A method and system for constructing a path for inheriting Yi folk songs

Abstract

The present invention belongs to the field of cultural inheritance modeling and provides a method and system for constructing an inheritance path of Yi folk songs, comprising obtaining multi-source cultural element data of Yi folk songs and parsing the cultural element data into multiple semantic element nodes; constructing a semantic graph of Yi folk songs based on the semantic element nodes; clustering the semantic graph to generate multiple semantic association clusters, defining each inheritance path as a structural unit containing the above-mentioned various nodes; constructing an adaptability function based on learner portrait information to evaluate the matching degree between each inheritance path and the learner, selecting the inheritance path with the greatest matching degree, and generating a personalized Yi folk song inheritance path; collecting learner behavior data at each node, and feeding back the function value to update the path structure; archiving the finally constructed inheritance path and the learner behavior results to generate a path resume and inheritance trajectory graph.

Description

Technical Field

[0001] The present invention belongs to the field of cultural heritage modeling, and specifically relates to a method and system for constructing a Yi folk song inheritance path. Background Art

[0002] At present, the following methods are mainly used to protect and inherit Yi folk songs: first, static archiving through recording and video recording to form a digital resource library; second, periodic display through folk activities or intangible cultural heritage exhibitions; third, course teaching and vocal training by colleges and universities or cultural institutions. However, the above methods generally have the following problems:

[0003] First, the existing inheritance methods are mainly centered on content preservation and lack path structure design oriented to learner characteristics, resulting in a lack of hierarchy and adaptability in the learning process, making it difficult to effectively support individualized and sustainable inheritance goals.

[0004] Secondly, although some university courses have introduced Yi folk song teaching modules, most of them are in the form of elective courses or special courses. The teaching path is not systematic and lacks deep integration with the folk song's own culture, singing context and semantic structure, making it difficult to achieve the simultaneous development of culture and ability.

[0005] Thirdly, the existing teaching and inheritance process lacks dynamic evaluation mechanisms and behavioral feedback mechanisms. Learners' performance in different links cannot be captured and archived in a timely manner, resulting in a disconnect between teaching strategies and path recommendations, making it difficult to form a closed-loop inheritance system.

[0006] In addition, the content of Yi folk songs often involves special elements such as Yi dialect, cultural etiquette and polyphonic singing, which are difficult to learn. Without targeted adaptation strategies and path adjustment mechanisms, it is very easy for learners to give up halfway, affecting the continuity of inheritance. Summary of the Invention

[0007] In order to solve the problems in the prior art, the present invention provides a method for constructing a Yi folk song inheritance path, comprising the following steps:

[0008] Step S10, obtaining multi-source cultural element data of Yi folk songs, wherein the cultural element data includes audio data, lyrics text, singing context, festival background, singer information, and Yi language translation content, and parsing the cultural element data into multiple semantic element nodes;

[0009] Step S20, constructing a semantic graph of Yi folk songs based on the semantic element nodes, wherein each node in the graph represents a folk song knowledge unit, and the connecting edges between the nodes represent logical dependency relationships;

[0010] Step S30: clustering the semantic graph to generate multiple semantic association clusters, and constructing inheritance node clusters including native nodes, education nodes, practice nodes, and evaluation nodes based on the semantic association clusters, defining each inheritance path as a structural unit including the above-mentioned nodes;

[0011] Step S40: Based on the learner profile information, construct an adaptability function to evaluate the matching degree between each inheritance path and the learner, select the inheritance path with the greatest matching degree, and generate a personalized Yi folk song inheritance path;

[0012] Step S50: During the execution of the inheritance path, the learner's behavior data at each node is collected, including the melody restoration degree, semantic understanding degree, and practice completion degree indicators, and the evaluation function value of each node is calculated. The function value is then fed back to update the path structure;

[0013] Step S60: Archive the final constructed inheritance path and learner behavior results to generate a path resume and inheritance trajectory map.

[0014] Furthermore, the step S10 includes:

[0015] Step S101, collecting original singing data of Yi folk songs, including audio data and video materials of multiple parts, and marking the corresponding singers, regions and genealogy information;

[0016] Step S102: using a speech recognition system to transcribe the audio data into lyrics text, and bilingual personnel perform Yi language semantic annotation and translation;

[0017] Step S103, recording the singing context of folk songs based on field surveys, and classifying and coding them using multi-level labels;

[0018] Step S104: extracting festival backgrounds by combining the Yi calendar and local chronicles, and setting festival type and cultural embedding level labels for each folk song;

[0019] Step S105, constructing a performer information file, collecting the performer's name, gender, age and teacher lineage;

[0020] Step S106, establishing a nested structure of Yi language paragraphs and syllable pinyins and corresponding Chinese meanings in the lyrics;

[0021] Step S107: Structure the collected data into a set of semantic element nodes, and record context attributes and label information.

[0022] Furthermore, the step S20 includes:

[0023] Step S201: Define a semantic graph model using a heterogeneous graph structure, including multiple types of nodes such as lyrics nodes, melody nodes, festival nodes, and singer nodes;

[0024] Step S202: number the semantic element nodes and create a node attribute table to record their types, semantic labels, and cultural level information;

[0025] Step S203: calling the association rule mining algorithm to calculate the semantic association between nodes based on the node occurrence frequency and co-occurrence frequency, and constructing a connecting edge if the association is greater than a set threshold;

[0026] Step S204: performing knowledge migration and manual completion operations on isolated nodes;

[0027] Step S205: Save the semantic graph using a graph database or RDF triple format.

[0028] Furthermore, the step S30 includes:

[0029] Step S301: extract the attributes of all nodes in the semantic graph and construct a node vector containing word embedding, structure encoding, function label and context level;

[0030] Step S302, clustering the node vectors using a graph clustering algorithm;

[0031] Step S303, dividing the nodes in the clustering result into native nodes, education nodes, practice nodes and evaluation nodes;

[0032] Step S304: setting a standard inheritance path structure unit, including at least one original node, one education node, one practice node, and one evaluation node;

[0033] Step S305 : Combining the path structure based on the node connection relationship, generating a path number and storing it in the path library.

[0034] Furthermore, the step S40 includes:

[0035] Step S401: Collect the learner's ethnic identity, language ability, vocal foundation, preference tags, and historical learning behavior to construct a multi-dimensional learner portrait vector;

[0036] Step S402: extracting content attributes, language type, rhythm characteristics, and teaching difficulty information from the path structure to construct a path attribute vector;

[0037] Step S403: Calculate the compatibility score between the learner profile and each path using a weighted cosine similarity function;

[0038] Step S404: Select the path with the highest adaptability score as the personalized inheritance path for the current learner;

[0039] Step S405: store the binding relationship between the learner and the path, and synchronize it to the system database.

[0040] Furthermore, the step S50 includes:

[0041] Step S501: Before the path is executed, a node behavior collection module is configured to collect melody imitation, question understanding, practice record, and score feedback behavior data according to node type;

[0042] Step S502: Calculate standardized index values for the collected data, including melody restoration, semantic understanding, and practical completion;

[0043] Step S503: Setting a weighted evaluation function to integrate multiple indicators to calculate the node comprehensive score;

[0044] Step S504: Bind the evaluation score to the path structure, and adjust the node recommendation weight or insert a supplementary node according to the score.

