Intelligent matching method and system for heterogeneous resource interaction information and public object model

Abstract

The invention discloses an intelligent matching method and system for heterogeneous resource interaction information and a public object model, and the method comprises the following steps: S1, carrying out the heterogeneous resource analysis of the GBB-based heterogeneous resource interaction information through structural body analysis, path generation and path information storage, converting type constraint and hierarchical association implied in the information into a structured symbol sequence, and outputting standardized attribute path information; s2, on the basis of standardized path information obtained by heterogeneous resource analysis, through coprocessing of abbreviation word semantic recognition and extension, semantic quantification and type verification, automatic matching is achieved under complex naming rules and structural differences, and an optimal mapping set meeting semantic association and technical feasibility is output; according to the method, a full-process processing mechanism of analysis-extension-matching-verification is constructed, and semantic analysis and structure optimization technologies are combined, so that automatic and accurate matching of heterogeneous data is realized, and the efficiency and precision bottlenecks of traditional manual configuration are solved.

Description

Technical Field

[0001] This invention relates to the field of heterogeneous resource interaction and public object model adaptation technology, and more specifically to an intelligent matching method and system for heterogeneous resource interaction information and public object model. Background Technology

[0002] In real-world engineering scenarios of heterogeneous system integration, different systems are often built independently based on different technical standards, resulting in significant heterogeneity in their data interaction. Differences in communication protocols, data formats, and naming rules create numerous obstacles to cross-system information sharing, becoming a key bottleneck restricting the efficiency of system collaboration.

[0003] In the traditional model, the mapping relationship between heterogeneous resource interaction information and the common object model relies entirely on manual configuration. Relevant technical personnel need to analyze the meaning of fields from different systems one by one, manually write matching rules, and integrate data from various sources into a unified data framework. This approach has obvious limitations: on the one hand, the data structures of many systems have complex hierarchical characteristics, and matching errors are easily caused by missing levels during manual sorting; on the other hand, there are a large number of custom abbreviations and mixed naming forms in different systems. These naming methods have strong semantic ambiguity, which can easily lead to deviations when interpreted manually, thus affecting the accuracy of data interaction.

[0004] As system complexity continues to increase, a single integration scenario may involve dozens of heterogeneous subsystems, generating a massive number of interactive data fields daily. The efficiency of traditional manual configuration methods can no longer meet actual needs. At the same time, the lag and instability of manual configuration make it difficult to support the automatic conversion of data by integration components with autonomous characteristics (such as autonomous gateways), thus restricting the level of automation and intelligence in heterogeneous system integration.

[0005] For example, in distributed system integration scenarios based on General Blackboard Technology (GBB), different subsystems share information through GBB data blackboards. However, the resources accessing GBB (such as sensors with different protocols, multi-format databases, and service nodes with different architectures) exhibit significant heterogeneity, and their data formats and naming rules have incompatibility differences with the common object model adapted to GBB. Traditional manual configuration methods cannot meet the dynamic interaction requirements of GBB and struggle to handle GBB's unique hierarchical data structure and abbreviation naming system, resulting in low matching accuracy and delayed response, thus hindering the collaborative efficiency of the GBB system.

[0006] Therefore, how to build an automated intelligent matching mechanism to solve the matching problem caused by structural differences and achieve accurate association of heterogeneous resource interaction information, so as to provide reliable technical support for the efficient integration of heterogeneous systems, is a problem that needs to be solved by those skilled in the art. Summary of the Invention

[0007] In view of this, the present invention provides an intelligent matching method and system for heterogeneous resource interaction information and common object models, which is applicable to solving the information interaction barriers between multi-source heterogeneous resources (such as sensors with different protocols, multi-format databases, and service nodes with different architectures) and standardized models in distributed systems. It can be widely applied to scenarios such as industrial internet, cross-domain data fusion, and intelligent collaborative systems to achieve efficient and accurate matching between the two, thereby improving system resource utilization and collaborative response speed.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] A method for intelligently matching heterogeneous resource interaction information with a common object model includes the following steps:

[0010] S1. For heterogeneous resource interaction information based on GBB, heterogeneous resource parsing is performed through structure parsing, path generation and path information storage. The implicit type constraints and hierarchical associations in the information are transformed into structured symbol sequences, and standardized attribute path information is output.

