars, identifying several different similar partitions. The specific method is as follows:

[0010] Based on the preset assignment table, confirm the assignment associated with different grammatical keywords, and sort the assignments corresponding to each position grammatical keyword according to the sorting method of the feature grammar, and confirm the assignment column associated with the corresponding feature grammar.

[0011] Confirm the positional features associated with each assignment column: Confirm the assignments associated with different positions within the corresponding assignment column, extract the first assignment associated with each assignment, sort the extracted first assignments, confirm the first assignment column, extract the second assignment associated with each assignment, sort the extracted second assignments, confirm the second assignment column, and so on, select and sort the assignments at each of the same arrangement positions, confirm the assignment column associated with the corresponding order;

[0012] Perform class-based verification on several feature syntaxes: verify the first assignment column associated with each feature syntax, and verify the value differences associated with each first assignment column. Denote the values ​​associated with different sorting positions within the first assignment column as S. i Where i represents different sorting positions, and then the values ​​S at the same sorting positions are... i Perform difference processing to confirm the undetermined difference. If the undetermined difference is ≥0, perform average processing on several groups of undetermined differences confirmed at each sorting position to lock the confirmed value. Divide the two groups of feature syntaxes with confirmed values ​​≤Y1 into the same type of syntax. If the confirmed value is >Y1, do not perform the labeling of the same type of syntax.

[0013] Then, the syntaxes classified into the same category are re-confirmed. This confirmation process combines the second-order assignment columns associated with different feature syntaxes and other assignment columns associated thereafter. The first assignment column is used to confirm the same confirmation method for the same type of syntaxes. The different feature syntaxes are re-classified, so that several feature syntaxes of the same type of syntax are divided into multiple different classification partitions. The feature syntaxes associated with each type of partition are marked.

[0014] Step 2: Construct a knowledge graph for the feature grammars associated with each similar partition. Based on the different feature grammars associated with each similar partition, and based on the different assignment columns associated with different feature grammars, classify and construct a knowledge graph for several feature grammars. This completes the knowledge graph construction process associated with each similar partition. The specific method is as follows:

[0015] The feature syntax of the same type of partition is confirmed, and the assignment column associated with each feature syntax is confirmed. The associated feature syntax is arranged in ascending order of the assignment column. The knowledge graph construction process of several feature syntaxes of the same type of partition is completed according to the determined arrangement order.

[0016] Then, the average value of several assigned columns associated with each partition of the same type is processed and used as the partition feature of the corresponding partition of the same type. The knowledge graphs of several partitions of the same type are integrated according to the sorting of partition features from smallest to largest to obtain a complete knowledge graph.

[0017] Step 3: Reprocess the constructed complete knowledge graph. Identify the columns associated with different feature syntaxes from top to bottom, and label the assigned columns associated with each feature syntax according to their column location. Determine the strongest association column and record it. The specific method is as follows:

[0018] From the complete knowledge graph, identify the columns containing different feature syntaxes from top to bottom; these columns are labeled Z. q Where q represents different feature syntaxes, and Z q =1, 2, ..., n;

[0019] Assign the column associated with the corresponding feature syntax and the column Z it belongs to. q Perform validation: Identify the assignments associated with different syntax keywords from the assignment column, randomly sum the associated assignments, confirm the total sum, execute several random summation processes, and compare the total sum associated with each random summation process with the corresponding column Z. q Perform interpolation using Cz. q =|Z q -Sum Values ​​| Identify the differences Cz associated with different sum values. q Then, from the confirmed differences Cz q In the middle, select the minimum value Cz q min, and Cz q The random summation process associated with min is called the confirmation process. The summation assignment in the confirmation process is called the selected assignment. The selected assignment is sorted from front to back to generate the strongest association column associated with the corresponding feature syntax.

[0020] First, identify the strongest correlation column recorded for each feature syntax, and sum the different assignments within the strongest correlation column to confirm the summation parameters;

[0021] Identify the column that matches the summation parameter from the complete knowledge graph, and use the column as the index source to perform grammatical indexing from top to bottom to lock the corresponding feature grammar.

[0022] Then, the relevant content associated with the locked feature syntax will be displayed.

[0023] This invention provides a method for constructing a knowledge graph based on grammar teaching content. Compared with existing technologies, it has the following advantages:

[0024] This invention introduces a multi-level assignment column classification mechanism. By using a preset assignment table, the feature syntax is transformed into a quantifiable assignment sequence. Then, through hierarchical comparison of the first, second, and subsequent assignment columns, combined with the mean difference judgment, a progressive classification of the feature syntax is achieved. This hierarchical classification logic not only preserves the integrity of the grammatical features but also accurately captures the subtle differences between different syntaxes, making the grouping of similar syntaxes more logical and consistent. It effectively avoids the boundary ambiguity problem caused by reliance on subjective judgment in traditional classification, laying a structured and standardized foundation for the construction of knowledge graphs.

