Method for constructing password product evaluation knowledge base
By building a cryptographic product evaluation knowledge base and combining a knowledge graph with entity model and relational model, the problem of lack of associated representation of cryptographic product evaluation knowledge is solved, and multi-dimensional testing support and technical upgrades are achieved.
Patent Information
- Application Number
- CN202510440621.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-09
- Publication Date
- 2025-08-15
AI Technical Summary
The existing cryptographic product evaluation knowledge lacks correlation and unified representation, and has not formed a complete knowledge base, which cannot effectively support multi-dimensional testing evaluation and technical upgrades.
Build a cryptographic product evaluation knowledge base, establish an evaluation knowledge base model through entity models and relationship models, use knowledge graphs to represent and enhance information, and establish association relationships based on document crawling and standard specifications to form a cryptographic product evaluation knowledge base.
It realizes formal, standardized representation and visual display of cryptographic product evaluation knowledge, provides highly targeted testing support, and improves the security performance evaluation and technical upgrade of cryptographic product.
Smart Images
Figure CN120494059A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of cryptographic product evaluation and relates to a method for constructing a cryptographic product evaluation knowledge base, specifically a method for constructing a cryptographic product evaluation knowledge base involving information extraction and knowledge graph. Background Art
[0002] With the development of network technology, business operations across all industries rely on the acquisition and transmission of network information. Therefore, information security is crucial to social stability and industry safety. Cryptographic products encrypt and protect user identities and data through cryptographic algorithms and protocols, making their security performance crucial to information security. Cryptographic product evaluations assess cryptographic products across multiple dimensions, including functionality and algorithms, through test cases. Through comprehensive testing and analysis, they aim to ensure that cryptographic products and services comply with relevant technical standards and specifications, thereby guiding technological upgrades and quality improvements.
[0003] Currently, existing cryptographic products have a wide variety of functions and types, encompassing multiple test dimensions, including chips, software modules, and complete devices. Therefore, a large number of test cases are required to ensure the security performance of cryptographic products from multiple perspectives. Currently, knowledge related to cryptographic product evaluation mostly resides in various online resources, such as literature. This information lacks connectivity and unified representation, and a comprehensive knowledge base has yet to be established to provide domain knowledge support for cryptographic product evaluation. Knowledge graphs, due to their graph-like representation, are capable of representing complex entities. Nodes represent entity concepts, where entity attributes further represent information about the entity in different dimensions. Edges represent relationships between entity concepts. The semantic relationship network formed by entities and relationships allows for a formalized and standardized representation of cryptographic product evaluation knowledge. The visualization of knowledge graphs clearly demonstrates the internal information relationships and logic of evaluation knowledge, enabling targeted testing of relevant products. Summary of the Invention
[0004] In view of the problems existing in the prior art, the present invention aims to provide a method for constructing a knowledge base for cryptographic product evaluation.
[0005] The technical solution of the present invention is:
[0006] A method for constructing a cryptographic product evaluation knowledge base, comprising the following steps:
[0007] 1) Establish an entity model for each entity type in the cryptographic product, the entity model including the entity name and entity attributes; construct several relationship models based on the association relationships between the various entities; and construct a cryptographic product evaluation knowledge base model based on the entity model and relationship model; the entity types include cryptographic product capabilities, cryptographic product functions, cryptographic product types, cryptographic product test types, cryptographic product standards and specifications, products, tools, use cases, and issues;
[0008] 2) Select cryptographic product capabilities, cryptographic product functions, cryptographic product types, and cryptographic product test types as target entity types; establish a hierarchical view of the corresponding target entity type based on the inclusion relationship between entities of the same target entity type; then store the structured information in the hierarchical view in the form of a graph to obtain a cryptographic product technology system graph;
[0009] 3) Crawling documents using the entity names in the graph as literature search terms to obtain the document titles, abstracts, keywords, and texts of the crawled documents; then enhancing the graph information of the cryptographic product technology system graph based on the obtained keywords and texts;
[0010] 4) Obtain attribute information of entities of cryptographic product standard specification type from online open source resources to form structured entities; establish association relationships between entities of cryptographic product standard specification type and entities in the cryptographic product technology system map through co-occurrence relationships;
[0011] 5) Using the entities of the cryptographic product test type in the cryptographic product technology system map as search terms, crawling test-related cryptographic product evaluation system knowledge; then extracting product, tool, use case, and problem-related entities from the literature documents, and constructing triples (product, association, use case), (use case, association, tool), (use case, association, problem) to obtain a cryptographic product evaluation system map;
[0012] 6) For the product entities in the cryptographic product evaluation system map, establish an association relationship between them and the entities of the cryptographic product type in the cryptographic product technology system map, and associate the cryptographic product evaluation system map with the cryptographic product technology system map through the triple (product, association relationship, cryptographic product type) to obtain the cryptographic product evaluation knowledge base.