[0011] S2. Based on the standardized path information obtained from heterogeneous resource parsing, through the collaborative processing of abbreviation semantic recognition and expansion, semantic quantization and type verification, automatic matching is achieved under complex naming rules and structural differences, and the optimal mapping set that satisfies semantic association and technical feasibility is output.

[0012] Preferably, the heterogeneous resource interaction information is in the form of unstructured text, which is sequentially associated with the information unit name and the field description contained in specific symbols. Each field description includes a description representing the information type and a name representing the information identifier.

[0013] The output attribute path information is an explicit mapping result of the logical structure of the input information. Each path starts with an information type description, sequentially connects the complete hierarchical relationship of the fields through a specific separator, and ends with a canonical symbol.

[0014] Preferably, step S1 includes the following:

[0015] S11. Parse the structure definitions in the GBB interactive file, use regular expressions to match structure keywords, structure names and field information, and construct a structure dictionary and a top-level structure set;

[0016] S12. For each top-level struct, use a recursive method to generate the complete paths of all its fields, and prepend the data type to the paths;

[0017] S13. Save the generated path information to a file in the specified directory. If the saving attempt fails, output an error message.

[0018] Preferably, in step S11, the content of constructing the structure dictionary and the top-level structure set is as follows:

[0019] Collect all structure names and parse the field information of each structure. When processing field rows, support cases where the type contains spaces. At the same time, identify the referenced structures and find the top-level structure, that is, the structure defined in the file but not referenced by other structures.

[0020] In step S12, the visited collection is used to record the visited structures. When processing each field, if the field is a nested structure, the nested structure is processed recursively; otherwise, the data type is added before the path.

[0021] Preferably, the specific content of step S2 is as follows:

[0022] S21. Split the standardized path data obtained from heterogeneous resource parsing and extract abbreviations;

[0023] S22. Through a three-level processing mechanism of explicit expansion, intelligent inference expansion, and contextual ambiguity resolution, the fuzzy abbreviation is converted into the accurate full name, resulting in the standard abbreviation path;

[0024] S23. For standard abbreviation paths, construct a multi-granularity evaluation system from the vocabulary, component level, and path level to perform semantic similarity matching and realize the quantification of deep semantic associations;

[0025] S24. Perform type compatibility filtering and output the optimal mapping set.

[0026] Preferably, the specific content of step S22 is as follows:

[0027] S221. Explicit Expansion: Based on an abbreviation-full name mapping table built by domain experts, abbreviations with a high degree of standardization are directly replaced, covering industry-wide common abbreviations and protocol-specific abbreviations. The mapping table is dynamically updated through incremental learning.

[0028] S222. Intelligent Inference Expansion: For non-standardized abbreviations not included in the mapping table, a dual-path inference model of semantic vectors and character features is constructed. In the semantic dimension, candidate full names are screened by calculating the cosine similarity between the high-dimensional semantic vectors of the abbreviation to be expanded and the vocabulary of the domain corpus. In the character dimension, character-level feature scores are constructed by combining word root matching, edit distance and suffix analysis. The results of the two channels are weighted and fused to generate a list of candidate full names with probability values.

[0029] S223. Contextual Ambiguity Resolution: For polysemous abbreviations, a secondary screening is performed based on the hierarchical semantics of the path in which they are located to determine the optimal full name.

[0030] Preferably, the specific content of step S23 is as follows:

[0031] At the lexical level, the relevance of expanded semantic units is calculated using the cosine similarity formula of word vectors;

[0032] At the component level, for the semantic unit sequence after splitting the composite field name, the optimal alignment method is found by dynamic programming algorithm. The weighted average of the maximum similarity of the aligned components is taken as the overall score and normalized by dividing by the length of the longer sequence.

[0033] At the path level, the differences in hierarchical structure are handled based on the idea of ​​dynamic programming. A dynamic programming matrix is ​​constructed for the two path sequences, and the maximum sum of hierarchical similarity between the two path sequences is calculated.