[0025] A two-step strategy of "partition construction - integration and optimization" is adopted. First, a tree-like branching graph is constructed in each similar partition according to the assignment column sorting. This makes the same grammatical content form a unified branch and the different content form hierarchical sub-branches, ensuring that the grammatical knowledge in a single partition is clear. Then, the average value of partition features is used to sort the graphs of multiple partitions to achieve organic integration. The final complete knowledge graph not only presents the independence of each grammatical unit, but also highlights the correlation between different grammatical systems. It provides users with a panoramic cognitive framework from micro grammatical points to macro grammatical systems, which greatly improves the systematicness and comprehensibility of grammatical knowledge.

[0026] The design of the strongest correlation column breaks through the traditional column-by-column retrieval mode. By calculating the difference between the random sum and the column it belongs to, the optimal correlation sequence is locked, giving each feature grammar a unique "retrieval label". This design transforms grammar retrieval from linear scanning to precise positioning based on feature values. Users can directly lock the target column by using the sum parameter of the strongest correlation column, and then quickly match feature grammar and related example content by combining upper and lower indexes, significantly reducing retrieval time. Especially for teaching scenarios containing a large number of grammar rules and examples, this mechanism can help learners or teachers quickly locate target content, effectively improve knowledge acquisition efficiency, and enhance the practicality and convenience of grammar teaching. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Please see Figure 1 This application provides a method for constructing a knowledge graph based on grammar teaching content, including the following steps:

[0030] Step 1: Assign values ​​to the characteristic grammars of different types of grammar teaching content, identify the different assignment columns associated with different characteristic grammars, classify the identified assignment columns, and simultaneously classify the associated characteristic grammars to identify several different similar partitions. Specifically, in the corresponding partition classification process, based on the correlation between corresponding grammars, multiple different characteristic grammars are classified. According to the specific classification process, different characteristic grammars can be effectively divided into different partitions to carry out the specific division of similar partitions.

[0031] The specific method for classifying feature syntax is as follows:

[0032] Based on the preset assignment table, confirm the assignment associated with different grammatical keywords, and sort the assignments corresponding to the grammatical keywords at each position according to the sorting method of the feature grammar, and confirm the assignment column associated with the corresponding feature grammar. Specifically, within the corresponding feature grammar, for example, the keywords of the grammar corresponding to "subject-verb-object" all correspond to different assignments, which can be confirmed in the corresponding assignment table. Assuming that the assignment associated with the subject is 11, the assignment associated with the predicate is 12, and the assignment associated with the subsequent object is 13, then the assignment column associated with the corresponding feature grammar is: 11-12-13.

[0033] Confirm the positional features associated with each assignment column: Confirm the assignments associated with different positions within the corresponding assignment column, and extract the first assignment associated with each assignment. Sort the extracted first assignments to confirm the first assignment column. Then extract the second assignment associated with each assignment and sort the extracted second assignments to confirm the second assignment column. And so on, select and sort the assignments at each position with the same arrangement to confirm the assignment column associated with the corresponding order. Specifically, taking the associated assignment column 11-12-13 as an example, first confirm the first assignment, which is "1, 1, 1". The confirmed first assignment column is: 111. Then confirm the second assignment column. The confirmed second assignment column is "123".

[0034] Perform class-based verification on several feature syntaxes: verify the first assignment column associated with each feature syntax, and verify the value differences associated with each first assignment column. Denote the values ​​associated with different sorting positions within the first assignment column as S. i Where i represents different sorting positions, and then the values ​​S at the same sorting positions are... iPerform difference processing to confirm the undetermined difference value. If the undetermined difference value is ≥0 (that is, after difference processing, absolute value processing is still required), perform mean processing on several groups of undetermined difference values ​​confirmed at each sorting position, lock the confirmed value, and classify the two groups of feature syntax with confirmed value ≤Y1 into the same type of syntax. Otherwise, do not perform the same type of syntax marking.

[0035] Then, the syntaxes classified into the same category are reconfirmed. This confirmation process combines the second-order assignment columns associated with different feature syntaxes and other assignment columns associated thereafter. The first assignment column is used to confirm the same confirmation method for syntaxes in the same category. Different feature syntaxes are reclassified again (that is, based on the first classification, they are divided again), so that several feature syntaxes in the same category are divided into multiple different classification partitions. The feature syntaxes associated with each category partition are marked.

[0036] Specifically, each feature grammar has different assignment columns. According to the confirmation process of the corresponding assignment, each assignment column has a first assignment column, a second assignment column, ..., and subsequent assignment columns of other orders. First, based on the first assignment column, the same type of grammar is initially labeled. Within the corresponding labeled partition, after the confirmation process of the subsequent second assignment columns, ... and other related assignment columns, different classification partitions are locked, so that several different feature grammars are divided into several different partitions of the same type. Each partition of the same type contains different grammatical content, so that each grammatical content can be effectively divided.