[0045] Furthermore, the step S60 includes:

[0046] Step S601, creating a path history data structure to record the learner ID, path number, node execution order, completion status, score value and timestamp;

[0047] Step S602, archiving the path history in a structured format into a path database;

[0048] Step S603: Constructing a inheritance trajectory graph, where nodes are inheritance nodes in the history, edges are execution order relationships between nodes, and execution time and state attributes are added to the edges;

[0049] Step S604 : Visualize the learning path based on the trajectory graph, identify learning bottlenecks, and feed the trajectory data back to the path recommendation module.

[0050] Another aspect of the present invention provides a system for constructing a Yi folk song inheritance path, comprising the following modules:

[0051] A multi-source cultural element acquisition module is used to obtain multi-source cultural element data of Yi folk songs, including audio data, lyrics, singing context, festival background, singer information and Yi language translation content, and parse the cultural elements into multiple semantic element nodes;

[0052] A semantic graph construction module is used to construct a semantic graph of Yi folk songs based on the semantic element nodes, wherein each node in the semantic graph represents a folk song knowledge unit, and the connecting edges between the nodes represent logical dependency relationships;

[0053] A node clustering and path construction module is used to cluster the semantic graph to generate multiple semantic association clusters, and to construct inheritance node clusters including native nodes, education nodes, practice nodes, and evaluation nodes based on the semantic association clusters, defining each inheritance path as a structural unit including the above-mentioned nodes;

[0054] The path matching and recommendation module is used to construct an adaptability function based on the learner profile information to evaluate the matching degree between each inheritance path and the learner, select the inheritance path with the greatest matching degree, and generate a personalized inheritance path for Yi folk songs;

[0055] The behavior collection and evaluation module is used to collect learners' behavior data at each node during the execution of the inheritance path, including the melody restoration degree, semantic understanding degree and practice completion degree indicators, calculate the evaluation function value of each node, and feed the function value back to the path structure;

[0056] The path archiving and trajectory mapping module is used to archive the final constructed inheritance path and learner behavior results, and generate a path resume and inheritance trajectory map.

[0057] Furthermore, the multi-source cultural element acquisition module includes:

[0058] Audio collection unit, used to collect audio and video materials of Yi folk songs with multiple voice parts, and annotate the singer, region and genealogy information;

[0059] The speech transcription and translation unit is used to transcribe the audio into lyrics using a speech recognition system, and bilingual personnel perform Yi language annotation and translation;

[0060] The context annotation unit is used to record the singing context and perform multi-level label encoding based on the field survey results;

[0061] Festival labeling unit, used to extract festival background and establish cultural level labels by combining the Yi calendar and local literature;

[0062] Singer file creation unit, used to build singer information files and record their individual attributes and pedigree;

[0063] Yi language nested structural units are used to establish the phonetic and semantic structure of Yi language paragraphs in lyrics;

[0064] The node generation unit is used to structure the data into a set of semantic element nodes and generate attribute labels.

[0065] Furthermore, the semantic graph construction module includes:

[0066] A graph structure definition unit is used to establish a graph model containing lyrics, melody, festival, and singer node types using a heterogeneous graph approach;

[0067] Node attribute encoding unit, used to number each node and establish its attribute table of type, label, and cultural level;

[0068] The semantic edge generation unit is used to call the association rule mining algorithm to calculate the semantic association between nodes based on the frequency of occurrence and co-occurrence frequency, and generate a connecting edge when the association exceeds the set threshold;

[0069] An isolated node processing unit, used to perform knowledge migration or manual completion on isolated nodes that have no connection;

[0070] The graph storage unit is used to save the constructed semantic graph as a graph database or RDF triple format.

[0071] Furthermore, the node clustering and path construction module includes:

[0072] A node vector generation unit, which extracts the attributes of all nodes in the graph and constructs node vectors containing word embeddings, structural encodings, functional labels, and contextual levels;

[0073] A clustering analysis unit, used for clustering node vectors using a graph clustering algorithm;

[0074] Node classification unit, used to divide the clustering results into native nodes, educational nodes, practice nodes and evaluation nodes;

[0075] A path template setting unit is used to set a standard inheritance path structure template, which includes at least one native node, one education node, one practice node and one evaluation node;

[0076] The path combination and numbering unit is used to generate a path structure by combining nodes based on the connection relationship, and assign path numbers to store them in the path library.

[0077] Furthermore, the path matching and recommendation module includes:

[0078] The portrait modeling unit is used to collect learner ethnicity, language ability, vocal foundation, learning preferences and historical behavior data to construct a learner portrait vector;

[0079] Path attribute modeling unit, used to extract content features, language types, rhythm styles and teaching difficulty in the path to construct a path attribute vector;

[0080] A fitness calculation unit, used to calculate the matching score between the learner and each path using a weighted cosine similarity function;

[0081] A path selection unit is used to select the path with the highest fitness score as the personalized recommendation path;

[0082] The path binding unit is used to bind the learner to the selected path and write the binding relationship into the system database.

[0083] Furthermore, the behavior collection and evaluation module includes:

[0084] The behavior collection unit is used to configure the node behavior collection mechanism before the path is executed, and collect the singing audio, answer text, performance record and scoring data according to the node type;

[0085] An indicator calculation unit is used to calculate standardized indicator values for the collected behavioral data, including melody restoration, semantic understanding, and practice completion;

[0086] Evaluation function unit, used to set weighted evaluation function and calculate the node comprehensive score based on multiple indicators;

[0087] The path adjustment unit is used to feed back the evaluation results to the path structure and adjust the node recommendation weight or add supplementary nodes according to the score.

[0088] Furthermore, the path archiving and trajectory mapping module includes:

[0089] A resume structure building unit, used to create a path resume data structure including learner ID, path number, node execution order, completion status, scoring result and timestamp;

[0090] A history archiving unit for saving the route history in a structured format to a route database;

[0091] A graph generation unit is used to construct a inheritance trajectory graph based on historical data. Nodes in the graph represent path nodes, and edges represent execution order and state attributes.

[0092] The graph visualization unit is used to display the trajectory graph as an interactive graph, identify path bottlenecks and provide feedback to the path recommendation module.

[0093] By constructing an inheritance path structure including native nodes, education nodes, practice nodes and evaluation nodes, the present invention realizes the organic integration of Yi folk song teaching content with cultural semantics, practical application and learning evaluation, overcomes the problems of fragmented teaching content and lack of systematic path in the traditional inheritance model, and improves the integrity and logic of the inheritance process.

[0094] The present invention introduces learner portraits and adaptability functions, and dynamically generates personalized inheritance paths based on learners' language ability, vocal foundation and learning preferences. This can significantly improve learning efficiency and content acceptance, and enhance the pertinence and intelligence of path recommendations.