[0034] Preferably, the specific content of step S24 is as follows:

[0035] The compatibility of basic data types is judged based on the rule that low precision is compatible with high precision. For custom data types, the compatibility of custom data types is determined by comparing whether the number of fields, field types, and field order of the structure are completely consistent, or by checking whether there is an inheritance relationship. Finally, the mapping pairs that simultaneously satisfy the semantic similarity greater than a preset threshold and are type compatible are retained to form the optimal mapping set.

[0036] Preferably, the intelligent matching method for heterogeneous resource interaction information and public object models further includes: S3. By constructing a heterogeneous data set covering multiple fields, validating its effectiveness, accuracy, and efficiency, and generating a matching report. The matching report includes the matching results, matching similarity and attribute compatibility descriptions, and whether the file extension path is correct. Users can modify the extension dictionary themselves based on whether the extension results are correct. When the file path does not match or the attribute type is incompatible, the report outputs the unmatched paths and classifies them according to the reason for the unmatch.

[0037] An intelligent matching system for heterogeneous resource interaction information and a common object model, based on the aforementioned intelligent matching method for heterogeneous resource interaction information and a common object model, includes: a heterogeneous resource parsing module, an intelligent matching optimization module, and a system report generation module;

[0038] The heterogeneous resource parsing module includes a structure parsing module and a path generation and file saving module. It is used to parse heterogeneous resource interaction information through structure parsing, path generation and path information saving. It transforms the implicit type constraints and hierarchical associations in the information into a structured symbol sequence and outputs standardized attribute path information.

[0039] The intelligent matching optimization module is used to automatically match based on the standardized path information obtained from the parsing of heterogeneous resources, through the collaborative processing of abbreviation semantic recognition and expansion, semantic quantization and type verification, under complex naming rules and structural differences, and output the optimal mapping set that satisfies semantic association and technical feasibility.

[0040] The system report generation module is used to generate matching reports for compatibility analysis with the GBB common object model and correctness explanations of GBB-specific abbreviation extensions.

[0041] As can be seen from the above technical solutions, compared with the prior art, the present invention discloses an intelligent matching method and system for heterogeneous resource interaction information and public object models. By constructing a full-process processing mechanism of parsing-expansion-matching-verification, and combining semantic analysis and structural optimization technology, it realizes automated and accurate matching of heterogeneous data, and solves the efficiency and accuracy bottlenecks of traditional manual configuration. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0043] Figure 1 A schematic diagram illustrating an intelligent matching method for heterogeneous resource interaction information and a public object model provided by the present invention;

[0044] Figure 2 This is a schematic diagram of the heterogeneous resource parsing process provided by the present invention;

[0045] Figure 3 This is a schematic diagram of the intelligent matching optimization process provided by the present invention;

[0046] Figure 4 The abbreviation and full name mapping provided for this invention is intended to represent the intent.

[0047] Figure 5 This is a schematic diagram of the word-level semantic similarity probability calculation provided by the present invention;

[0048] Figure 6A schematic diagram of a report segment for structure path matching provided by the present invention;

[0049] Figure 7 This is a schematic diagram illustrating the results of unmatched paths provided by the present invention. Detailed Implementation

[0050] 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.

[0051] This invention discloses an intelligent matching method for heterogeneous resource interaction information and a common object model, such as... Figure 1 This includes the following steps:

[0052] S1. For heterogeneous resource interaction information based on GBB, heterogeneous resource parsing is performed through structure parsing, path generation and path information storage. The implicit type constraints and hierarchical associations in the information are transformed into structured symbol sequences, and standardized attribute path information is output.

[0053] S2. Based on the standardized path information obtained from heterogeneous resource parsing, through the collaborative processing of abbreviation semantic recognition and expansion, semantic quantization and type verification, automatic matching is achieved under complex naming rules and structural differences, and the optimal mapping set that satisfies semantic association and technical feasibility is output.

[0054] In this embodiment, the common object model refers to a standardized data model adapted to the GBB interaction protocol. It is constructed based on the hierarchical data interaction specifications of GBB and includes three core modules: data type definition, field semantic encoding, and hierarchical association rules. It is used to realize unified data access and interaction of multi-source heterogeneous resources under the GBB framework.