[0037] Step 2: Construct a graph for the feature grammar associated with each similar partition. Based on the different feature grammars associated with each similar partition and the different assignment columns associated with different feature grammars, construct a graph for several feature grammars, thus completing the knowledge graph construction process associated with each similar partition. Specifically, different feature grammars are associated with different partitions. In order to complete the corresponding graph classification construction process between different feature grammars, the graph is constructed according to the assignment columns associated with each feature grammar.

[0038] The specific method for constructing a graph classification of several feature grammars is as follows:

[0039] The feature syntax belonging to the same partition is identified, and the assignment column associated with each feature syntax is identified. The associated feature syntaxes are then arranged according to the order of the assignment columns from smallest to largest. The knowledge graph construction process of several feature syntaxes in the same partition is completed according to the determined arrangement order. Specifically, when constructing the knowledge graph, branches are constructed sequentially from front to back for each feature syntax. The same syntax content is identified by the same branch. Subsequent different syntax content with the same syntax content in the previous order is sorted by multiple branches. The overall assignment column associated with the feature syntax at the top of the order is smaller, and the assignment column gradually increases as it goes down. This process is repeated to complete the knowledge graph construction process of several feature syntaxes. Since the knowledge graph construction process is similar to the existing tree diagram, it will not be described in detail here.

[0040] Then, the average value of several assigned columns associated with each partition of the same type is processed and used as the partition feature of the corresponding partition of the same type. The knowledge graphs of several partitions of the same type are integrated according to the sorting of partition features from smallest to largest to obtain a complete knowledge graph.

[0041] Step 3: Reprocess the constructed complete knowledge graph (that is, the overall graph associated with all similar partitions), identify the columns associated with different feature syntaxes from top to bottom, and mark the assigned columns associated with each feature syntax according to the columns they are in, determine the strongest association column, and record the strongest association column to facilitate fast indexing of feature syntaxes in the later stage and improve its indexing speed.

[0042] The specific method for confirming the strongest association column for each feature syntax association is as follows:

[0043] From the complete knowledge graph, identify the columns containing different feature syntaxes from top to bottom; these columns are labeled Z. q Where q represents different feature syntaxes, and Z q =1, 2, ..., n, Z q When =1, it means the corresponding feature syntax is in the first column, Z q When =2, it means that the corresponding feature syntax is located in the second column. According to the top-to-bottom sorting method, the different columns associated with different feature syntaxes can be identified in turn (that is, the different Arabic numerals they correspond to).

[0044] Assign the column associated with the corresponding feature syntax and the column Z it belongs to. q Perform validation: Identify the assignments associated with different syntax keywords from the assignment column, randomly sum the associated assignments, confirm the total sum, execute several random summation processes, and compare the total sum associated with each random summation process with the corresponding column Z. qPerform interpolation using Cz. q =|Z q -Sum Values ​​| Identify the differences Cz associated with different sum values. q Then, from the confirmed differences Cz q In the middle, select the minimum value Cz q min, and Cz q The random summation process associated with min is called the confirmation process. The summation assignments in the confirmation process are called the selected assignments. The selected assignments are sorted from front to back to generate the strongest association column associated with the corresponding feature syntax.

[0045] Once the strongest related column for each feature syntax is identified, the specific methods for subsequent content indexing include:

[0046] When indexing, first identify the strongest related column recorded by each feature syntax, and sum the different assignments in the strongest related column to confirm the summation parameters;

[0047] Identify the column that matches the summation parameter from the complete knowledge graph, and use the column as the index source to perform grammatical indexing from top to bottom to lock the corresponding feature grammar.

[0048] Next, the relevant content associated with the locked feature syntax will be displayed (specifically, each feature syntax is not just a single feature syntax, but also includes a large number of example contents, which are displayed here to facilitate the understanding of the feature syntax by search personnel).

[0049] Specifically, the original retrieval method involves sequentially verifying different feature syntaxes from top to bottom within the corresponding knowledge graph, identifying whether the verified feature syntaxes match the retrieved syntax content. If they match, the result is output; otherwise, the retrieval continues. However, this top-down retrieval method is too slow and cannot achieve a fast and effective retrieval result. According to the retrieval method in this embodiment, the column is quickly identified based on the strongest associated column, and the associated feature syntax is quickly identified by indexing up and down, thereby quickly locking in the associated content. Its indexing speed is faster and more efficient, achieving the fastest and best retrieval result.

