A defense intelligence classification management method and system based on natural language processing

The importance of key words and named entities of defense intelligence are determined through natural language processing technology, which solves the problem that existing technology cannot be classified and managed, and realizes the accurate and comprehensive classification of defense intelligence, which improves the scientificity and accuracy of management.

CN118779456BActive Publication Date: 2024-11-29XIAMEN YUANTING INFORMATION TECH CO LTD +1
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202411248698.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-06
Publication Date
2024-11-29
Estimated Expiration
2044-09-06

AI Technical Summary

Technical Problem

The prior art cannot be classified and managed based on the importance and content of defense intelligence, and lacks considerations on the importance and impact of information.

Method used

Natural language processing technology is used to determine keywords and named entities in defense intelligence, calculate their importance scores, and classify them based on this, including keyword modules, word frequency modules, importance scoring modules and named entity modules.

Benefits of technology

Improve the accuracy and comprehensiveness of defense intelligence classification management, ensure scientific and accurate assessment of the importance of intelligence, and provide a data basis for subsequent management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118779456B_ABST
    Figure CN118779456B_ABST
Patent Text Reader

Abstract

The present invention provides a defense intelligence classification management method and system based on natural language processing, which relates to the field of intelligence management technology. The method includes: determining defense intelligence keywords according to defense intelligence; determining the defense intelligence keyword frequency of the defense intelligence keywords; obtaining known important intelligence keywords and determining the importance of the known important intelligence keywords; determining the defense intelligence importance score according to the defense intelligence keywords, the defense intelligence keyword frequency, the known important intelligence keywords and the importance of the known important intelligence keywords; obtaining named entities; determining key named entity recognition results according to the named entities; and classifying and managing defense intelligence according to the defense intelligence importance score and the key named entity recognition results. According to the present invention, defense intelligence can be classified and managed according to the importance of defense intelligence and the key named entities in the defense intelligence, thereby improving the accuracy of defense intelligence classification management.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of intelligence management, and in particular to a defense intelligence classification management method and system based on natural language processing. Background Art

[0002] In related technologies, CN110135711A discloses an intelligence management method and device, which relates to the field of security technology and can calculate the value of intelligence based on relevant data of intelligence in multiple dimensions, thereby more accurately and objectively evaluating the effectiveness of intelligence. The method includes: obtaining parameters of intelligence, operational data of intelligence, and the reputation level of the user; the parameters of intelligence include at least the completeness level of intelligence, the type level of intelligence, the proportion level of the preset core part of intelligence, and the evaluation level of intelligence; the operational data of intelligence include at least the value feedback value level of intelligence, the transaction volume behavior value level of intelligence, the completeness value level of intelligence, and the assertion mean level of intelligence; generating an initial value vector based on the parameters of intelligence and the reputation level of the user; generating an evaluation value vector based on the reputation level of the user of the operational data of intelligence; and calculating the value of intelligence based on the initial value vector and the evaluation value vector. This embodiment of the solution is applied to a network system.

[0003] CN105243568A discloses a sales information classification management system and search method. The sales information collection unit transmits the collected information to corresponding sales information classification units with decreasing priorities. When searching for sales information, the sales information classification units are gradually filtered in order of priority according to the priority of the sales information being searched. Each time predetermined sales information is found, the corresponding sales information classification units are sorted in order of the number of times they have been searched. This solution provides a sales information classification management system and search method, which can achieve precise search and push based on the priority of sales information.

[0004] Based on the above-mentioned related technologies, information can be classified and managed according to its priority level. However, the related technologies do not consider the impact of the importance and content of information on the classification and management of information, that is, it is impossible to classify and manage defense intelligence according to its importance and content.

[0005] The information disclosed in the background technology section of this application is only intended to deepen the understanding of the general background technology of this application, and should not be regarded as an admission or any form of suggestion that the information constitutes the prior art already known to those skilled in the art. Summary of the Invention

[0006] The present invention provides a defense intelligence classification management method and system based on natural language processing, which can solve the technical problem that related technologies cannot classify and manage defense intelligence according to its importance and content.

[0007] According to a first aspect of the present invention, a defense intelligence classification management method based on natural language processing is provided, comprising:

[0008] Determine defense intelligence keywords based on defense intelligence;

[0009] Determining the frequency of the defense intelligence keyword appearing in the defense intelligence;

[0010] Obtain known important intelligence keywords and determine the importance of the known important intelligence keywords;

[0011] Determining a defense intelligence importance score based on the defense intelligence keyword, the defense intelligence keyword frequency, the known important intelligence keyword, and the importance of the known important intelligence keyword;

[0012] Acquire named entities of defense intelligence, wherein the named entities include: person name entities, place name entities, and organization name entities;

[0013] Determining a key named entity recognition result based on the named entity;

[0014] The defense intelligence is classified and managed according to the defense intelligence importance score and the key named entity recognition result.