[0055] To further implement the above technical solution, the heterogeneous resource interaction information is in the form of unstructured text, which is sequentially associated with the information unit name and the field description contained in specific symbols. Each field description contains a description representing the information type and a name representing the information identifier. Essentially, it is a logical encoding of the information unit and its internal structure. The sources of GBB-based heterogeneous resource interaction information include GBB data blackboard, GBB node interaction interface, etc.

[0056] The output attribute path information is an explicit mapping result of the logical structure of the input information. Its generation follows the dual encoding principle of "type-level": each path starts with an information type description, sequentially connects the complete hierarchical relationship of the fields through a specific separator, and ends with a standard symbol. The core function of this representation is to transform the implicit type constraints and hierarchical associations in the input information into a structured symbol sequence that can be directly processed.

[0057] In this embodiment, the structured coding features of the input information and the type and hierarchical mapping features of the output path together determine that the processing process must meet two requirements: on the one hand, it is necessary to extract and restore the internal logic of information units from unstructured text through accurate text parsing; on the other hand, it is necessary to transform the extracted logical structure into a unified symbolic expression through standardized path generation. These two steps together constitute the basis for the transformation from the original representation of information to a computable representation.

[0058] To further implement the above technical solutions, such as Figure 2 The specific content of step S1 includes:

[0059] S11. Parse the structure definitions in the GBB interactive file, use regular expressions to match structure keywords, structure names and field information, and construct a structure dictionary and a top-level structure set;

[0060] S12. For each top-level struct, use a recursive method to generate the complete paths of all its fields, and prepend the data type to the paths;

[0061] S13. Save the generated path information to a file in the specified directory. If the saving attempt fails, output an error message.

[0062] To further implement the above technical solution, step S11 involves constructing the structure dictionary and the top-level structure set as follows:

[0063] Collect all structure names and parse the field information of each structure. When processing field rows, support cases where the type contains spaces. At the same time, identify the referenced structures and find the top-level structure, that is, the structure defined in the file but not referenced by other structures.

[0064] In step S12, the visited collection is used to record the visited structures. When processing each field, if the field is a nested structure, the nested structure is processed recursively; otherwise, the data type is added before the path.

[0065] In this embodiment, the intelligent matching optimization algorithm in step S2 breaks through the reliance on experience in traditional manual matching. Through multi-level semantic computation and dynamic programming technology, it effectively solves the matching problem of heterogeneous forms but homogeneous semantics.

[0066] To further implement the above technical solutions, such as Figure 3 The specific content of step S2 is as follows:

[0067] S21. Split the standardized path data obtained from heterogeneous resource parsing and extract abbreviations;

[0068] S22. Through a three-level processing mechanism of explicit expansion, intelligent inference expansion, and contextual ambiguity resolution, the fuzzy abbreviation is converted into the accurate full name, resulting in the standard abbreviation path;

[0069] S23. For standard abbreviation paths, construct a multi-granularity evaluation system from the vocabulary, component level, and path level to perform semantic similarity matching and realize the quantification of deep semantic associations;

[0070] S24. Perform type compatibility filtering and output the optimal mapping set.

[0071] To further implement the above technical solution, the specific content of step S22 is as follows:

[0072] S221. Explicit Expansion: Based on an abbreviation-to-full name mapping table built by domain experts, highly standardized abbreviations are directly replaced, covering industry-standard abbreviations and protocol-specific abbreviations. The mapping table is dynamically updated through incremental learning to incorporate newly emerging standardized abbreviations, ensuring that the expansion accuracy of known abbreviations reaches 100%. Custom abbreviation mapping representations include, for example... Figure 4 ;

[0073] S222. Intelligent Inference Expansion: For non-standardized abbreviations not included in the mapping table, a dual-path inference model of semantic vectors and character features is constructed. In the semantic dimension, candidate full names are screened by calculating the cosine similarity between the high-dimensional semantic vectors of the abbreviation to be expanded and the vocabulary of the domain corpus. In the character dimension, character-level feature scores are constructed by combining word root matching, edit distance and suffix analysis. The results of the two channels are weighted and fused to generate a list of candidate full names with probability values.