[0013] Furthermore, the method for enhancing the graph information based on the obtained document keywords is as follows: merging the obtained document keywords to form a keyword set, and counting the word frequency of each document keyword in the keyword set; if the word frequency of the document keyword exceeds the set threshold N1, the corresponding document keyword is used as a high-frequency keyword; and associating the high-frequency keyword with the entity used for retrieval to form a triple (keyword, association relationship, retrieval entity).
[0014] Furthermore, the method for enhancing the graph information based on the obtained document text is as follows: first, the document text content crawled in step 3) is segmented, and then the associated words that have potential relationships with the entity names already existing in the cryptographic product technology system map are extracted through Chinese grammatical analysis. Finally, the tf-idf analysis method is used to count the word frequency of the potential associated words. If the word frequency of the potential associated words is greater than the set threshold N2, the corresponding potential associated words are retained as key associated words and written into the cryptographic product technology system map.
[0015] Furthermore, the method of extracting related words that have potential relationships with existing entity names in the graph through Chinese grammatical analysis is as follows:
[0016] 31) For each sentence in a document, construct a directed graph, where words are nodes and dependencies are edges, to show the grammatical relationships between words in the sentence; the dependency result in the directed graph can be expressed as l(id,w,head,v), where w is the current word, id is the index position of the current word, head is the index of the parent dependency word, and v is the dependency relationship with the parent word;
[0017] 32) Use the following three structures to obtain the associated words with potential relationships;
[0018] a) Subject-verb-object structure
[0019] If there are three participles, the dependency analysis results are expressed as s=(id s ,w s ,head s ,v s ), v=(id v ,w v ,head v ,v v ), o=(id o ,w o ,head o ,v o ), if the following conditions are met, the three participles have a subject-verb-object relationship, and s is the subject, v is the predicate, and o is the object;
[0020]
[0021] Among them, SVB is the subject-verb relationship, HED is the core relationship of the sentence, and VOB is the direct object;
[0022] If the subject participle or object participle is an existing entity in the cryptographic product technology system map, then the corresponding object participle or subject participle is added to the cryptographic product technology system map and an association relationship is established with the existing word;
[0023] b) Parallel structure
[0024] If there are two participles, their dependency analysis results are expressed as b=(id b ,w b ,head b ,v b ), d=(id d ,w d ,head d ,v d ), if the following conditions are met, the two participles have a parallel relationship;
[0025]
[0026] Among them, COO is a parallel relationship;
[0027] If one of the segmented words is an entity that already exists in the cryptographic product technology system map, then the other segmented word is added to the cryptographic product technology system map and an association relationship is established with the existing word;
[0028] c) Object-attributive relationship
[0029] If there are two participles o and t, where o is the object participle, the dependency analysis results are expressed as o=(id o ,w o ,head o ,v o ), t=(id t ,w t ,head t ,v t ), if the following conditions are met, then the two participles have an object-attributive-predicate relationship;
[0030]
[0031] Among them, ATT is a subject-predicate relationship, and it is known that o is a meaningless object participle, then t is the attribute description of the object. The attribute description t of the object o can be concatenated with the object o and stored in the cryptographic product technology system map.
[0032] Furthermore, the TF-IDF analysis method is used to count the frequency of potential related words; the calculation formula of TF-IDF is TF-IDF=TF*IDF, TF is the word frequency, IDF is the inverse document frequency; the word frequency of the i-th word n i,jIndicates that the i-th word is in the document d j The number of times it appears in the file d, the denominator is j The total number of times all words appear in n i,j =∑1 / d, d is the number of times the i-th word appears in document d j The degree of the core word in sentence c, n represents the total number of documents, n i The number of documents containing the i-th word.