[0015] According to a second aspect of the present invention, there is provided a defense intelligence classification management system based on natural language processing, comprising:

[0016] A keyword module is used to determine defense intelligence keywords based on defense intelligence;

[0017] A word frequency module, used to determine the frequency of the defense intelligence keyword appearing in the defense intelligence;

[0018] A known keyword module is used to obtain known important intelligence keywords and determine the importance of known important intelligence keywords;

[0019] an importance scoring module, configured to determine an importance score of defense intelligence based on the defense intelligence keyword, the frequency of the defense intelligence keyword, the known important intelligence keyword, and the importance of the known important intelligence keyword;

[0020] A named entity module, used to obtain named entities of defense intelligence, wherein the named entities include: person name entities, place name entities and organization name entities;

[0021] An entity recognition module, configured to determine a key named entity recognition result based on the named entity;

[0022] A classification management module is used to classify and manage defense intelligence based on the defense intelligence importance score and the key named entity recognition result.

[0023] Technical Effect: According to the present invention, defense intelligence keywords and named entities can be extracted from defense intelligence based on natural language processing technology. Furthermore, the importance of defense intelligence can be accurately analyzed based on defense intelligence keywords, and key named entities in defense intelligence can be determined. Then, based on the importance of defense intelligence and key named entities in defense intelligence, defense intelligence can be classified and managed, thereby improving the accuracy of defense intelligence classification management. When determining the stability coefficient of sample defense intelligence terms, the stability coefficient of the sample defense intelligence terms can be determined based on the word frequency of the sample defense intelligence terms. During the calculation process, the importance and stability of the sample defense intelligence terms in defense intelligence can be evaluated based on the word frequency of the sample defense intelligence terms in each paragraph of the defense intelligence, thereby improving the scientificity and accuracy of the stability coefficient of the sample defense intelligence terms. When determining the importance score of defense intelligence, the importance score of defense intelligence can be determined based on defense intelligence keywords, defense intelligence keyword frequency, known important intelligence keywords, and the importance of known important intelligence keywords. During the calculation process, the impact of multiple aspects on the importance of defense intelligence, such as the similarity between defense intelligence keywords and known important intelligence keywords, the importance of known important intelligence keywords, the frequency of defense intelligence keywords, and the number of defense intelligence keywords, is fully considered, thereby improving the comprehensiveness and accuracy of the defense intelligence importance score. When determining the key named entity recognition results, the first key named entity recognition result, the second key named entity recognition result, and the third key named entity recognition result can be determined based on the frequency of person name entities, the frequency of place name entities, the frequency of organization name entities, the location of person name entities, the location of place name entities, and the location of organization name entities. During the calculation process, the impact of the frequency of named entity appearance and the location of named entity appearance on the importance of named entities in defense intelligence is fully considered, providing a data foundation for the subsequent classification and management of defense intelligence.

[0024] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and not limiting of the present invention. Other features and aspects of the present invention will become more apparent from the following detailed description of exemplary embodiments with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can derive other embodiments based on these drawings without inventive efforts.

[0026] Figure 1 A flowchart of a defense intelligence classification management method based on natural language processing according to an embodiment of the present invention is exemplarily shown;

[0027] Figure 2 A block diagram of a defense intelligence classification management system based on natural language processing according to an embodiment of the present invention is exemplarily shown. DETAILED DESCRIPTION

[0028] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0029] The following specific embodiments are used to describe the technical solution of the present invention in detail. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0030] Figure 1 A flowchart of a defense intelligence classification management method based on natural language processing according to an embodiment of the present invention is exemplarily shown. The method includes:

[0031] Step S101, determining defense intelligence keywords based on defense intelligence;

[0032] Step S102, determining the frequency of the defense intelligence keyword in the defense intelligence;

[0033] Step S103, obtaining known important intelligence keywords and determining the importance of the known important intelligence keywords;

[0034] Step S104, determining a defense intelligence importance score based on the defense intelligence keyword, the defense intelligence keyword frequency, the known important intelligence keyword, and the importance of the known important intelligence keyword;

[0035] Step S105: Acquire named entities of defense intelligence, wherein the named entities include: person name entities, place name entities, and organization name entities;

[0036] Step S106, determining a key named entity recognition result based on the named entity;

[0037] Step S107 : Classify and manage the defense intelligence according to the defense intelligence importance score and the key named entity recognition result.

