Method, device and storage medium for constructing digital battlefield knowledge graph ontology

Abstract

The present invention provides a method, device, and storage medium for constructing a digital battlefield knowledge graph ontology. The method comprises: S1, forming a term dictionary for a selected domain; S2, preprocessing the original corpus to obtain a concept vocabulary; S3, using the TF-IWF algorithm to determine the core concept vocabulary; S4, clustering similar concepts using the ART network based on the core concept vocabulary to obtain cluster groups; S5, selecting the word with the highest frequency from the cluster group according to the word frequency as the ontology concept of the class; S6, clustering the cluster group again based on a finer classification threshold to obtain a cluster group of finer classes, and repeating S5-S6 to obtain the hierarchical relationship of the cluster concepts; S7, serializing the ontology concepts and the hierarchical relationship of the concepts into an OWL file. The above technical solution can semi-automatically convert domain documents into a computer-processable domain ontology, effectively improving the efficiency of constructing the digital battlefield domain ontology.

Description

Technical Field

[0001] The present invention relates to the technical field of ontology construction, and in particular to a method, device and storage medium for constructing a digital battlefield knowledge graph ontology. Background Art

[0002] An ontology is a model of concepts and the relationships between them within a domain, enabling a formal description of that domain. The digital battlefield ontology formally describes concepts and their relationships within the digital battlefield (military) domain. Domain ontologies construct a unified set of concepts, overcoming communication barriers between people, organizations, and systems due to differing backgrounds, languages, and technologies, enabling the sharing and reuse of domain knowledge. However, the construction of digital battlefield ontologies is highly specialized and difficult, often relying on manual construction by domain experts, resulting in low efficiency.

[0003] The existing ontology construction methods at home and abroad mainly include TOVE method, skeleton method, IDEF5 method, seven-step method, etc., but there is no complete set of ontology construction engineering methods, nor is there a mature method for ontology construction specifically for domain Chinese document knowledge. Summary of the Invention

[0004] Embodiments of the present invention provide a method, device, and storage medium for constructing an ontology of a digital battlefield knowledge graph to improve the efficiency of ontology construction in the digital battlefield field.

[0005] To achieve the above objectives, a method for constructing a digital battlefield knowledge graph ontology is provided, including:

[0006] Step S1: Based on the selected domain of the digital battlefield knowledge graph ontology, combined with the collected literature and terminology related to the selected domain, a term dictionary of the selected domain is formed, and the term dictionary is added as a custom dictionary to the selected natural language processing tool;

[0007] Step S2: preprocessing the original corpus using the selected natural language processing tool and custom dictionary to obtain the concept vocabulary related to the selected field in the original corpus, wherein the preprocessing includes word segmentation, part-of-speech tagging, and removal of interference items, including modal particles, prepositions, and quantifiers;

[0008] Step S3, using the Feature Term Frequency-Inverse Word Frequency (TF-IWF) algorithm to calculate the domain weight of each concept word in the obtained concept words, and determining the core concept words in the obtained concept words according to the calculated domain weight;

[0009] Step S4, based on the core concept vocabulary, similar concepts are clustered using a recursive adaptive resonance theory ART network to obtain a cluster group;

[0010] Step S5, selecting a candidate word with the largest frequency from the cluster group according to the frequency size as the ontology concept representing the class corresponding to the cluster group, and removing the selected candidate word from the cluster group;

[0011] Step S6: cluster the clusters that have been removed from the candidate words again based on the selected finer classification threshold to obtain clusters at a finer level. Then, the process goes to step S5 and loops through steps S5-S6 until the clusters cannot be further subdivided, thereby obtaining a hierarchical relationship of cluster concepts.

[0012] Step S7: Serialize the obtained ontology concepts and hierarchical relationships of the concepts into a computer-processable network ontology language OWL file.

[0013] Preferably, in the method, step S7 comprises:

[0014] The API provided by Jena semantic web framework is used to serialize ontology concepts and hierarchical relationships of concepts into computer-processable OWL files in the form of resource description framework RDF and OWL format.