[0033] Furthermore, the relationship model includes a relationship model between cryptographic product capabilities and cryptographic product standards and specifications, a relationship model between cryptographic product functions and cryptographic product standards and specifications, a relationship model between cryptographic product types and cryptographic product standards and specifications, a relationship model between cryptographic product test types and cryptographic product standards and specifications, a relationship model between products and cryptographic product types, a relationship model between products and use cases, a relationship model between use cases and tools, and a relationship model between use cases and problems.
[0034] Furthermore, the method of establishing an association relationship between entities of the cryptographic product standard specification type and entities in the cryptographic product technology system map through co-occurrence relationships is: taking the entities in the cryptographic product technology system as a dictionary set, performing entity recognition on the text corresponding to the summary attribute field in the cryptographic product standard specification; if the recognized entity is an entity in the cryptographic product technology system in the dictionary set, then establishing an association relationship between the entity of the cryptographic product standard specification type and the corresponding entity in the cryptographic product technology system.
[0035] A server, characterized in that it includes a memory and a processor, the memory stores a computer program, the computer program is configured to be executed by the processor, and the computer program includes instructions for executing the above method.
[0036] A computer-readable storage medium stores a computer program thereon, wherein the computer program implements the above method when executed by a processor.
[0037] The advantages of the present invention are as follows:
[0038] (1) Designed a process for building a knowledge base for cryptographic product evaluation, providing process guidance for knowledge base construction from the perspectives of domain modeling and data extraction.
[0039] (2) The knowledge base model structure was designed based on the business knowledge in the field of cryptographic product evaluation, and the domain knowledge was modeled and described through the ontology model and ontology relationship.
[0040] (3) The password evaluation knowledge base is comprehensively constructed from the product technology system and the evaluation system. The present invention not only provides a knowledge base model, but also provides the specific construction process of the two. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] Figure 1 A flow chart for constructing the cryptographic product evaluation knowledge base of the present invention.
[0042] Figure 2 This is a model diagram of the knowledge base for password product evaluation.
[0043] Figure 3 A hierarchical view structured diagram.
[0044] Figure 4 This is a flowchart for enhancing the atlas processing in the main text of the document.
[0045] Figure 5 This is the dependency analysis result diagram of "SPECK-type algorithms are a class of lightweight block cipher algorithms."
[0046] Figure 6 This is the dependency result diagram of "cryptographic product capabilities including cryptographic operations, cryptographic anti-counterfeiting, electromagnetic compatibility, etc." DETAILED DESCRIPTION
[0047] The present invention will be described in further detail below with reference to the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.
[0048] In order to build a knowledge base for cryptographic product evaluation, information is extracted from literature containing evaluation knowledge to form structured data. The graph database is used as the storage medium to manage and visualize the structured data and provide related search functions. The specific implementation steps are as follows: Figure 1 As shown, we first need to design a cryptographic product evaluation knowledge base model, then extract knowledge based on the model structure, and construct the cryptographic product technology system map and the cryptographic product evaluation system map in turn. Finally, we integrate the knowledge of the two parts of the map to form a cryptographic product evaluation knowledge base.
[0049] 1. Design a knowledge base model for password product evaluation
[0050] The password product evaluation knowledge base model needs to comprehensively summarize a large amount of domain business knowledge and extract and abstract the formalized and unified representation of concepts, examples and other information related to the password product evaluation field. This patent uses knowledge graphs to store and represent evaluation knowledge, so the password evaluation knowledge base model is constructed through entity models and relationship models. The specific entity model and relationship model definitions are shown in Table 1 and Table 2, and the overall structure of the model is shown in Figure 2 shown.
[0051] The entity models include the following 9 types:
[0052] Cryptographic product capabilities: cryptographic product encryption and decryption capabilities such as cryptographic operations, cryptographic anti-counterfeiting, and electromagnetic compatibility.
[0053] Cryptographic product functions: Cryptographic security functions such as physical and environmental security, network communication security, computing security, and data security.