[0038] According to the defense intelligence classification management method based on natural language processing of an embodiment of the present invention, defense intelligence keywords and named entities in defense intelligence can be extracted based on natural language processing technology. Furthermore, the importance of defense intelligence can be accurately analyzed based on defense intelligence keywords, and key named entities in defense intelligence can be determined. Then, according to the importance of defense intelligence and the key named entities in defense intelligence, defense intelligence can be classified and managed, thereby improving the accuracy of defense intelligence classification management.

[0039] According to an embodiment of the present invention, in step S101, defense intelligence keywords are determined based on defense intelligence.

[0040] According to one embodiment of the present invention, step S101 includes:

[0041] Segment defense intelligence and determine all defense intelligence terms;

[0042] Get known stop words;

[0043] determining sample defense intelligence terms based on all defense intelligence terms and the known stop words;

[0044] determining the word frequency of the sample defense intelligence terms;

[0045] determining a stability coefficient of the sample defense intelligence terms according to the word frequency of the sample defense intelligence terms;

[0046] Defense intelligence keywords are determined based on the stability coefficients of the sample defense intelligence terms.

[0047] For example, split the defense intelligence into separate words or phrases to obtain all defense intelligence words for subsequent processing; obtain known stop words (e.g., "of", "is", "in"), and determine the set of known stop words; based on all defense intelligence words, determine the set of all defense intelligence words, and take the difference between the set of all defense intelligence words and the intersection of the set of all defense intelligence words and the set of known stop words as the set of sample defense intelligence words. The elements in the set of sample defense intelligence words are sample defense intelligence words, and determine the word frequency of the sample defense intelligence words appearing in each paragraph of the defense intelligence; based on the word frequency of the sample defense intelligence words, evaluate the appearance status of the sample defense intelligence words in each paragraph of the defense intelligence, determine the stability coefficient of the sample defense intelligence words, and based on the stability coefficient of the sample defense intelligence words, determine whether the sample defense intelligence word is a defense intelligence keyword.

[0048] According to an embodiment of the present invention, determining the stability coefficient of the sample defense intelligence word according to the word frequency of the sample defense intelligence word includes: determining the stability coefficient of the i-th sample defense intelligence word according to formula (1) ,

[0049] (1),

[0050] Where, is the word frequency of the i-th sample defense intelligence word appearing in the j-th paragraph of the defense intelligence, m is the number of paragraphs in the defense intelligence, j ≤ m, and both j and m are positive integers.

[0051] According to an embodiment of the present invention, is the average word frequency of the i-th sample defense intelligence word appearing in the m paragraphs of the defense intelligence, )]] is the standard deviation of the word frequency of the i-th sample defense intelligence word appearing in the m paragraphs of the defense intelligence, <( is the ratio of the standard deviation of the word frequency of the i-th sample defense intelligence word to the average word frequency, indicating the discreteness of the i-th sample defense intelligence word in the defense intelligence. The larger this ratio is, the stronger the discreteness of the i-th sample defense intelligence word in the defense intelligence. The smaller the value of, the worse the stability of the i-th sample defense intelligence word in the defense intelligence, and the smaller the possibility that the i-th sample defense intelligence word is a defense intelligence keyword. The smaller this ratio is, the larger the value of, the higher the stability of the i-th sample defense intelligence word in the defense intelligence, and the greater the possibility that the i-th sample defense intelligence word is a defense intelligence keyword. [[ID = 28]] [[ID = 29]]

[0052] In this way, the stability coefficient of the sample defense intelligence terms can be determined based on the word frequency of the sample defense intelligence terms. During the calculation process, the importance and stability of the sample defense intelligence terms in the defense intelligence can be evaluated based on the word frequency of the sample defense intelligence terms in each paragraph in the defense intelligence, thereby improving the scientificity and accuracy of the stability coefficient of the sample defense intelligence terms.

[0053] According to one embodiment of the present invention, determining defense intelligence keywords based on the stability coefficient of the sample defense intelligence terms includes:

[0054] If the stability coefficient of the sample defense intelligence term is greater than or equal to the set threshold, the sample defense intelligence term is determined to be a defense intelligence keyword;

[0055] If the stability coefficient of the sample defense intelligence term is less than the set threshold, it is determined that the sample defense intelligence term is not a defense intelligence keyword.

[0056] According to an embodiment of the present invention, in step S102, the frequency of the defense intelligence keyword in the defense intelligence is determined.