[0095] By collecting behavioral data and providing evaluation feedback during the path execution process, the present invention realizes the dynamic adjustment of the path structure and the archiving of resumes, and at the same time constructs a visual inheritance trajectory map, providing an operational quantitative basis for teaching management, system optimization and cultural research, and has good adaptability, scalability and promotion value. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0097] Figure 1 is a flow chart of the method of the present invention;

[0098] Figure 2 It is a system block diagram of the present invention. DETAILED DESCRIPTION

[0099] The invention is preferably described below in conjunction with the accompanying drawings and specific embodiments.

[0100] This embodiment solves the above problem through the following steps:

[0101] In one embodiment, reference Figure 1 The present invention provides a method for constructing a Yi folk song inheritance path, which is used to systematically establish a personalized inheritance path covering multiple dimensions such as native resources, educational content, practical scenarios, and evaluation feedback in the context of intangible cultural digitization, thereby realizing the dynamic, accurate, and continuous inheritance of Yi folk songs in university teaching and regional communities. The method specifically includes the following steps:

[0102] Step S10, obtaining multi-source cultural element data of Yi folk songs, wherein the cultural element data includes audio data, lyrics text, singing context, festival background, singer information and Yi language translation content, and parsing the cultural element data into multiple semantic element nodes.

[0103] In the process of constructing the inheritance pathways of Yi folk songs, to ensure that the generated pathways accurately reflect the multidimensional cultural characteristics of the folk songs, it is necessary to first comprehensively and systematically analyze the original cultural resources and transform them into knowledge nodes with semantic expression capabilities. Because Yi folk songs not only possess musical characteristics but also embed complex contexts such as language habits, cultural etiquette, social functions, and festival customs, it is necessary to establish a data structure that integrates audio, text, semantics, and background information to support the subsequent semantic graph construction and pathway modeling process.

[0104] The multi-source cultural element data refers to a data set with Yi folk songs as the core, originating from different media and cultural levels. The element content contained is used to construct node information in the knowledge graph and is the basic material for constructing the entire inheritance path.

[0105] In the cultural element data:

[0106] Audio data refers to digital audio files that record the actual performance of Yi folk songs, including acoustic features such as melody, rhythm, and singing intonation.

[0107] Lyrics text refers to the Chinese and Yi language lyrics corresponding to Yi folk songs, including poetic structure and language style.

[0108] The singing context refers to the social and cultural context in which folk songs are used, including the singing location and purpose (such as weddings).

[0109] Festival background refers to whether the folk song is associated with a specific festival or ceremony.

[0110] Singer information refers to the name, ethnicity, gender, age, inheritance lineage, etc. of the traditional singer.

[0111] Yi translation content refers to translating lyrics or semantic content from Yi into common language so that non-Yi speakers can understand and learn.

[0112] In a specific implementation of step S10, step S10 specifically includes the following sub-steps:

[0113] Step S101, collect the original singing data of Yi folk songs, obtain the audio files including the complete songs and the corresponding singing video materials. The collection of the audio files should include multiple voice versions. If there are multiple singing versions of the singing lineage, the source ethnic group, region and inheritor information should be marked separately.

[0114] Step S102: extract the lyrics text data and transcribe the audio data into text content through an automatic speech recognition system. If the system's automatic recognition is inaccurate, the Yi language lyrics are manually dictated and semantically annotated and translated by researchers with bilingual skills to form a lyrics text structure in a comparative format.

[0115] Step S103, recording singing context information, obtaining the usage context of folk songs through field surveys or folk literature, and classifying them into types including but not limited to weddings, funerals, farming, blessings, drinking songs, etc., and encoding them using a standardized label system.

[0116] Step S104, obtain festival background information, combine the Yi calendar, local chronicles and cultural department information to determine the national festivals, traditional activities or cultural rituals associated with the folk song, and set multi-level labels to identify its importance, spread and ritual embedding degree.

[0117] Step S105, constructing a singer information file, identifying the singers and archiving the data of the collected folk song materials, collecting basic information such as their life experience, village they belong to, and teacher-student relationship, and forming a "singer-song" mapping table.

[0118] Step S106, sorting out the Yi language translation content, marking the corresponding syllable pinyin and literal translation content for the Yi language paragraphs in the lyrics text, and recording them in a nested structure to facilitate subsequent machine processing and language teaching system calls.

[0119] Step S107: organize the collected and sorted information into a structured data format to construct a set of semantic element nodes. Each node is accompanied by a corresponding label, context identifier and semantic dimension information, which is input into subsequent steps as the basic elements of graph construction.

[0120] The implementation of the above steps not only achieves the systematic collection of Yi folk songs from a multi-dimensional cultural perspective, but also ensures the accuracy of subsequent semantic modeling and the cultural depth of path construction. By transforming raw unstructured data into structured, semantic knowledge nodes, the system greatly improves the parsability and inheritability of traditional folk songs, while also providing complete data support for building paths for personalized learning.

[0121] In an exemplary implementation, the wedding folk song "Ashiqie" widely sung during the "Torch Festival" in Liangshan area was collected. Through audio transcription and text comparison, the following node structure was formed: "Ashiqie (node 1)" is associated with "wedding scene (node 2)", "Torch Festival (node 3)", "Singer A (node 4)", "Yi sentence pattern B (node 5)", etc., forming multilateral connections in the graph, providing complete cultural semantic support for subsequent path selection and learning recommendations.

[0122] Step S20: constructing a semantic graph of Yi folk songs based on the semantic element nodes, wherein each node in the graph represents a folk song knowledge unit, and the connecting edges between the nodes represent logical dependency relationships.

[0123] To achieve structured, personalized, and dynamic recommendations for Yi folk song inheritance pathways, the semantic element nodes parsed in step S10 need to be further organized into a knowledge graph structure with associated semantics and contextual logic. Due to the nonlinear, multi-dependent semantic relationships between the various cultural elements in Yi folk songs, a single data hierarchy cannot support pathway modeling and recommendation engines. Therefore, it is necessary to construct a semantic graph, organically connecting discrete nodes with logical dependencies to achieve knowledge networking, association visualization, and semantic reasoning. This provides a topological foundation and semantic connectivity support for subsequent pathway construction.

[0124] In a specific implementation of step S20, step S20 specifically includes the following sub-steps:

[0125] Step S201, define the structural model of the semantic graph, using a heterogeneous graph structure, which contains multiple types of nodes and edges. The node types include but are not limited to lyrics nodes, melody nodes, festival nodes, scene nodes, singer nodes and language nodes. The edge types include causal relationships, homology relationships, similarity relationships and temporal relationships.

[0126] In step S202, the semantic element nodes in step S10 are mapped to a node set of the graph structure according to their attribute classification, an ID number is assigned to each node using a unique identifier, and a node attribute table is established to record the node type, language category, information source, semantic label and cultural level.

[0127] Step S203: construct the logical dependency relationship between the nodes. For each pair of nodes that may have a semantic relationship, call the association rule mining algorithm to calculate the association index. The association calculation can be performed in the following form:

[0128]

[0129] in, Representation node With node The semantic relevance of and Represents nodes respectively and nodes The frequency of occurrence of Indicates the frequency of the two occurring together.