[0054] Cryptographic product type: Cryptographic products are classified into product types based on algorithms, chips, modules, boards, systems, and complete machines.
[0055] Cryptographic product testing types: Cryptographic product testing types are divided from the perspectives of software, algorithm, hardware, attack penetration, electromagnetic compatibility, etc.
[0056] Cryptographic product standards and specifications: Cryptographic product standards and specifications are the basis for ensuring the security of cryptographic products and systems, and also stipulate technical requirements and testing methods for cryptographic products.
[0057] Product: Specific cryptographic product.
[0058] Tools: Testing tools required to perform the cryptographic product testing process.
[0059] Use case: Cryptography product test case.
[0060] Problem: Product problems discovered when executing the test task of the cryptographic product.
[0061] The relationship model is the following 8 models, which are mainly the association relationships of related entities, and the association relationships here are bidirectional relationships between subjects and objects: (cryptographic product capabilities, association relationships, cryptographic product standards and specifications), (cryptographic product functions, association relationships, cryptographic product standards and specifications), (cryptographic product types, association relationships, cryptographic product standards and specifications), (cryptographic product test types, association relationships, cryptographic product standards and specifications), (products, association relationships, cryptographic product types), (products, association relationships, use cases), (use cases, association relationships, tools), (use cases, association relationships, problems).
[0062] Table 1 Entity model information table
[0063]
[0064] Table 2 Relationship model information table
[0065]
[0066]
[0067] 2. Construction of cryptographic product technology system map
[0068] The construction of the cryptographic product technology system map includes three steps: (1) the systematization and storage of the cryptographic product technology system; (2) the enhancement of the cryptographic product technology system map; and (3) the structured extraction and association of cryptographic product standards and specifications.
[0069] (1) Cryptographic product technology systematization and storage
[0070] The cryptographic product technology system includes entity types such as cryptographic product capability, cryptographic product function, cryptographic product type, cryptographic product test type and cryptographic product standard specification. For entities of entity types such as cryptographic product capability, cryptographic product function, cryptographic product type and cryptographic product test type, this patent constructs hierarchical views such as Figure 3 The information in the hierarchical view is manually edited to form entities of corresponding types, and the information relationship between high-level entities and low-level entities is represented by the "include" relationship, as shown in Figure 3 Then the structured information triples are stored in the neo4j graph database in the form of entities and relationships.
[0071] (2) Enhanced cryptographic product technology system map
[0072] Through the systematization and storage of the aforementioned cryptographic product technology architecture, hierarchical knowledge related to cryptographic product capabilities, functions, types, and test types is stored in the form of a graph. Because the aforementioned entity and relationship data is insufficient to encompass all domain knowledge, this patent utilizes open-source literature to enhance the information in the constructed graph.
[0073] First, we crawled the documents using the entity names in the graph as search terms, obtaining the document title, abstract, keywords, and text of the crawled documents, and then enhanced the graph information based on the obtained keywords and text.
[0074] The obtained document keywords are merged into a keyword set and their frequency is counted, that is, the number of times each keyword appears as a keyword in the crawled documents. If the frequency of a keyword exceeds a set threshold N1, the corresponding keyword is considered a high-frequency keyword. Finally, the high-frequency keywords are associated with the entities used for retrieval to form a triple (keyword, association, retrieval entity), completing the graph information enhancement.
[0075] The enhanced processing of the crawled document text content on the graph is achieved by using a method based on Chinese grammatical analysis and word frequency statistics. First, the document text content is preprocessed, and then Chinese grammatical analysis is used to extract the associated words that have potential relationships with the entity names that already exist in the cryptographic product technology system graph. Finally, the tf-idf analysis method is used to count the word frequencies of the potential associated words. If the word frequency of the potential associated words is greater than the set threshold N2, the corresponding potential associated words will be retained as key associated words and written into the graph. The process is as follows: Figure 4 shown.