[0057] For example, through text editing software and office software, find the frequency of defense intelligence keywords appearing in various paragraphs of defense intelligence.

[0058] According to an embodiment of the present invention, in step S103, known important intelligence keywords are obtained, and the importance of the known important intelligence keywords is determined.

[0059] For example, by establishing a database of known important intelligence keywords, known important intelligence keywords are obtained, and the importance of each known important intelligence keyword assessed by experts or scholars in related fields is obtained. This importance is in the form of a score. The higher the importance of the known important intelligence keyword, the greater the score.

[0060] According to an embodiment of the present invention, in step S104, a defense intelligence importance score is determined based on the defense intelligence keyword, the defense intelligence keyword frequency, the known important intelligence keyword, and the importance of the known important intelligence keyword.

[0061] For example, defense intelligence keywords and known important intelligence keywords are input into the natural language processing model to obtain keyword semantic vectors and known keyword semantic vectors. Based on the keyword semantic vectors, known keyword semantic vectors, defense intelligence keyword frequency and the importance of known important intelligence keywords, the importance of defense intelligence is evaluated to determine the defense intelligence importance score.

[0062] According to one embodiment of the present invention, step S104 includes:

[0063] Determining a keyword semantic vector based on the defense intelligence keyword;

[0064] Determining a semantic vector of a known keyword based on the known important intelligence keyword;

[0065] Determine the defense intelligence importance score of the a-th defense intelligence according to formula (2) ,

[0066] (2),

[0067] Among them, if is a conditional function, is the preset similarity threshold, max is the maximum value function, is the frequency of the k-th defense intelligence keyword in the j-th paragraph of defense intelligence, is the importance of the rth known important intelligence keyword, is the keyword semantic vector of the k-th defense intelligence keyword, for The transposed vector of is the known keyword semantic vector of the rth known important intelligence keyword, K is the number of defense intelligence keywords, R is the number of known important intelligence keywords, m is the number of paragraphs in defense intelligence, j≤m, k≤K, r≤R, j, m, k, K, r and R are all positive integers.

[0068] According to one embodiment of the present invention, is the cosine similarity between the keyword semantic vector of the k-th defense intelligence keyword and the known keyword semantic vector of the r-th known important intelligence keyword. The larger the cosine similarity, the closer the semantics of the k-th defense intelligence keyword is to the semantics of the r-th known important intelligence keyword. The maximum value of the cosine similarity between the keyword semantic vector of the k-th defense intelligence keyword and the known keyword semantic vectors of R known important intelligence keywords is taken. The above-mentioned maximum value-taking process can determine whether the k-th defense intelligence keyword belongs to the known important intelligence keyword.

[0069] According to one embodiment of the present invention, in formula (2), the conditional function The value of includes the following two cases, when it satisfies When the condition is met, the maximum value of the cosine similarity between the keyword semantic vector of the k-th defense intelligence keyword and the known keyword semantic vector of R known important intelligence keywords is less than the preset similarity threshold, indicating that the k-th defense intelligence keyword does not belong to the known important intelligence keywords, and the value of the condition function is 0. When it is not met When the condition is met, the maximum value of the cosine similarity between the keyword semantic vector of the k-th defense intelligence keyword and the known keyword semantic vector of R known important intelligence keywords is greater than the preset similarity threshold, indicating that the k-th defense intelligence keyword belongs to the known important intelligence keyword. The value of this condition function is .

[0070] According to one embodiment of the present invention, represents the importance of the k-th defense intelligence keyword, is the importance score of the jth paragraph of the defense intelligence determined by the importance of the kth defense intelligence keyword and the defense intelligence keyword frequency of the kth defense intelligence keyword in the jth paragraph of the defense intelligence, The importance score of defense intelligence is determined based on the importance of defense intelligence keywords, which is the sum of the number of defense intelligence keywords and the number of defense intelligence paragraphs.

[0071] In this way, the defense intelligence importance score can be determined based on defense intelligence keywords, defense intelligence keyword frequency, known important intelligence keywords, and the importance of known important intelligence keywords. During the calculation process, the impact of multiple aspects such as the similarity between defense intelligence keywords and known important intelligence keywords, the importance of known important intelligence keywords, the defense intelligence keyword frequency, and the number of defense intelligence keywords on the importance of defense intelligence are fully considered, thereby improving the comprehensiveness and accuracy of the defense intelligence importance score.

[0072] According to an embodiment of the present invention, in step S105, named entities of defense intelligence are acquired, wherein the named entities include: person name entities, place name entities, and organization name entities.