[0074] In this embodiment, in the semantic dimension, the Word2Vec model is used to transform the abbreviation to be expanded and the words in the domain corpus into 300-dimensional high-dimensional semantic vectors, and the candidate full names are screened by calculating the cosine similarity of the vectors; in the character dimension, character-level feature scoring is constructed by combining root matching, edit distance and suffix analysis to make up for the defect of semantic vectors being too sensitive to spelling irregularities.

[0075] S223. Contextual Ambiguity Resolution: For polysemous abbreviations, a secondary screening is performed by combining the hierarchical semantics of the path in which the abbreviation is located. Information such as the parent level and sibling fields of the field in which the abbreviation is located is transformed into auxiliary features, and the optimal full name is determined accordingly.

[0076] In this embodiment, Word2Vec is a distributed word representation model based on the distributed assumption that "the semantics of a word is determined by its context." This model is pre-trained on a large general corpus and learns vector representations of words through two main architectures: Skip-gram and ContinuousBag-of-Words.

[0077] Skip-gram works by predicting the surrounding vocabulary of a target word. This structure excels at learning complex patterns, especially low-frequency words.

[0078] The content of ContinuousBag-of-Words is: given the context vocabulary of the target word, predict the target word itself. This structure has higher training efficiency.

[0079] After training, the Word2Vec model generates a dense real vector of fixed dimension for each word in the vocabulary. These word vectors have important characteristics: semantically similar words (or words that usually appear in similar contexts) are close in direction in the vector space, and the cosine value of their angle is close to 1; conversely, semantically unrelated words have large differences in vector direction, and the cosine similarity is close to 0 or negative; linear operations between vectors (such as V(King)-V(Man)+V(Woman)≈V(Queen)) can contain rich semantic and analogical relationships.

[0080] To further implement the above technical solution, the specific content of step S23 is as follows:

[0081] At the lexical level, the relevance of expanded semantic units is calculated using the cosine similarity formula of word vectors;

[0082] The formula is:

[0083]

[0084] Where w1·w2 is the vector dot product, |w1| is the vector magnitude, and the closer the cosine value is to 1, the more similar the semantics of the words are. Figure 5 ;

[0085] At the component level, for the semantic unit sequence after splitting the composite field name, the optimal alignment method is found through dynamic programming algorithm. The weighted average of the maximum similarity of the aligned components is taken as the overall score, and normalization is performed by dividing by the length of the longer sequence to avoid deviation caused by the difference in the number of components.

[0086] At the path level, the differences in hierarchical structure are handled based on the idea of ​​dynamic programming. A dynamic programming matrix is ​​constructed for the two path sequences, and the maximum sum of hierarchical similarity between the two path sequences is calculated.

[0087] In this embodiment, for a path sequence of length m and path sequence of length n The constructed dynamic programming matrix dp is given, where dp[i][j] represents the maximum sum of similarities between the first i layers of P1 and the first j layers of P2;

[0088] The initial state is dp[0][0] = 0, dp[0][j] = 0, dp[i][0] = 0 (the matching degree of the empty path is 0);

[0089] The value of dp[i][j] is derived from the results of the three subproblems, and the maximum value is taken as the optimal solution:

[0090] Match the current level: in, Component-level similarity;

[0091] Skip the i-th level of P1: dp[i][j] = dp[i-1][j];

[0092] Skip the j-th level of P2: dp[i][j] = dp[i][j-1];

[0093] After the matrix is ​​filled, dp[m][n] is the sum of similarity. By backtracking from dp[m][n] to dp[0][0], the direction of each step of the transition (match / skip) is recorded, the path alignment method is clarified, and an intuitive basis is provided for manual verification.

[0094] To further implement the above technical solution, the specific content of step S24 is as follows:

[0095] The compatibility of basic data types is judged based on the rule that low precision is compatible with high precision. For custom data types, the compatibility of custom data types is determined by comparing whether the number of fields, field types and field order of the structure are completely consistent, or by checking whether there is an inheritance relationship. Finally, the mapping pairs that simultaneously meet the requirements of semantic similarity greater than the preset threshold (default 0.8, which can be customized by the user based on the GBB scenario) and type compatibility are retained to form the optimal mapping set.