[0076] 1) Document preprocessing
[0077] For batch documents (n is the number of documents) preprocessing, with a single document d i As the processing unit. First, the document is segmented into paragraphs and sentences in turn. The paragraph segmentation is marked with "\n", and the sentence segmentation is marked with ".", "?", "!", "...", and the single sentence set S is obtained. i ={s i1 ,s i2 ,…,s im}, where m is d i The total number of sentences included in . Then, the word segmenter is used to segment the sentences. Before segmentation, the entities in the knowledge base are added to the word segmentation dictionary to retain the integrity of the word segmentation information of the general dictionary in a specific field. Finally, according to the content of the stop word list, stop words are removed from the word segmentation to obtain the word segmentation set C i ={c i1 ,c i2 ,…,c il}, where l is d i The total number of participles included in .
[0078] 2) Mining potential related words based on grammatical analysis
[0079] Dependency parsing is a language processing technology that reveals the syntactic structure of a language unit by analyzing the dependency relationships between components within the unit. The core idea of this analysis method is to regard a sentence as a directed graph, where words are nodes and dependency relationships are edges. By constructing such a graph structure, the grammatical relationship between each word in the sentence can be clearly displayed. The smallest granularity unit in dependency syntactic analysis is the segmentation word, and segmentations with dependency relationships have direct or indirect associations on the grammatical graph. This patent analyzes the grammatical structure components of the cryptographic evaluation text and summarizes three templates to expand and enhance the information in the cryptographic product evaluation knowledge base. Here, the templates are introduced based on the dependency analysis results of the text "SPECK-type algorithms are a type of lightweight block cipher algorithm" and "Cryptographic product capabilities include cryptographic operations, cryptographic anti-counterfeiting, electromagnetic compatibility, etc."
[0080] Dependency analysis uses a segmentation as the smallest unit. The dependency result can be expressed as l(id, w, head, v), where w is the current segmentation, id is the index position of the current segmentation, head is the index of the parent dependency segmentation, and v is the dependency relationship with the parent segmentation. By analyzing and summarizing the dependency analysis structure, we extract potentially related business concepts for cryptographic product evaluation. The following three structures are mainly used.
[0081] a) Subject-verb-object structure
[0082] The subject-verb-object structure is mainly defined by dependency relations. If there are three participles, the dependency analysis results are expressed as s=(id s ,w s ,head s ,v s ), v=(id v ,w v ,head v ,v v ), o=(id o ,w o ,head o ,v o ), if the following conditions are met, the three participles have a subject-verb-object relationship, with s as the subject, v as the predicate, and o as the object. At the same time, this condition can be used to extract information about the subject-verb-object structure.
[0083]
[0084] Among them, SVB is the subject-verb relationship, HED is the core relationship of the entire sentence, and VOB is the direct object. If the subject participle or object participle of the triple extracted by this rule template is an existing entity in the knowledge graph, then its corresponding object participle or subject participle will be added to the knowledge graph and an association relationship will be established with the existing word. Figure 5 (SPECK-type algorithm is a block cipher algorithm) can be obtained by extracting the dependency analysis result of "SPECK-type algorithm is a class of lightweight block cipher algorithms" through the above rules.
[0085] b) Parallel structure
[0086] The parallel relationship describes the dependency structure of two participles and can be defined as follows: if there are two participles, the dependency analysis results are expressed as b=(id b ,w b ,head b ,v b ), d=(id d ,w d ,head d ,vd ), if the following conditions are met, the two participles have a parallel relationship, and the information extraction of the parallel relationship structure can be realized through this condition.
[0087]
[0088] Among them, COO is a parallel relationship. If one of the word segments in the two-tuple extracted by this rule template is an entity that already exists in the knowledge graph, the other word segment will be added to the knowledge graph and an association relationship will be established with the existing word. The above "cryptographic product capabilities include cryptographic operations, cryptographic anti-counterfeiting, electromagnetic compatibility, etc." is analyzed by dependency and the results are as follows: Figure 6 As shown, the parallel structure information that can be extracted through the above rules is (cryptographic operation, cryptographic anti-counterfeiting) and (cryptographic operation, electromagnetic compatibility).
[0089] c) Object-attributive relationship
[0090] For the above statement "Cryptographic product capabilities include cryptographic operations, cryptographic anti-counterfeiting, and electromagnetic compatibility," the information extracted using the subject-verb-object rule is (cryptographic product capabilities, including, capabilities), where the object is "capability." Object segmentations without practical meaning, such as "capability," "structure," and "function," cannot be directly stored in the knowledge graph. Further analysis of their attributive-predicate relationships, or attribute description relationships, is required. Specific object segmentations with attributive-predicate relationships can provide more business value.