[0015] Preferably, in the method, in step S3, the following TF-IWF formula is used to calculate the domain weight of the concept vocabulary:

[0016]

[0017] Among them TF i,j Expressing word t i The frequency of feature terms in document j, n i,j Expressing word t i The frequency of occurrence in text j, ∑ k n k,j Represents the sum of the frequencies of all k words in text j, IWF i Expressing word t i The inverse frequency of a word in a corpus of m words, Represents the sum of the frequencies of all m words in the corpus, nt i Expressing word t i The total frequency of occurrences in the corpus.

[0018] Preferably, in the method, step S3 comprises:

[0019] Step S31, inputting a document set D = {di, i = 1, 2, ..., N} consisting of the obtained concept words di, where N is the number of concept words;

[0020] Step S32, calculating the domain weight of each concept word according to the TF-IWF formula, that is, the TF-IWF value;

[0021] Step S33 , traversing the TF-IWF value of each concept word, and extracting the concept words with TF-IWF values ​​greater than a predetermined threshold as core concept words.

[0022] Preferably, in the method, the selected natural language processing tool is HanLP.

[0023] Preferably, the method further comprises, after step S7:

[0024] Step S8: Use the Protege tool to open the OWL file and perform visual management of the OWL ontology.

[0025] Preferably, in the method, the term dictionary is a Chinese term dictionary, and the original corpus is Chinese original corpus.

[0026] On the other hand, a device for constructing a digital battlefield knowledge graph ontology is provided, comprising a memory and a processor, wherein the memory stores at least one program, and the at least one program is executed by the processor to implement any of the methods described above.

[0027] On the other hand, a computer-readable storage medium is provided, wherein the storage medium stores at least one program, and the at least one program is executed by a processor to implement any one of the methods described above.

[0028] The above technical solution has the following technical effects:

[0029] The technical solution of the embodiment of the present invention uses natural language preprocessing tools to perform word segmentation and part-of-speech tagging on document knowledge in the digital battlefield field, such as Chinese, and uses statistical algorithms to mine core concepts. The ontology concepts and concept hierarchical relationships are extracted using a clustering algorithm, and then the ontology concepts and concept hierarchical relationships are serialized into OWL ontology files. This achieves the semi-automatic conversion of domain documents into domain ontologies with computer-processable ontology structures, thereby improving the efficiency of constructing digital battlefield domain ontologies.

[0030] In a further improvement plan, ontology modeling tools can be used to visually manage OWL ontology, which is convenient for the improvement and correction of the ontology. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A flowchart of a method for constructing a digital battlefield knowledge graph ontology according to an embodiment of the present invention;

[0032] Figure 2 This is an example of the effect of word segmentation and part-of-speech tagging on the original corpus in the method of one embodiment of the present invention;

[0033] Figure 3This is a structural diagram of an apparatus for constructing a digital battlefield knowledge graph ontology according to an embodiment of the present invention. DETAILED DESCRIPTION

[0034] To further illustrate various embodiments, the present invention is provided with accompanying drawings. These drawings form part of the present disclosure and are primarily used to illustrate the embodiments and, in conjunction with the relevant description in the specification, to explain the operating principles of the embodiments. By referring to these drawings, one of ordinary skill in the art will understand other possible embodiments and the advantages of the present invention. The components in the figures are not drawn to scale, and similar reference numerals are generally used to represent similar components.

[0035] The present invention will now be further described with reference to the accompanying drawings and specific embodiments.

[0036] Example 1:

[0037] Figure 1 This is a flow chart of a method for constructing a digital battlefield knowledge graph ontology according to an embodiment of the present invention. Figure 1 , the method of this example includes the following steps:

[0038] Step S1: Based on the selected field of the digital battlefield knowledge graph ontology, combined with the collected literature and terms related to the selected field, a term dictionary of the selected field is formed, and the term dictionary is added as a custom dictionary to the selected natural language processing tool.