[0096] In this embodiment, the basic data types of the public object model include GBB-compatible int32, float64, etc. Custom types must follow the GBB structure inheritance specification, that is, the subtype must be fully compatible with the field structure of the parent type.

[0097] To further implement the above technical solution, an intelligent matching method for heterogeneous resource interaction information and public object models further includes: S3. By constructing a heterogeneous data set covering multiple fields, validating its effectiveness, accuracy, and efficiency, and generating a matching report. The matching report includes the matching results, matching similarity and attribute compatibility descriptions, and whether the file extension path is correct. Users can modify the extension dictionary themselves based on whether the extension results are correct. When the file path does not match or the attribute type is incompatible, the report will output the unmatched path and classify it according to the reason for the unmatch.

[0098] In this embodiment, the effectiveness of the algorithm is verified based on a self-built dataset, which covers 1740 heterogeneous interaction data based on GBB, including 418 sets of compound names containing abbreviations (78 known abbreviations and 340 unknown abbreviations). The similarity threshold used in the experiment is 0.8, the Word2Vec model dimension is 300, and the training corpus is a domain-general corpus. The comparison methods include traditional manual configuration, mapping table-based matching, and simple character matching.

[0099] The matching report output by the system integrates multi-dimensional information, including both technical details and engineering readability, such as... Figure 6 This report snippet, generated based on a real dataset, shows the structure path matching results, similarity scores, attribute compatibility information, and whether the file extension path is correct. Users can modify the extension dictionary based on the accuracy of the extension results. The similarity threshold can be customized by the user to improve matching accuracy. When file paths do not match or attribute types are incompatible, the report will also output the unmatched paths and categorize them according to the reason for the mismatch, providing a more intuitive view of the matching results. Figure 7 As shown.

[0100] Based on this self-built dataset, the four core technologies integrated by the system—input data parsing and data path generation, abbreviation semantic recognition, path-level semantic matching, and type filtering—were tested and verified. The results show that:

[0101] Regarding abbreviation recognition and expansion, for the 78 abbreviations entered into the dictionary, this invention achieved a 100% matching accuracy rate thanks to its precise mapping mechanism; while for the 340 non-standardized abbreviations not entered into the dictionary, 311 were successfully matched through an intelligent inference strategy that fuses semantic vectors and character features, achieving an accuracy rate of 91.5%, effectively breaking through the dependence of traditional methods on predefined dictionaries.

[0102] Furthermore, to address the challenging scenario of polysemous abbreviations, 62 typical test cases covering different contexts were selected for further verification: traditional methods rely solely on character matching or a single semantic model, achieving a matching accuracy of 68%. However, this technology, through a contextual path semantic association mechanism, can resolve ambiguities by combining the overall path information of the field, increasing the matching accuracy to 95%, with an overall matching correctness of 93.1%. This result fully demonstrates its adaptability and accuracy in complex naming scenarios.

[0103] Regarding the interaction information of heterogeneous resources and the matching of common object models, in 1740 test data, the accuracy rate of attribute abbreviation names expanded after being entered into the dictionary reached 100%; the accuracy rate of abbreviations not entered into the dictionary reached 91.5% through intelligent expansion; the path matching success rate reached 97.8%; and the type compatibility check accuracy rate reached 100%. The matching report generated by the system integrates semantic relevance, structural matching degree and type compatibility analysis, providing interpretable and reliable results support for engineering applications.

[0104] The generated report combines semantic relevance, structural matching, and type compatibility analysis, providing "interpretable and reliable" matching results for engineering practice. Experimental verification shows that the method proposed in this invention not only solves the core pain points of traditional methods, but also achieves a leap from theoretical matching to engineering applicability through module collaboration, providing reliable technical support for scenarios such as code refactoring and cross-system data integration.