[0130] when Exceeding the set threshold When , add the node Pointing to a node , and label the type and weight of the edge.

[0131] Step S204: Execute graph structure optimization processing and perform semantic completion on isolated nodes. If a node does not establish an effective connection with any other node, add an inference connection to it through the knowledge transfer mechanism or expert annotation method to improve the connectivity of the overall graph.

[0132] Step S205 completes the storage and call configuration of the graph structure data, saves the constructed semantic graph structure in the graph database format, supports subsequent queries, path deduction and visualization, and optional implementation solutions include: using the Neo4j graph database to achieve persistent storage of nodes and edges, or using triple knowledge expression based on the RDF format to achieve cross-platform compatibility.

[0133] Through the above steps, a knowledge graph system covering the semantic relationships throughout the entire life cycle of Yi folk songs can be constructed. This system makes the complex semantic structure between raw data nodes explicit, networked, and structured, thereby providing a foundational topological structure, contextual semantic support, and dynamic expansion capabilities for path generation. Furthermore, the graph structure is continuously updated, possessing adaptability and evolutionary capabilities, providing a semantic reasoning foundation for personalized path recommendations.

[0134] In a specific example, in the collection example of the folk song "Ashiqie", its lyrics node can be connected to the "wedding scene" node and the "Torch Festival" node through edges, the melody node is connected to the "female solo" node and the "flat tune style" node, the singer node is connected to the "inheritance genealogy branch X" node, and the language node is connected to the "Southern Yi dialect" node. Through the construction of these nodes and edges, a complete multi-dimensional semantic association subgraph can be formed in the semantic graph for subsequent path construction and evaluation.

[0135] Step S30, clustering the semantic graph to generate multiple semantic association clusters, and constructing an inheritance node cluster including native nodes, education nodes, practice nodes and evaluation nodes based on the semantic association clusters, defining each inheritance path as a structural unit including the above-mentioned types of nodes.

[0136] In the process of constructing a Yi folk song inheritance pathway, to ensure that the pathway possesses complete cultural expression, pedagogical adaptation, practical scenario mapping, and learning assessment feedback capabilities, a structured analysis of the constructed Yi folk song semantic map is required. Clustering is used to identify semantically adjacent node groups, which are then abstracted into functional nodes within the pathway and combined into inheritance pathway units with stable structure, clear goals, and actionable guidance. Because nodes within the semantic map vary in expression content, usage scenarios, and cultural functions, clustering is employed to uniformly categorize them at the semantic level, ensuring that each pathway encompasses a continuous cognitive chain from "cultural starting point" to "competence verification."

[0137] In this step:

[0138] Native nodes represent the nodes of folk song content, such as lyrics, melody segments, singing styles, etc.

[0139] Education nodes are nodes associated with teaching activities, such as course modules, teaching objectives, and teaching material locations.

[0140] Practice nodes represent nodes of folk song application scenarios such as actual performances, festival performances, and folk activities.

[0141] Evaluation nodes are used to carry the results of learning behaviors, such as semantic understanding accuracy, tone matching, and situation recurrence index.

[0142] In a specific implementation, step S30 includes the following sub-steps:

[0143] Step S301: Node vector feature construction

[0144] Extract attributes from all nodes in the semantic graph and construct a vectorized representation of each node. The vector contains:

[0145] Word embedding encoding reflects the semantic distribution of node names;

[0146] Adjacency structure encoding, indicating the connection status of the node in the graph;

[0147] Feature label vector, indicating the category it belongs to (native, educational, etc.);

[0148] Context level identifier, encoding the context level of the node in the inheritance logic.

[0149] The implementation method is to splice the above elements into a fixed-dimensional vector

[0150] in:

[0151] Number the node;

[0152] is the vector dimension, for example, set to 128 dimensions;

[0153] Word embedding uses pre-trained models such as Word2Vec or BERT;

[0154] Adjacency coding uses a combination of node degree and PageRank value;

[0155] Function labels are one-hot encoded;

[0156] The level identifier is represented by the level number.

[0157] Step S302: semantic clustering analysis

[0158] All node vectors Input into the graph clustering algorithm for clustering. Optional implementation solutions include:

[0159] K-means clustering;

[0160] Hierarchical clustering;

[0161] Louvain community partitioning algorithm;

[0162] HDBSCAN density clustering algorithm.

[0163] Take K-means as an example:

[0164] Set the number of clusters ;

[0165] initialization cluster centers;

[0166] Iterative updates:

[0167] Assignment: Each vector is assigned to the nearest cluster center;

[0168] Update: Recalculate the center of each cluster as the mean of the vector to which it belongs;

[0169] The process terminates when the cluster assignments no longer change or the maximum number of iterations is reached.

[0170] The output is semantic association clusters .

[0171] Step S303: Functional node classification

[0172] The function of the nodes in each semantic cluster is determined by combining the original label of the node and the contextual attributes of the subgraph in which it is located to determine its functional type. The classification criteria are as follows:

[0173] If the node is directly related to audio, lyrics, or rhythm, it is marked as a "native node";

[0174] If the node is linked to a course number, textbook chapter, or learning objective, it is marked as an "education node";

[0175] If the node is associated with a festival, event venue, or community performance, it is marked as a "practice node";

[0176] If a node participates in the scoring mechanism, behavioral indicators, and learning feedback, it is marked as an "evaluation node".

[0177] Each type of node set is recorded as:

[0178] Native Node Set ;

[0179] Education Node Set ;

[0180] Practice Node Set ;

[0181] Evaluation node set .

[0182] Step S304: inheritance path structure definition

[0183] Construct standard inheritance path structural units and set paths is a four-tuple:

[0184]

[0185] in:

[0186] , represents the native node;

[0187] , represents the education node;

[0188] , represents a practice node;

[0189] , represents the evaluation node.

[0190] If a path does not contain any of the above four types of nodes, it will not be included in the valid path set.

[0191] Step S305: Path combination and storage

[0192] The four types of nodes with connected relationships are extracted from the graph as specific path units, and their node numbers, connection order, and path IDs are recorded. Path information is stored in a structured path library to support subsequent path evaluation and personalized matching.

[0193] By implementing the above steps, we can uncover a culturally complete, pedagogically systematic, and practically feasible inheritance pathway structure from the complex Yi folk song knowledge graph, effectively avoiding the problems of "isolated content, disconnected teaching, and lack of feedback" in traditional folk song inheritance. This method is adaptable to different teaching groups and supports dynamic path expansion and version updates, making it highly adaptable and valuable for promotion.

[0194] In a specific example, still taking the folk song "Ashiqie" as an example, the following nodes are extracted from the semantic graph:

[0195] Lyrics melody node "Ashiqie Lyrics Section 1" is classified as a native node ;

[0196] The textbook chapter node "Vocal Music Basics Course Chapter 3" is classified as an education node ;

[0197] The practical task node "2024 Torch Festival Campus Performance" is classified as a practical node ;

[0198] The scoring node "Pitch Restoration Rate Evaluation Standard" is classified as an evaluation node .