[0039] The digital battlefield involves a wide range of fields, including personnel, weapons, environments, tactics, and facilities. The data involved is extensive and complex. Analysis of digital battlefield data includes the following: military personnel data, weapons and equipment data, military facility data, battlefield environment data, and military tactical data. The scope of the field and the data involved must be determined based on needs.

[0040] A terminology dictionary for the selected field is generated based on relevant terms from military literature such as the "People's Liberation Army Military Language," the "National Defense Science and Technology Thesaurus," and the "National Defense Science and Technology Terminology Dictionary." This dictionary is then added as a custom dictionary to the custom dictionary of a selected natural language processing tool, such as HanLP, to prevent incorrect segmentation of specialized terms in subsequent word segmentation.

[0041] Preferably, the embodiment of the present invention constructs the ontology for Chinese corpus; wherein the term dictionary is a Chinese term dictionary, and the original corpus is Chinese original corpus.

[0042] Step S2, preprocessing the original corpus using the selected natural language processing tool and the above-mentioned custom dictionary to obtain conceptual vocabulary related to the selected field in the original corpus, wherein the preprocessing includes word segmentation, part-of-speech tagging, and removal of interference items, which include modal particles, prepositions, quantifiers, etc. Figure 2 The present invention provides an example of the effect of word segmentation and part-of-speech tagging on a piece of original corpus. In a specific implementation, the word segmentation and part-of-speech tagging methods of the prior art can be used to pre-process the original corpus.

[0043] Step S3, using the feature term frequency-inverse word frequency TF-IWF algorithm to calculate the domain weight of each concept word in the above-mentioned concept words, and determine the core concept words in the obtained concept words according to the calculated domain weight; wherein the domain weight is represented by the TF-IWF value.

[0044] The TF-IWF (Term Frequency-Inverse Word Frequency) algorithm is a relatively mature improved algorithm based on the TF-IDF algorithm in statistical methods. The term frequency TF (TermFrequency) refers to the number of times a feature term, i.e., a concept word, appears in a document, i.e., its frequency. The inverse word frequency IWF (InverseWord Frequency) refers to the quantification of the frequency distribution of feature terms in a document set. The main idea of ​​this algorithm is: if a word or phrase has a high frequency TF in a document and rarely appears in the entire corpus, i.e., its IWF is small, then it is considered that this word or phrase has good category discrimination ability and is suitable for extracting key concept words. The calculation formula of the TF-IWF value is as follows:

[0045]

[0046] Among them TF i,j Expressing word t i The frequency of the feature term in text j, the numerator n of the TF part i,j Expressing word t i The number of times it appears in text j is the frequency, ∑ k n k,j Represents the sum of the frequencies of all words in text j; IWF i Expressing word t i The inverse frequency of a word in a corpus containing m words is the logarithm of the ratio of the total frequency of the words in the corpus to the number of times the word appears in the corpus in the query text. The numerator of the fraction in the logarithmic symbol of the IWF part is Represents the sum of the frequencies of all words in the corpus. In this example, there are m words in the corpus, nt i Expressing word ti The total frequency of occurrences in the corpus.

[0047] Specifically, in step S3, the process of extracting domain core concept vocabulary using the TF-IWF algorithm includes the following steps:

[0048] S31, input the pre-processed document set D = {d i , i=1,2,…,N}, in this example, the corpus selected for processing is the Chinese domain. The preprocessed Chinese domain document set is the Chinese domain document set D composed of the concept vocabulary di obtained in the above steps, N is the number of concept vocabulary, and N is a natural number;

[0049] S32, calculating the domain weight of each conceptual vocabulary obtained according to the TF-IWF formula described above, i.e., the TF-IWF value;

[0050] S33, traversing the TF-IWF value of each concept word, judging whether its TF-IWF value is greater than a preset threshold, if so, extracting the concept word as the core concept word of the processed original corpus.