[0091] The object-attributive relation rule can be defined as follows: if there are two participles o and t, where o is the object participle, the dependency analysis results are expressed as o=(id o ,w o ,head o ,v o ), t=(id t ,w t ,head t ,v t ), if the following conditions are met, the two participles have an object-attributive-predicate relationship, and this condition can be used to extract information about the parallel relationship structure.
[0092]
[0093] Here, ATT represents a subject-attribute relationship. Given that o is a meaningless object participle, t represents the attribute description of the object. The attribute description t of object o can be concatenated with the object o and stored in the knowledge graph. For example, the object of the text "Cryptographic product capabilities include cryptographic operations, cryptographic anti-counterfeiting, and electromagnetic compatibility" is "capability." Using the above rules, the object-attribute relationship tuple extracted is (cryptographic operations, capability), where "cryptographic operations" is the attribute description of the object "capability." After concatenation, "cryptographic operations capability" is obtained.
[0094] 3) Analysis of Key Associated Words Based on Word Frequency
[0095] A large number of potential associated words are obtained from the full texts of a batch of documents through the method of mining potential associated words based on syntactic analysis. However, not all associated words can supplement valuable information to the evaluation knowledge base. Therefore, this method evaluates the importance of the associated words in the field by analyzing their word frequencies through TF-IDF.
[0096] TF-IDF (Term Frequency–Inverse Document Frequency) is a statistics-based technology for retrieval and text mining, used to evaluate the importance of a word for a document in a document set or a corpus. This patent optimizes the calculation method of TF-IDF by combining the dependency degree between the associated word and the core word. For example Figure 5 The relationships between "SPECK algorithm" and "block cipher algorithm" and the core word "is" are both of degree one. Therefore, this method not only represents the importance of the associated words in terms of frequency but also combines the importance of the associated words in terms of semantics.
[0097] The calculation formula of TF-IDF is TF-IDF = TF * IDF, where TF is the term frequency and IDF is the inverse document frequency.
[0098] The term frequency represents the frequency of a segmented word in the text:
[0099]
[0100] where, n i,j represents the number of times the i-th segmented word appears in document d j and the denominator is the total number of times all words appear in document d j This patent comprehensively considers the actual number of times a segmented word appears and the degree between the segmented word and the core word of the sentence where it is located when counting the number of times n i,j of the segmented word in the text, which can be expressed as: n i,j = ∑1 / d, where d is the degree between the i-th segmented word and the core word each time it appears in sentence c of document d j
[0101] The inverse document frequency indicates that the fewer the number of documents containing a certain word, the higher the importance of the word:
[0102]
[0103] where, n represents the total number of documents, and n i is the number of documents containing the i-th segmented word.
[0104] 4) Importing Key Associated Words into the Knowledge Base
[0105] After frequency segmentation of the associated words, the words with the top 30% TF-IDF values are retained and stored in the knowledge graph, and association relationships are established with the entities associated with them that already exist in the graph.
[0106] (3) Structural extraction of cryptographic product standards and specifications and their association with the technical system
[0107] According to the standard specification type in the model designed by this patent, the attributes of the standard specification entity need to include the specification name, specification number, summary, release date, and implementation date. The standard text of the cryptographic product specification is obtained from open source resources on the Internet, and its specification name, specification number, summary, release date, and implementation date information is obtained to form a structured entity.
[0108] Entities of the cryptographic product standard specification type and entities in the cryptographic product technical system are associated through co-occurrence relationships. The specific method is: using the entities in the cryptographic product technical system as a dictionary set, entity recognition is performed on the text corresponding to the summary attribute field in the cryptographic product standard specification. If the recognized entity is an entity in the cryptographic product technical system in the dictionary set, an association relationship is established between the entity of the cryptographic product standard specification type and the corresponding entity in the cryptographic product technical system.