[0199] Combine the above four types of nodes to generate a path quadruple:

[0200]

[0201] The system stores this path in the path number library, labeled "Path P-1003," and makes it a candidate for subsequent learner path matching. If the subsequent learner's profile matches the path's functionality well, this path will be recommended as a preferred inheritance solution.

[0202] Step S40: Based on the learner portrait information, an adaptability function is constructed to evaluate the matching degree between each inheritance path and the learner, and the inheritance path with the greatest matching degree is selected to generate a personalized Yi folk song inheritance path.

[0203] To achieve personalized recommendations and adaptive matching for Yi folk song transmission pathways, it is necessary to establish individual characteristic models based on learners' knowledge background, cultural awareness, language proficiency, and learning preferences. Combined with the previously constructed transmission pathway structure, the adaptability of each pathway to the learner can be assessed. Because learners vary significantly in language foundation, vocal experience, and cultural familiarity, adopting a fixed pathway can lead to poor teaching outcomes and cultural misunderstandings. Therefore, a quantifiable adaptability function is needed. By aligning the features between the learner profile and the pathway attributes, the most suitable pathway can be selected, thereby generating a precise, efficient, and dynamically adjustable personalized Yi folk song transmission pathway.

[0204] In this step:

[0205] Learner profile information refers to a multidimensional feature vector generated based on the learner's basic attributes, historical behavior, preference labels and other information, which is used to describe the learner's ability status and learning needs.

[0206] The adaptability function refers to the function form used to calculate the degree of match between the learner profile and the path structure, which is usually expressed in the form of a similarity measure or a scoring function.

[0207] The inheritance path refers to a complete structural unit consisting of original nodes, educational nodes, practice nodes and evaluation nodes, which is used to describe the implementation process of a specific folk song in teaching.

[0208] A personalized path refers to the path that best fits the profile of a specific learner, selected based on the adaptability function, and is the only output result of the "person-path" mapping relationship.

[0209] In a specific implementation, step S40 specifically includes the following sub-steps:

[0210] Step S401: Construct a learner profile feature model

[0211] Extract learner profiles from registration information, course records, and interaction behaviors to form a feature vector containing the following elements:

[0212] Ethnic identity coding, used to determine their affinity to the Yi language and culture;

[0213] Language proficiency level, based on language proficiency test scores, is divided into three levels: high, medium, and low;

[0214] Vocal basics are graded as beginner, intermediate, or advanced based on parameters such as pitch, rhythm, and range;

[0215] Learning preference tags, including preferences for strong beats, popular lyrics, and female vocals;

[0216] Historical performance indicators, including behavioral indicators such as completion, engagement, and duration.

[0217] The above features are encoded into vector form as follows:

[0218]

[0219] in, For learner portrait vector, Indicates the The eigenvalues of the dimensions, is the total number of features.

[0220] Step S402: Path attribute vectorization modeling

[0221] For each path constructed in step S30 , extract the attribute information of the nodes, such as content labels, language characteristics, rhythm style, teaching difficulty, and practice methods, and construct the path attribute vector:

[0222]

[0223] in, Indicates the The characteristic vector of the path, Indicates that the path is The attribute value of the dimension corresponds to the first Features.

[0224] Step S403: Calculation of adaptability function

[0225] Use the vector similarity function to calculate the matching degree between the path and the learner profile:

[0226]

[0227] in:

[0228] For learners and paths The fitness score of

[0229] For the The weight coefficient of each dimension;

[0230] and The learners and paths are The values in the dimensions.

[0231] Step S404: Select the optimal path

[0232] For all candidate path sets Calculate the matching score and get the path number with the highest score:

[0233]

[0234] Will The path referred to serves as the personalized Yi folk song inheritance path for current learners.

[0235] Step S405: Path binding and system recording

[0236] The path With learners Bind them one by one, record their path number, execution time and path version, and synchronize them to the learning platform database for subsequent behavior feedback and dynamic adjustment.

[0237] By implementing the above steps, we can create detailed profiles of learners with different backgrounds, abilities, and preferences, ensuring that the generated inheritance paths are highly consistent with their cognitive structures and learning objectives, thereby significantly improving learning efficiency, cultural cognition accuracy, and the stability of inheritance quality. This mechanism is not only real-time and scalable, but also provides path evaluation and feedback capabilities, allowing for path optimization as learners' behavior changes.

[0238] In a specific example:

[0239] For a learner L, the system extracts the following profile information:

[0240] Ethnic identity: Han;

[0241] Language proficiency: If you are completely unfamiliar with Yi language, the score is 1;

[0242] Vocal Fundamentals: Intermediate, score 3;

[0243] Preferred tags: strong sense of rhythm, female singing style;

[0244] Historical performance: High completion but weak cultural understanding.

[0245] Construct the learner vector:

[0246]

[0247] Extracting paths from the path candidate library (As in the aforementioned example of “Ashiqie”), its attribute vector is:

[0248]

[0249] Setting weights , then the similarity calculation result is: ,Based on the principle of maximum score, the system selects path number P-1032 as the personalized inheritance path of learner L, and pushes the corresponding courses,,singing videos and practice plans to achieve a complete closed loop from portrait recognition to,path construction.

[0250] Step S50: During the execution of the inheritance path, the learner's behavior data at each node is collected, including the melody restoration degree, semantic understanding degree and practice completion degree indicators, the evaluation function value of each node is calculated, and the function value is fed back to update the path structure.

[0251] After completing the learner path recommendation and entering the path execution phase, in order to dynamically monitor the learning process and adaptively adjust the path structure, it is necessary to collect learner behavioral performance data in real time during the execution of the path nodes. Because learning Yi folk songs involves not only melodic imitation and language memorization, but also cultural understanding and practical performance, a single-dimensional assessment is insufficient to reflect the learning status. Therefore, representative behavioral indicators should be collected during the execution of the original node, education node, practice node, and evaluation node, and a multidimensional evaluation function should be constructed. The evaluation results should be fed back to the path model in real time for subsequent path adjustment and learning plan optimization.

[0252] In this step:

[0253] Melody restoration refers to the degree to which learners restore the standard pitch, rhythm and musical pattern when singing, and is an indicator for evaluating the quality of folk song imitation.

[0254] Semantic comprehension refers to the depth of learners' understanding of the lyrics content, cultural background and semantic structure, which can be achieved through answering tests or voice retelling.

[0255] Practice completion refers to whether learners have completed practical tasks, such as performance participation, recording submission, situational demonstration, etc. The completion status is used to evaluate learning outcomes.

[0256] In a specific implementation, step S50 includes the following sub-steps:

[0257] Step S501: Initialization of behavior data collection mechanism

[0258] Before each node in the path is executed, the behavior collection module is bound. The module sets the corresponding collection plan according to the node type, including:

[0259] For native nodes, collect audio input and perform melody comparison analysis;

[0260] For educational nodes, collect question-answer data, semantic annotations, or oral summaries;

[0261] For practice nodes, record activity participation status and achievement upload records;

[0262] For evaluation nodes, integrate system scoring and instructor scoring.