[0105] An intelligent matching system for heterogeneous resource interaction information and a common object model, based on an intelligent matching method for heterogeneous resource interaction information and a common object model, includes: a heterogeneous resource parsing module, an intelligent matching optimization module, and a system report generation module;

[0106] The heterogeneous resource parsing module includes a structure parsing module and a path generation and file saving module. It is used to parse heterogeneous resource interaction information through structure parsing, path generation and path information saving. It transforms the implicit type constraints and hierarchical associations in the information into a structured symbol sequence and outputs standardized attribute path information.

[0107] The intelligent matching optimization module is used to automatically match based on the standardized path information obtained from the parsing of heterogeneous resources, through the collaborative processing of abbreviation semantic recognition and expansion, semantic quantization and type verification, under complex naming rules and structural differences, and output the optimal mapping set that satisfies semantic association and technical feasibility.

[0108] The system report generation module is used to generate matching reports for compatibility analysis with the GBB common object model and correctness explanations of GBB-specific abbreviation extensions.

[0109] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements an intelligent matching method for heterogeneous resource interaction information and a common object model.

[0110] A processing terminal includes a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the computer program, it implements an intelligent matching method for heterogeneous resource interaction information and a common object model.

[0111] In another implementation, an intelligent matching method for heterogeneous resource interaction information and common object models is applicable to solving the information interaction barriers between multi-source heterogeneous resources (such as sensors with different protocols, multi-format databases, and heterogeneous service nodes) and standardized models in other distributed systems. It can be widely applied to scenarios such as industrial internet, cross-domain data fusion, and intelligent collaborative systems to achieve efficient and accurate matching between the two, thereby improving system resource utilization and collaborative response speed.

[0112] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0113] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for intelligently matching heterogeneous resource interaction information with a common object model, characterized in that, Includes the following steps: S1. For heterogeneous resource interaction information based on GBB, heterogeneous resource parsing is performed through structure parsing, path generation and path information storage. The implicit type constraints and hierarchical associations in the information are transformed into structured symbol sequences, and standardized attribute path information is output. S2. Based on the standardized path information obtained from heterogeneous resource parsing, through the collaborative processing of abbreviation semantic recognition and expansion, semantic quantization and type verification, automatic matching is achieved under complex naming rules and structural differences, and the optimal mapping set that satisfies semantic association and technical feasibility is output.

2. The intelligent matching method for heterogeneous resource interaction information and public object model according to claim 1, characterized in that, The heterogeneous resource interaction information is in the form of unstructured text, which is sequentially associated with the information unit name and the field description contained in specific symbols. Each field description includes a description representing the information type and a name representing the information identifier. The output attribute path information is an explicit mapping result of the logical structure of the input information. Each path starts with an information type description, sequentially connects the complete hierarchical relationship of the fields through a specific separator, and ends with a canonical symbol.

3. The intelligent matching method for heterogeneous resource interaction information and public object model according to claim 1, characterized in that, The specific content of step S1 includes: S11. Parse the structure definitions in the GBB interactive file, use regular expressions to match structure keywords, structure names and field information, and construct a structure dictionary and a top-level structure set; S12. For each top-level struct, use a recursive method to generate the complete paths of all its fields, and prepend the data type to the paths; S13. Save the generated path information to a file in the specified directory. If the saving attempt fails, output an error message.

4. The intelligent matching method for heterogeneous resource interaction information and public object model according to claim 3, characterized in that, In step S11, the contents of constructing the structure dictionary and the top-level structure collection are as follows: Collect all structure names and parse the field information of each structure. When processing field rows, support cases where the type contains spaces. At the same time, identify the referenced structures and find the top-level structure, that is, the structure defined in the file but not referenced by other structures. In step S12, the visited collection is used to record the visited structures. When processing each field, if the field is a nested structure, the nested structure is processed recursively; otherwise, the data type is added before the path.

5. The intelligent matching method for heterogeneous resource interaction information and public object model according to claim 1, characterized in that, The specific content of step S2 is as follows: S21. Split the standardized path data obtained from heterogeneous resource parsing and extract abbreviations; S22. Through a three-level processing mechanism of explicit expansion, intelligent inference expansion, and contextual ambiguity resolution, the fuzzy abbreviation is converted into the accurate full name, resulting in the standard abbreviation path; S23. For standard abbreviation paths, construct a multi-granularity evaluation system from the vocabulary, component level, and path level to perform semantic similarity matching and realize the quantification of deep semantic associations; S24. Perform type compatibility filtering and output the optimal mapping set.