[0109] 3. Construction of Cryptographic Product Evaluation System Map
[0110] The construction of the cryptographic product evaluation system map includes two steps: (1) acquisition and structuring of cryptographic product evaluation knowledge; (2) integration of cryptographic product evaluation system and cryptographic product technology system knowledge.
[0111] (1) Acquisition and structuring of knowledge of the cryptographic product evaluation system to obtain a cryptographic product evaluation system map
[0112] According to the test-related entity model in the cryptography product evaluation knowledge base model, the cryptography product evaluation system includes information on products, tools, use cases, and issues. Literature containing relevant information needs to be obtained from open source resources. This patent uses the entity type "cryptography product testing" in the cryptography product technology system as a search term to crawl test-related literature documents. Entities related to products, tools, use cases, and issues are extracted from these documents, and relationships such as (product, association, use case), (use case, association, tool), and (use case, association, issue) are constructed to obtain a map of the cryptography product evaluation system.
[0113] (2) Integrate the cryptographic product evaluation system map with the cryptographic product technology system map
[0114] For product entities in the evaluation system, an association relationship is established with the cryptographic product types in the technical system, namely (product, association relationship, cryptographic product type). After the establishment of this relationship, the cryptographic product evaluation system and the cryptographic product technical system are associated through the triple (product, association relationship, cryptographic product type). The two form a connected information domain, resulting in a cryptographic product evaluation knowledge base presented in the form of a graph.
[0115] While specific embodiments of the present invention have been disclosed for illustrative purposes, intended to facilitate understanding and implementation of the present invention, those skilled in the art will appreciate that various substitutions, variations, and modifications are possible without departing from the spirit and scope of the present invention and the appended claims. Therefore, the present invention should not be limited to the disclosure of the preferred embodiments, and the scope of protection claimed in the present invention shall be determined by the scope of the claims.
Claims
1. A method for constructing a cryptographic product evaluation knowledge base, comprising the following steps: 1) Establish an entity model for each entity type in the cryptographic product, the entity model including the entity name and entity attributes; construct several relationship models based on the association relationships between the various entities; and construct a cryptographic product evaluation knowledge base model based on the entity model and relationship model; the entity types include cryptographic product capabilities, cryptographic product functions, cryptographic product types, cryptographic product test types, cryptographic product standards and specifications, products, tools, use cases, and issues; 2) Select cryptographic product capabilities, cryptographic product functions, cryptographic product types, and cryptographic product test types as target entity types; establish a hierarchical view of the corresponding target entity type based on the inclusion relationship between entities of the same target entity type; then store the structured information in the hierarchical view in the form of a graph to obtain a cryptographic product technology system graph; 3) Crawling documents using the entity names in the graph as literature search terms to obtain the document titles, abstracts, keywords, and texts of the crawled documents; then enhancing the graph information of the cryptographic product technology system graph based on the obtained keywords and texts; 4) Obtain attribute information of entities of cryptographic product standard specification type from online open source resources to form structured entities; establish association relationships between entities of cryptographic product standard specification type and entities in the cryptographic product technology system map through co-occurrence relationships; 5) Using the entities of the cryptographic product test type in the cryptographic product technology system map as search terms, crawling test-related cryptographic product evaluation system knowledge; then extracting product, tool, use case, and problem-related entities from the literature documents, and constructing triples (product, association, use case), (use case, association, tool), (use case, association, problem) to obtain a cryptographic product evaluation system map; 6) For the product entities in the cryptographic product evaluation system map, establish an association relationship between them and the entities of the cryptographic product type in the cryptographic product technology system map, and associate the cryptographic product evaluation system map with the cryptographic product technology system map through the triple (product, association relationship, cryptographic product type) to obtain the cryptographic product evaluation knowledge base.
2. The method according to claim 1, characterized in that The method for enhancing graph information based on the obtained document keywords is as follows: merging the obtained document keywords into a keyword set, and counting the word frequency of each document keyword in the keyword set; if the word frequency of a document keyword exceeds the set threshold N1, the corresponding document keyword is used as a high-frequency keyword; and associating the high-frequency keyword with the entity used for retrieval to form a triple (keyword, association relationship, retrieval entity).