[0051] In step S4, based on the obtained core concept vocabulary, a recursive Adaptive Resonance Theory (ART) network is used to cluster similar concepts, and similar concept vocabulary is aggregated into the same group to obtain a cluster group. Specifically, a word frequency method is used to determine a concept word in each cluster to represent the cluster group. When the word is selected as an ontology concept, it will be removed from the cluster group, and then the group will continue to be clustered at a finer level to obtain a cluster group at the next concept level. In this way, all cluster groups can be obtained. See steps S5 and S6 below for details.

[0052] Step S5, selecting a candidate word with the largest word frequency from the cluster group according to the word frequency as the ontology concept representing the class corresponding to the cluster group, and removing the selected candidate word from the cluster group;

[0053] In step S6, based on the selected finer classification threshold, the cluster group that has been removed from the candidate word is clustered again to obtain a cluster group at a finer level, and then the process goes to step S5, and steps S5-S6 are executed in a loop until the cluster cannot be further subdivided, and the hierarchical relationship of the cluster concepts is obtained.

[0054] For example, if a total of 100 concept words are obtained after analysis and preprocessing of the original corpus, and 50 of these concept words have a high degree of similarity, then they will be clustered into a group. A word frequency count is performed on these 50 concept words, and the word with the highest frequency is taken as the ontological concept representing the class; then, this word is removed from the concept group obtained by clustering, and the remaining 49 concept words are clustered based on a predetermined finer classification threshold. If the result of this re-clustering is that 20 concept words with high similarity are clustered into one group, and the other 29 concept words are clustered into another group, then the cluster group of 50 concept words contains the cluster groups of 20 and 29 concept words at the conceptual level, thereby obtaining a hierarchical relationship between the concept words.

[0055] Step S7: Serialize the obtained ontology concepts and hierarchical relationships of the concepts into a computer-processable network ontology language OWL file.

[0056] Specifically, the API provided by the Jena Semantic Web framework can be used to serialize ontology concepts and concept hierarchies in the form of Resource Description Framework (RDF) and OWL into a computer-processable OWL file, thereby achieving the construction of an OWL ontology.

[0057] After step S7, the method further includes:

[0058] Step S8: Use Protege to open the OWL file and perform visual management on the constructed OWL file, i.e., the OWL ontology. Domain experts can improve and modify the constructed OWL ontology through the visual process.

[0059] Based on the above description, it can be seen that the technical solution of the embodiment of the present invention uses natural language processing, statistical methods and clustering algorithms to semi-automatically convert domain documents into domain ontologies that can be processed by computers, effectively improving the efficiency of digital battlefield domain ontology construction.

[0060] Example 2:

[0061] The present invention also provides a device for constructing a digital battlefield knowledge graph ontology, such as Figure 3 As shown, the device includes a processor 301, a memory 302, a bus 303, and a computer program stored in the memory 302 and executable on the processor 301. The processor 301 includes one or more processing cores. The memory 302 is connected to the processor 301 via the bus 303. The memory 302 is used to store program instructions. When the processor executes the computer program, the steps in the above-mentioned method embodiment of the present invention are implemented.

[0062] Furthermore, as an executable solution, the device for constructing the digital battlefield knowledge graph ontology can be a computer unit, which can be a computing device such as a desktop computer, laptop, PDA, or cloud server. The computer unit may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above-mentioned computer unit structure is merely an example of a computer unit and does not constitute a limitation of the computer unit. The computer unit may include more or fewer components than those described above, or a combination of certain components, or different components. For example, the computer unit may also include input and output devices, network access devices, buses, etc., but this is not limited in the embodiments of the present invention.

[0063] Furthermore, as an executable solution, the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the computer unit and connects various parts of the entire computer unit using various interfaces and lines.

[0064] The memory can be used to store the computer programs and / or modules, and the processor implements the various functions of the computer unit by running or executing the computer programs and / or modules stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system and at least one application required for a function; the data storage area can store data created based on the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0065] Example 3:

[0066] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method in the embodiment of the present invention are implemented.