6. The intelligent matching method for heterogeneous resource interaction information and public object model according to claim 5, characterized in that, The specific content of step S22 is as follows: S221. Explicit Expansion: Based on an abbreviation-full name mapping table built by domain experts, abbreviations with a high degree of standardization are directly replaced, covering industry-wide common abbreviations and protocol-specific abbreviations. The mapping table is dynamically updated through incremental learning. S222. Intelligent Inference Expansion: For non-standardized abbreviations not included in the mapping table, a dual-path inference model based on semantic vectors and character features is constructed, covering GBB domain-specific abbreviations. In the semantic dimension, candidate full names are screened by calculating the cosine similarity between the high-dimensional semantic vectors of the abbreviation to be expanded and the vocabulary in the domain corpus. In the character dimension, character-level feature scoring is constructed by combining root matching, edit distance, and suffix analysis. The results of the two channels are weighted and fused to generate a list of candidate full names with probability values. S223. Contextual Ambiguity Resolution: For polysemous abbreviations, a secondary screening is performed based on the hierarchical semantics of the path in which they are located to determine the optimal full name.

7. The intelligent matching method for heterogeneous resource interaction information and public object model according to claim 5, characterized in that, The specific content of step S23 is as follows: At the lexical level, the relevance of expanded semantic units is calculated using the cosine similarity formula of word vectors; At the component level, for the semantic unit sequence after splitting the composite field name, the optimal alignment method is found by dynamic programming algorithm. The weighted average of the maximum similarity of the aligned components is taken as the overall score and normalized by dividing by the length of the longer sequence. At the path level, the differences in hierarchical structure are handled based on the idea of ​​dynamic programming. A dynamic programming matrix is ​​constructed for the two path sequences, and the maximum sum of hierarchical similarity between the two path sequences is calculated.

8. The intelligent matching method for heterogeneous resource interaction information and public object model according to claim 5, characterized in that, The specific content of step S24 is as follows: The compatibility of basic data types is judged based on the rule that low precision is compatible with high precision. For custom data types, the compatibility of custom data types is determined by comparing whether the number of fields, field types, and field order of the structure are completely consistent, or by checking whether there is an inheritance relationship. Finally, the mapping pairs that simultaneously satisfy the semantic similarity greater than a preset threshold and are type compatible are retained to form the optimal mapping set.

9. The intelligent matching method for heterogeneous resource interaction information and public object model according to claim 1, characterized in that, Also includes: S3. By constructing a heterogeneous dataset covering GBB scenarios across multiple fields, the effectiveness, accuracy, and efficiency are verified, and a matching report is generated. The matching report includes the matching results, matching similarity and attribute compatibility descriptions, and whether the file extension path is correct. Users can modify the extended dictionary themselves based on whether the extension results are correct. When the file path does not match or the attribute type is incompatible, the report will output the unmatched path and categorize it according to the reason for the mismatch.

10. An intelligent matching system for heterogeneous resource interaction information and common object models, characterized in that, The intelligent matching method for heterogeneous resource interaction information and public object model according to any one of claims 1-9 includes: a heterogeneous resource parsing module, an intelligent matching optimization module, and a system report generation module; The heterogeneous resource parsing module, including the structure parsing module and the path generation and file saving module, is used to parse heterogeneous resource interaction information based on GBB through structure parsing, path generation and path information saving. It transforms the implicit type constraints and hierarchical associations in the information into a structured symbol sequence and outputs standardized attribute path information. The intelligent matching optimization module is used to automatically match based on the standardized path information obtained from the parsing of heterogeneous resources, through the collaborative processing of abbreviation semantic recognition and expansion, semantic quantization and type verification, under complex naming rules and structural differences, and output the optimal mapping set that satisfies semantic association and technical feasibility. The system report generation module is used to generate matching reports for compatibility analysis with the GBB common object model and correctness explanations of GBB-specific abbreviation extensions.

Images