3. The method according to claim 1, characterized in that The method for enhancing the graph information based on the obtained document text is as follows: first, the document text content crawled in step 3) is segmented, and then the associated words that have potential relationships with the entity names already existing in the cryptographic product technology system map are extracted through Chinese grammatical analysis. Finally, the tf-idf analysis method is used to count the word frequencies of the potential associated words. If the word frequency of the potential associated words is greater than the set threshold N2, the corresponding potential associated words are retained as key associated words and written into the cryptographic product technology system map.
4. The method according to claim 3, characterized in that The method of extracting related words that have potential relationships with existing entity names in the graph through Chinese grammatical analysis is: 31) For each sentence in a document, construct a directed graph, where words are nodes and dependencies are edges, to show the grammatical relationships between words in the sentence; the dependency result in the directed graph can be expressed as l(id,w,head,v), where w is the current word, id is the index position of the current word, head is the index of the parent dependency word, and v is the dependency relationship with the parent word; 32) Use the following three structures to obtain the associated words with potential relationships; a) Subject-verb-object structure If there are three participles, the dependency analysis results are expressed as s=(id s ,w s ,head s ,v s ), v=(id v ,w v ,head v ,v v ), o=(id o ,w o ,head o ,v o ), if the following conditions are met, the three participles have a subject-verb-object relationship, and s is the subject, v is the predicate, and o is the object; Among them, SVB is the subject-verb relationship, HED is the core relationship of the sentence, and VOB is the direct object; If the subject participle or object participle is an existing entity in the cryptographic product technology system map, then the corresponding object participle or subject participle is added to the cryptographic product technology system map and an association relationship is established with the existing word; b) Parallel structure If there are two participles, their dependency analysis results are expressed as b=(id b ,w b ,head b ,v b ), d=(id d ,w d ,head d ,v d ), if the following conditions are met, the two participles have a parallel relationship; Among them, COO is a parallel relationship; If one of the segmented words is an entity that already exists in the cryptographic product technology system map, then the other segmented word is added to the cryptographic product technology system map and an association relationship is established with the existing word; c) Object-attributive relationship If there are two participles o and t, where o is the object participle, the dependency analysis results are expressed as o=(id o ,w o ,head o ,v o ), t=(id t ,w t ,head t ,v t ), if the following conditions are met, then the two participles have an object-attributive-predicate relationship; Among them, ATT is a subject-predicate relationship, and it is known that o is a meaningless object participle, then t is the attribute description of the object. The attribute description t of the object o can be concatenated with the object o and stored in the cryptographic product technology system map.
5. The method according to claim 3, characterized in that The TF-IDF analysis method is used to count the frequency of potential related words; the calculation formula of TF-IDF is TF-IDF=TF*IDF, TF is the word frequency, IDF is the inverse document frequency; the word frequency of the i-th word n i,j Indicates that the i-th word is in the document d j The number of times it appears in the file d, the denominator is j The total number of times all words appear in n i,j =∑1 / d, d is the number of times the i-th word appears in document d j The degree of the core word in sentence c, n represents the total number of documents, n i The number of documents containing the i-th word.
6. The method according to claim 1, characterized in that The relationship model includes the relationship model between cryptographic product capabilities and cryptographic product standards and specifications, the relationship model between cryptographic product functions and cryptographic product standards and specifications, the relationship model between cryptographic product types and cryptographic product standards and specifications, the relationship model between cryptographic product test types and cryptographic product standards and specifications, the relationship model between products and cryptographic product types, the relationship model between products and use cases, the relationship model between use cases and tools, and the relationship model between use cases and problems.
7. The method according to claim 6, characterized in that The method for establishing an association relationship between entities of the cryptographic product standard specification type and entities in the cryptographic product technology system map through co-occurrence relationships is as follows: using entities in the cryptographic product technology system as a dictionary set, performing entity recognition on the text corresponding to the summary attribute field in the cryptographic product standard specification; if the recognized entity is an entity in the cryptographic product technology system in the dictionary set, then establishing an association relationship between the entity of the cryptographic product standard specification type and the corresponding entity in the cryptographic product technology system.
8. A server, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, the computer program is configured to be executed by the processor, and the computer program comprises instructions for executing the method according to any one of claims 1 to 7.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.