[0263] Optional implementations include: using speech recognition algorithms to compare pitch deviations, using natural language processing to score text understanding, and using task management systems to quantify the status of practice completion.

[0264] Step S502: Index calculation and standardization

[0265] For each type of behavioral data, define the normalized index , indicating the The raw performance scores of these indicators are normalized to the range [0, 1] and used for subsequent comprehensive scoring. Indicators include, but are not limited to, melodic restoration, semantic understanding, and practical completion. Alternatively, pitch matching can be used for scoring.

[0266] Step S503: Evaluation function construction and calculation

[0267] Set up multi-index weighted evaluation function , of the following form:

[0268]

[0269] in:

[0270] Score for comprehensive assessment;

[0271] is the number of indicator types;

[0272] For the The weight of the indicators meets the ;

[0273] For the Standardized scores of behavioral indicators.

[0274] The system dynamically sets the path type according to different For example, the cultural understanding path can increase the weight of semantic understanding, while the performance-oriented path can increase the weight of melody restoration and practical completion.

[0275] Step S504: Path feedback mechanism execution

[0276] Bind the node evaluation function value to the path structure, and the system decides whether to fine-tune the path based on the score of each node. The adjustments include:

[0277] Lower the recommendation priority of nodes with lower scores;

[0278] Add supplementary nodes targeting weak links;

[0279] Rearrange the order of node execution;

[0280] Mark as "Path needs to be strengthened" or "Path adaptation is successful".

[0281] The updated path results are synchronously written into the database for recommendation optimization or individual evolution path construction in the subsequent learning phase.

[0282] This step establishes a behavioral feedback loop within the Yi folk song inheritance pathway. This not only accurately reflects learners' actual performance at different stages but also provides a quantitative basis for the dynamic reconstruction and personalized iteration of the pathway. The evaluation process comprehensively encompasses the three dimensions of musical skill, language comprehension, and practical completion. With the advantages of definable indicators, observable behaviors, and adjustable pathways, it significantly enhances the intelligence, adaptability, and stability of the inheritance pathway.

[0283] For example, taking the inheritance path P-1032 of the aforementioned "Ashiqie" as an example, after learner L executes the native node "Melody Imitation", the system collects his recording and analyzes the melody deviation to obtain the restoration score. ; Answer 4 out of 6 questions correctly in the education node "Lyrics Understanding Questions" to calculate the semantic understanding degree ; In the practice node "Torch Festival Performance", the recording video was not submitted, and the practice completion . Set the weight w=[0.4,0.4,0.2], then the evaluation function value is:

[0284]

[0285] The system identifies it as a "partially completed path", marks the path as "recommended for strengthening the practical link", and inserts a "simple performance task" node in the path database for strengthening practical tasks in subsequent learning stages to ensure the complete closure of the inheritance chain.

[0286] Step S60: Archive the final constructed inheritance path and learner behavior results to generate a path resume and inheritance trajectory map.

[0287] After completing the execution and behavioral assessment of the inheritance pathway, to ensure the traceability of the learning process and sustainable optimization of the pathway system, each learner's pathway structure and behavioral assessment results must be archived. Based on this data, a transmission pathway history record and a trajectory map for analysis and presentation are generated. This archiving mechanism enables process-based documentation and phased analysis of individual transmission processes. The construction of a trajectory map facilitates understanding the node distribution, bottlenecks, and dynamic evolution trends of Yi folk song transmission from a holistic perspective of the pathway space. This archiving and mapping mechanism not only provides a basis for subsequent teaching decisions and inheritance interventions, but also provides data support for knowledge map updates, curriculum system transformation, and policy planning.

[0288] In this step:

[0289] Path history refers to recording the inheritance path completed by the learner in chronological order as structured data, including node number, completion status, timestamp and evaluation score.

[0290] The inheritance trajectory map refers to a visual display of the learning trajectory, path jumps and stage status of a learner or group in the entire inheritance network in the form of a graph structure.

[0291] In a specific implementation, step S60 includes the following sub-steps:

[0292] Step S601: Initialize the archive data structure

[0293] After the path execution is completed, a structured archive object is created, defining the following fields:

[0294] Learner Identifier (UID);

[0295] Path ID (PID);

[0296] Path version (VersionID);

[0297] NodeList, which contains the number, type and execution order of each node;

[0298] Node completion status (StatusList), records the completion status of each node, with values including "completed", "uncompleted" and "partially completed";

[0299] Node evaluation value (EvalList), corresponding to the behavior evaluation score of the node in step S50;

[0300] Execution time (TimeStampList), the entry and exit time of each node;

[0301] The overall score (PathScore) is the performance score of the entire path based on the comprehensive weight.

[0302] The structure is organized and stored in JSON or relational table structure. Optional implementation solutions include: using MongoDB database for document storage, or using PostgreSQL to record in a standardized form.

[0303] Step S602: Generate path history record

[0304] Assemble the above fields into learner-path history entries and store them in the path history table to form a complete process record. For example:

[0305] {

[0306] "UID": "L-102",

[0307] "PID": "P-1032",

[0308] "NodeList": ["A1", "E3", "P2", "V5"],

[0309] "StatusList": ["Completed", "Completed", "Uncompleted", "Partially Completed"],

[0310] "EvalList": [0.92, 0.67, 0.00, 0.55],

[0311] "TimeStampList": ["2025-02-10T08:30", ..., "2025-02-10T10:15"],

[0312] "PathScore": 0.636

[0313] }

[0314] Furthermore, if there are multiple path versions, the version history can be recorded and supported for subsequent comparative analysis.

[0315] Step S603: Generate trajectory graph nodes and edges

[0316] Construct a graph node set, where the nodes are the path nodes of each execution, and the edges are the logical order and time sequence between the nodes. The graph structure is represented as:

[0317]

[0318] in:

[0319] For users Trajectory map of

[0320] is a set of path nodes;

[0321] is the time order edge between nodes;

[0322] Each edge is accompanied by execution time and completion status attributes.

[0323] If the same learner executes multiple paths, a multi-path trajectory overlay graph can be constructed.

[0324] Step S604: Integration of trajectory map visualization and system feedback

[0325] Use a visualization engine (such as D3.js, ECharts) to display the trajectory map in the form of a dynamic interactive graph, supporting the following operations:

[0326] Demonstrate the learner path execution chain;

[0327] Highlight low-scoring nodes;

[0328] Display overlapping areas of multiple user paths;

[0329] Identify high-frequency failure areas of nodes.

[0330] Write graph data back to the inheritance system backend to support subsequent path recommendation optimization and graph evolution algorithm application.