[0067] If the module / unit integrated in the computer unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the process in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement the steps of the above-mentioned various method embodiments. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM) and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction.

[0068] Although the present invention has been particularly shown and described in conjunction with preferred embodiments, it will be understood by those skilled in the art that various changes in form and details may be made to the present invention without departing from the spirit and scope of the invention as defined in the appended claims, and all such changes are within the scope of protection of the present invention.

Claims

1. A method for constructing a digital battlefield knowledge graph ontology, characterized in that: include: Step S1: Based on the selected field of the digital battlefield knowledge graph ontology, combined with the collected literature and terminology related to the selected field, a term dictionary of the selected field is formed, and the term dictionary is added as a custom dictionary to the selected natural language processing tool; Step S2, preprocessing the original corpus using the selected natural language processing tool and the custom dictionary to obtain concept vocabulary related to the selected field in the original corpus, wherein the preprocessing includes word segmentation, part-of-speech tagging, and removal of interference items, wherein the interference items include modal particles, prepositions, and quantifiers; Step S3, using the Feature Term Frequency-Inverse Word Frequency (TF-IWF) algorithm to calculate the domain weight of each concept word in the obtained concept words, and determining the core concept words in the obtained concept words according to the calculated domain weight; Step S4, clustering similar concepts based on the core concept vocabulary using a recursive adaptive resonance theory (ART) network to obtain a cluster group; Step S5, selecting a candidate word with the largest word frequency from the cluster group according to the word frequency as the ontology concept representing the class corresponding to the cluster group, and removing the selected candidate word from the cluster group; Step S6: clustering the cluster groups removed from the candidate words again based on the selected finer classification threshold to obtain cluster groups at a finer level, and then going back to step S5, looping through steps S5-S6 until the clusters cannot be further subdivided, thereby obtaining a hierarchical relationship of cluster concepts; Step S7: Serialize the obtained ontology concepts and hierarchical relationships of the concepts into a computer-processable network ontology language OWL file.

2. The method according to claim 1, characterized in that The step S7 comprises: The API provided by Jena semantic web framework is used to serialize ontology concepts and hierarchical relationships of concepts into computer-processable OWL files in the form of resource description framework RDF and OWL format.

3. The method according to claim 1, characterized in that In step S3, the following TF-IWF formula is used to calculate the domain weight of the concept vocabulary: Among them TF i,j Expressing word t i The frequency of feature terms in document j, n i,j Expressing word t i The frequency of occurrence in text j, ∑ k n k,j Represents the sum of the frequencies of all k words in text j, IWF i Expressing word t i The inverse frequency of a word in a corpus of m words, represents the sum of the frequencies of all m words in the corpus, nt i Expressing word t i The total frequency of occurrences in the corpus.

4. The method according to claim 3, characterized in that The step S3 comprises: Step S31, inputting a document set D = {di, i = 1, 2, ..., N} consisting of the obtained concept words di, where N is the number of concept words; Step S32, calculating the domain weight of each concept word according to the TF-IWF formula, that is, the TF-IWF value; Step S33 , traversing the TF-IWF value of each concept word, and extracting the concept words with TF-IWF values ​​greater than a predetermined threshold as core concept words.

5. The method according to claim 1, wherein The selected natural language processing tool is HanLP.

6. The method according to claim 1, characterized in that After step S7, the following steps are further included: Step S8: Use Protege to open the OWL file and perform visual management on the OWL ontology.

7. The method according to claim 1, characterized in that The term dictionary is a Chinese term dictionary, and the original corpus is Chinese original corpus.

8. A device for constructing a digital battlefield knowledge graph ontology, characterized in that: The method comprises a memory and a processor, wherein the memory stores at least one program, and the at least one program is executed by the processor to implement the method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that The storage medium stores at least one program, and the at least one program is executed by a processor to implement the method according to any one of claims 1 to 7.

Images