[0050] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0051] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for constructing a knowledge graph based on grammar teaching content, characterized in that, Includes the following steps: Step 1: Assign values ​​to the characteristic grammars of different types of grammar teaching content, identify the different assignment columns associated with different characteristic grammars, classify the identified assignment columns by feature, and simultaneously classify the associated characteristic grammars to identify several different similar partitions. Step 2: Construct a graph for the feature grammar associated with each similar partition. Based on the different feature grammars associated with each similar partition and the different assignment columns associated with different feature grammars, construct a graph for several feature grammars, thus completing the knowledge graph construction process associated with each similar partition. Step 3: Reprocess the constructed complete knowledge graph, identify the columns associated with different feature syntaxes from top to bottom, mark the assignment columns associated with each feature syntax according to the column they are in, determine the strongest association column, and record the strongest association column.

2. The knowledge graph construction method based on grammar teaching content according to claim 1, characterized in that, In step one, the specific method for classifying feature syntax is as follows: Based on the preset assignment table, confirm the assignment associated with different grammatical keywords, and sort the assignments corresponding to each position grammatical keyword according to the sorting method of the feature grammar, and confirm the assignment column associated with the corresponding feature grammar. Confirm the positional features associated with each assignment column: Confirm the assignments associated with different positions within the corresponding assignment column, extract the first assignment associated with each assignment, sort the extracted first assignments, confirm the first assignment column, extract the second assignment associated with each assignment, sort the extracted second assignments, confirm the second assignment column, and so on, select and sort the assignments at each of the same arrangement positions, confirm the assignment column associated with the corresponding order; Perform class-based verification on several feature syntaxes: verify the first assignment column associated with each feature syntax, and verify the value differences associated with each first assignment column. Denote the values ​​associated with different sorting positions within the first assignment column as S. i Where i represents different sorting positions, and then the values ​​S at the same sorting positions are... i Perform difference processing to confirm the undetermined difference value. If the undetermined difference value is ≥0, perform average processing on several groups of undetermined difference values ​​confirmed at each sorting position to lock the confirmed value. Divide the two groups of feature syntax with confirmed value ≤Y1 into the same type of syntax. Then, the syntaxes classified into the same category are re-confirmed. This confirmation process combines the second-order assignment columns associated with different feature syntaxes and other subsequent assignment columns. The first assignment column is used to confirm the same confirmation method for syntaxes in the same category. Different feature syntaxes are re-classified, so that several feature syntaxes in the same category are divided into multiple different classification partitions. The feature syntaxes associated with each category partition are marked.

3. The knowledge graph construction method based on grammar teaching content according to claim 2, characterized in that, If the confirmation value is greater than Y1, then no labeling of the same syntax will be performed.

4. The knowledge graph construction method based on grammar teaching content according to claim 1, characterized in that, In step two, the specific method for constructing a graph classification of several feature grammars is as follows: The feature syntax of the same type of partition is confirmed, and the assignment column associated with each feature syntax is confirmed. The associated feature syntax is arranged in ascending order of the assignment column. The knowledge graph construction process of several feature syntaxes of the same type of partition is completed according to the determined arrangement order. Then, the average value of several assigned columns associated with each partition of the same type is processed and used as the partition feature of the corresponding partition of the same type. The knowledge graphs of several partitions of the same type are integrated according to the sorting of partition features from smallest to largest to obtain a complete knowledge graph.

5. The knowledge graph construction method based on grammar teaching content according to claim 1, characterized in that, In step three, the specific method for confirming the strongest related column is as follows: From the complete knowledge graph, identify the columns containing different feature syntaxes from top to bottom; these columns are labeled Z. q Where q represents different feature syntaxes, and Z q =1, 2, ..., n; Assign the column associated with the corresponding feature syntax and the column Z it belongs to. q Perform validation: Identify the assignments associated with different syntax keywords from the assignment column, randomly sum the associated assignments, confirm the total sum, execute several random summation processes, and compare the total sum associated with each random summation process with the corresponding column Z. q Perform interpolation using Cz. q =|Z q -Sum Values ​​| Identify the differences Cz associated with different sum values. q Then, from the confirmed differences Cz q In the middle, select the minimum value Cz q min, and Cz q The random summation process associated with min is called the confirmation process. The summation assignments in the confirmation process are called the selected assignments. The selected assignments are sorted from front to back to generate the strongest association column associated with the corresponding feature syntax.

6. The knowledge graph construction method based on grammar teaching content according to claim 5, characterized in that, Step three also includes subsequent retrieval steps, which specifically include: First, identify the strongest correlation column recorded for each feature syntax, and sum the different assignments within the strongest correlation column to confirm the summation parameters; Identify the column that matches the summation parameter from the complete knowledge graph, and use the column as the index source to perform grammatical indexing from top to bottom to lock the corresponding feature grammar. Then, the relevant content associated with the locked feature syntax will be displayed.

Images