[0331] See also Figure 2 In another embodiment, the present invention further provides a system for constructing a Yi folk song inheritance path, comprising:

[0332] A multi-source cultural element acquisition module is used to obtain multi-source cultural element data of Yi folk songs, including audio data, lyrics, singing context, festival background, singer information and Yi language translation content, and parse the cultural elements into multiple semantic element nodes;

[0333] A semantic graph construction module is used to construct a semantic graph of Yi folk songs based on the semantic element nodes, wherein each node in the semantic graph represents a folk song knowledge unit, and the connecting edges between the nodes represent logical dependency relationships;

[0334] A node clustering and path construction module is used to cluster the semantic graph to generate multiple semantic association clusters, and to construct inheritance node clusters including native nodes, education nodes, practice nodes, and evaluation nodes based on the semantic association clusters, defining each inheritance path as a structural unit including the above-mentioned nodes;

[0335] The path matching and recommendation module is used to construct an adaptability function based on the learner profile information to evaluate the matching degree between each inheritance path and the learner, select the inheritance path with the greatest matching degree, and generate a personalized inheritance path for Yi folk songs;

[0336] The behavior collection and evaluation module is used to collect learners' behavior data at each node during the execution of the inheritance path, including the melody restoration degree, semantic understanding degree and practice completion degree indicators, calculate the evaluation function value of each node, and feed the function value back to the path structure;

[0337] The path archiving and trajectory mapping module is used to archive the final constructed inheritance path and learner behavior results, and generate a path resume and inheritance trajectory map.

[0338] In a further embodiment, the multi-source cultural element acquisition module includes:

[0339] Audio collection unit, used to collect audio and video materials of Yi folk songs with multiple voice parts, and annotate the singer, region and genealogy information;

[0340] The speech transcription and translation unit is used to transcribe the audio into lyrics using a speech recognition system, and bilingual personnel perform Yi language annotation and translation;

[0341] The context annotation unit is used to record the singing context and perform multi-level label encoding based on the field survey results;

[0342] Festival labeling unit, used to extract festival background and establish cultural level labels by combining the Yi calendar and local literature;

[0343] Singer file creation unit, used to build singer information files and record their individual attributes and pedigree;

[0344] Yi language nested structural units are used to establish the phonetic and semantic structure of Yi language paragraphs in lyrics;

[0345] The node generation unit is used to structure the data into a set of semantic element nodes and generate attribute labels.

[0346] In a further embodiment, the semantic graph building module includes:

[0347] A graph structure definition unit is used to establish a graph model containing lyrics, melody, festival, and singer node types using a heterogeneous graph approach;

[0348] Node attribute encoding unit, used to number each node and establish its attribute table of type, label, and cultural level;

[0349] The semantic edge generation unit is used to call the association rule mining algorithm to calculate the semantic association between nodes based on the frequency of occurrence and co-occurrence frequency, and generate a connecting edge when the association exceeds the set threshold;

[0350] An isolated node processing unit, used to perform knowledge migration or manual completion on isolated nodes that have no connection;

[0351] The graph storage unit is used to save the constructed semantic graph as a graph database or RDF triple format.

[0352] In a further embodiment, the node clustering and path building module includes:

[0353] A node vector generation unit, which extracts the attributes of all nodes in the graph and constructs node vectors containing word embeddings, structural encodings, functional labels, and contextual levels;

[0354] A clustering analysis unit, used for clustering node vectors using a graph clustering algorithm;

[0355] Node classification unit, used to divide the clustering results into native nodes, educational nodes, practice nodes and evaluation nodes;

[0356] A path template setting unit is used to set a standard inheritance path structure template, which includes at least one native node, one education node, one practice node and one evaluation node;

[0357] The path combination and numbering unit is used to generate a path structure by combining nodes based on the connection relationship, and assign path numbers to store them in the path library.

[0358] In a further embodiment, the path matching and recommendation module includes:

[0359] The portrait modeling unit is used to collect learner ethnicity, language ability, vocal foundation, learning preferences and historical behavior data to construct a learner portrait vector;

[0360] Path attribute modeling unit, used to extract content features, language types, rhythm styles and teaching difficulty in the path to construct a path attribute vector;

[0361] A fitness calculation unit, used to calculate the matching score between the learner and each path using a weighted cosine similarity function;

[0362] A path selection unit is used to select the path with the highest fitness score as the personalized recommendation path;

[0363] The path binding unit is used to bind the learner to the selected path and write the binding relationship into the system database.

[0364] In a further embodiment, the behavior collection and evaluation module includes:

[0365] The behavior collection unit is used to configure the node behavior collection mechanism before the path is executed, and collect the singing audio, answer text, performance record and scoring data according to the node type;

[0366] An indicator calculation unit is used to calculate standardized indicator values for the collected behavioral data, including melody restoration, semantic understanding, and practice completion;

[0367] Evaluation function unit, used to set weighted evaluation function and calculate the node comprehensive score based on multiple indicators;

[0368] The path adjustment unit is used to feed back the evaluation results to the path structure and adjust the node recommendation weight or add supplementary nodes according to the score.

[0369] In a further embodiment, the path archiving and trajectory mapping module includes:

[0370] A resume structure building unit, used to create a path resume data structure including learner ID, path number, node execution order, completion status, scoring result and timestamp;

[0371] A history archiving unit for saving the route history in a structured format to a route database;

[0372] A graph generation unit is used to construct a inheritance trajectory graph based on historical data. Nodes in the graph represent path nodes, and edges represent execution order and state attributes.

[0373] The graph visualization unit is used to display the trajectory graph as an interactive graph, identify path bottlenecks and provide feedback to the path recommendation module.

[0374] It should be noted that the explanation of the aforementioned embodiment of the method for constructing the Yi folk song inheritance path is also applicable to the device of the embodiment of the present application and will not be repeated here.

[0375] Those skilled in the art will appreciate that the various units and algorithm steps described in the embodiments disclosed herein can be implemented using a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0376] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0377] In the several embodiments provided in this application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of this application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory; hereinafter referred to as: ROM), random access memory (Random Access Memory; hereinafter referred to as: RAM), magnetic disk or optical disk, and other media that can store program code.

[0378] The above is only a specific embodiment of the present application. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in this application, which should be included in the scope of protection of this application. For some module structures that are not particularly clear in the present invention, the content recorded in the prior art shall prevail. The prior art mentioned in the above background technology section and the specific embodiment section of the present invention can be regarded as part of the present invention and is used to understand the meaning of some technical features or parameters.

Claims

1. A method for constructing a Yi folk song inheritance path, characterized in that: The method comprises the following steps: Step S10, obtaining multi-source cultural element data of Yi folk songs, wherein the cultural element data includes audio data, lyrics text, singing context, festival background, singer information, and Yi language translation content, and parsing the cultural element data into multiple semantic element nodes; Step S20, constructing a semantic graph of Yi folk songs based on the semantic element nodes, wherein each node in the graph represents a folk song knowledge unit, and the connecting edges between the nodes represent logical dependency relationships; Step S30: clustering the semantic graph to generate multiple semantic association clusters, and constructing inheritance node clusters including native nodes, education nodes, practice nodes, and evaluation nodes based on the semantic association clusters, defining each inheritance path as a structural unit including the above-mentioned nodes; Step S40: Based on the learner profile information, construct an adaptability function to evaluate the matching degree between each inheritance path and the learner, select the inheritance path with the greatest matching degree, and generate a personalized Yi folk song inheritance path; Step S50: During the execution of the inheritance path, the learner's behavior data at each node is collected, including the melody restoration degree, semantic understanding degree, and practice completion degree indicators, and the evaluation function value of each node is calculated. The function value is then fed back to update the path structure; Step S60: Archive the final constructed inheritance path and learner behavior results to generate a path resume and inheritance trajectory map.

2. The method for constructing the inheritance path of Yi folk songs according to claim 1, characterized in that: The step S10 includes: Step S101, collecting original singing data of Yi folk songs, including audio data and video materials of multiple voice parts, and marking the corresponding singers, regions and genealogy information; Step S102: using a speech recognition system to transcribe the audio data into lyrics text, and bilingual personnel perform Yi language semantic annotation and translation; Step S103, recording the singing context of folk songs based on field surveys, and classifying and coding them using multi-level labels; Step S104: extracting festival backgrounds by combining the Yi calendar and local chronicles, and setting festival type and cultural embedding level labels for each folk song; Step S105, constructing a performer information file, collecting the performer's name, gender, age, and teacher lineage; Step S106, establishing a nested structure of Yi language paragraphs and syllable pinyins and corresponding Chinese meanings in the lyrics; Step S107: Structure the collected data into a set of semantic element nodes, and record context attributes and label information.

3. The method for constructing a Yi folk song inheritance path according to claim 1, wherein: The step S20 includes: Step S201: Define a semantic graph model using a heterogeneous graph structure, including multiple types of nodes such as lyrics nodes, melody nodes, festival nodes, and singer nodes; Step S202: number the semantic element nodes and create a node attribute table to record their types, semantic labels, and cultural level information; Step S203: calling the association rule mining algorithm to calculate the semantic association between nodes based on the node occurrence frequency and co-occurrence frequency, and constructing a connecting edge if the association is greater than a set threshold; Step S204: performing knowledge migration and manual completion operations on isolated nodes; Step S205: Save the semantic graph using a graph database or RDF triple format.

4. The method for constructing a Yi folk song inheritance path according to claim 1, wherein: The step S30 includes: Step S301: extract the attributes of all nodes in the semantic graph and construct a node vector containing word embedding, structure encoding, function label and context level; Step S302, clustering the node vectors using a graph clustering algorithm; Step S303, dividing the nodes in the clustering result into native nodes, education nodes, practice nodes and evaluation nodes; Step S304: setting a standard inheritance path structure unit, including at least one original node, one education node, one practice node, and one evaluation node; Step S305 : Combining the path structure based on the node connection relationship, generating a path number and storing it in the path library.

5. The method for constructing a Yi folk song inheritance path according to claim 1, characterized in that: The step S40 includes: Step S401: Collect the learner's ethnic identity, language ability, vocal foundation, preference tags, and historical learning behavior to construct a multi-dimensional learner portrait vector; Step S402: extracting content attributes, language type, rhythm characteristics, and teaching difficulty information from the path structure to construct a path attribute vector; Step S403: Calculate the compatibility score between the learner profile and each path using a weighted cosine similarity function; Step S404: Select the path with the highest adaptability score as the personalized inheritance path for the current learner; Step S405: store the binding relationship between the learner and the path, and synchronize it to the system database.

6. The method for constructing a Yi folk song inheritance path according to claim 1, characterized in that: The step S50 includes: Step S501: Before the path is executed, a node behavior collection module is configured to collect melody imitation, question understanding, practice record, and score feedback behavior data according to node type; Step S502: Calculate standardized index values for the collected data, including melody restoration, semantic understanding, and practical completion; Step S503: Setting a weighted evaluation function to integrate multiple indicators to calculate the node comprehensive score; Step S504: Bind the evaluation score to the path structure, and adjust the node recommendation weight or insert a supplementary node according to the score.

7. The method for constructing a Yi folk song inheritance path according to claim 1, characterized in that: The step S60 includes: Step S601, creating a path history data structure to record the learner ID, path number, node execution order, completion status, score value and timestamp; Step S602, archiving the path history in a structured format into a path database; Step S603: Constructing a inheritance trajectory graph, where nodes are inheritance nodes in the history, edges are execution order relationships between nodes, and execution time and state attributes are added to the edges; Step S604 : Visualize the learning path based on the trajectory graph, identify learning bottlenecks, and feed the trajectory data back to the path recommendation module.

8. A system for constructing a path for inheriting Yi folk songs, characterized in that: The system comprises: A multi-source cultural element acquisition module is used to obtain multi-source cultural element data of Yi folk songs, including audio data, lyrics, singing context, festival background, singer information and Yi language translation content, and parse the cultural elements into multiple semantic element nodes; A semantic graph construction module is used to construct a semantic graph of Yi folk songs based on the semantic element nodes, wherein each node in the semantic graph represents a folk song knowledge unit, and the connecting edges between the nodes represent logical dependency relationships; A node clustering and path construction module is used to cluster the semantic graph to generate multiple semantic association clusters, and to construct inheritance node clusters including native nodes, education nodes, practice nodes, and evaluation nodes based on the semantic association clusters, defining each inheritance path as a structural unit including the above-mentioned nodes; The path matching and recommendation module is used to construct an adaptability function based on the learner profile information to evaluate the matching degree between each inheritance path and the learner, select the inheritance path with the greatest matching degree, and generate a personalized inheritance path for Yi folk songs; The behavior collection and evaluation module is used to collect learners' behavior data at each node during the execution of the inheritance path, including the melody restoration degree, semantic understanding degree and practice completion degree indicators, calculate the evaluation function value of each node, and feed the function value back to the path structure; The path archiving and trajectory mapping module is used to archive the final constructed inheritance path and learner behavior results, and generate a path resume and inheritance trajectory map.

9. The Yi folk song inheritance path construction system according to claim 8, characterized in that: The multi-source cultural elements acquisition module includes: Audio collection unit, used to collect audio and video materials of Yi folk songs with multiple voice parts, and annotate the singer, region and genealogy information; The speech transcription and translation unit is used to transcribe the audio into lyrics using a speech recognition system, and bilingual personnel perform Yi language annotation and translation; The context annotation unit is used to record the singing context and perform multi-level label encoding based on the field survey results; Festival labeling unit, used to extract festival background and establish cultural level labels by combining the Yi calendar and local literature; Singer file creation unit, used to build singer information files and record their individual attributes and pedigree; Yi language nested structural units are used to establish the phonetic and semantic structure of Yi language paragraphs in lyrics; The node generation unit is used to structure the data into a set of semantic element nodes and generate attribute labels.

10. The Yi folk song inheritance path construction system according to claim 8, characterized in that: The semantic graph construction module includes: A graph structure definition unit is used to establish a graph model containing lyrics, melody, festival, and singer node types using a heterogeneous graph approach; Node attribute encoding unit, used to number each node and establish its attribute table of type, label, and cultural level; The semantic edge generation unit is used to call the association rule mining algorithm to calculate the semantic association between nodes based on the frequency of occurrence and co-occurrence frequency, and generate a connecting edge when the association exceeds the set threshold; An isolated node processing unit, used to perform knowledge migration or manual completion on isolated nodes that have no connection; The graph storage unit is used to save the constructed semantic graph as a graph database or RDF triple format.

Images