[0073] For example, defense intelligence can be processed through the entity recognition module in natural language processing technology, named entities in defense intelligence can be extracted, and the person name entities, place name entities, and organization name entities in defense intelligence can be determined.

[0074] According to an embodiment of the present invention, in step S106, a key named entity recognition result is determined based on the named entity.

[0075] According to one embodiment of the present invention, step S106 includes:

[0076] Determining the frequency of occurrence of the personal name entity in defense intelligence;

[0077] Determining the frequency of occurrence of the place-name entity in defense intelligence;

[0078] Determining the frequency of occurrence of the organizational name entity in defense intelligence;

[0079] Determine the location of the named entity in defense intelligence;

[0080] Determining the location of the geographical entity in the defense intelligence;

[0081] Determining the location of the organization name entity in the defense intelligence;

[0082] A first key named entity recognition result, a second key named entity recognition result and a third key named entity recognition result are determined based on the person name entity word frequency, the place name entity word frequency, the organization name entity word frequency, the person name entity position, the place name entity position and the organization name entity position.

[0083] For example, through text editing software and office software, the word frequency of personal name entities, place name entities and organization name entities appearing in each paragraph of defense intelligence is determined respectively; the personal name entity position where the personal name entity appears in defense intelligence is determined (for example, the first personal name entity appears at the first character of defense intelligence); the place name entity position of the place name entity in defense intelligence is determined; the organization name entity position of the organization name entity in defense intelligence is determined; based on the personal name entity word frequency, place name entity word frequency, organization name entity word frequency, personal name entity position, place name entity position and organization name entity position, the first key named entity recognition result, the second key named entity recognition result and the third key named entity recognition result are determined, the first key named entity recognition result represents the most important personal name entity appearing in defense intelligence, the second key named entity recognition result represents the most important place name entity appearing in defense intelligence, and the third key named entity recognition result represents the most important organization entity appearing in defense intelligence.

[0084] According to one embodiment of the present invention, according to the person name entity word frequency, the place name entity word frequency, the organization name entity word frequency, the person name entity position, the place name entity position and the organization name entity position, determining the first key named entity recognition result, the second key named entity recognition result and the third key named entity recognition result, including: determining the first key named entity recognition result according to formula (3), formula (4) and formula (5) , the second key named entity recognition results And the third key named entity recognition results ,

[0085] (3),

[0086] (4),

[0087] (5),

[0088] in, and is the preset weight, max is the maximum value function, is the frequency of the bth person name entity, is the location of the bth person name entity at the eth occurrence in defense intelligence, is the place name entity word frequency of the c-th place name entity, is the location of the cth place name entity in the defense intelligence for the qth time it appears. is the organizational name entity frequency of the d-th organizational name entity, is the position of the yth occurrence of the dth organization name entity in the defense intelligence. is the preset key location information, B is the number of personal name entities in defense intelligence, E is the number of times the b-th personal name entity appears in defense intelligence, C is the number of place name entities in defense intelligence, Q is the number of times the c-th place name entity appears in defense intelligence, D is the number of organization name entities in defense intelligence, and Y is the number of times the d-th organization name entity appears in defense intelligence, b≤B, e≤E, c≤C, q≤Q, d≤D, y≤Y, b, B, e, E, c, C, q, Q, d, D, y and Y are all positive integers.

[0089] According to one embodiment of the present invention, in formula (3), is the average frequency of name entity words in B, It represents the relative difference between the frequency of the bth personal name entity and the average frequency of the Bth personal name entity. The larger the ratio, the more frequent the bth personal name entity appears in defense intelligence, and the higher the importance of the bth personal name entity in defense intelligence content. The relative difference between the position of the bth name entity in the defense intelligence and the preset key position information is as follows: The larger the value, the smaller the difference between the position of the bth name entity in the defense intelligence and the preset key position information, and the higher the importance of the bth name entity in the defense intelligence content. The preset key position information is usually set to the beginning or end position of the defense intelligence. When the name entity appears at the beginning or end position of the defense intelligence, it means that the name entity is more important in the defense intelligence. The average operation is performed based on the number of times the bth person name entity appears, indicating the importance of the bth person name entity in defense intelligence determined by the position of the person name entity. The weighted sum of the importance of the name entity in defense intelligence determined by the frequency of the name entity's appearance and the importance of the name entity in defense intelligence determined by the position of the name entity's appearance represents the overall importance of the bth name entity in defense intelligence. It can be used to determine the overall importance of the most important personal name entities in defense intelligence. In other words, the personal name entity with the highest overall importance can be determined as the most important personal name entity.

[0090] According to one embodiment of the present invention, in formula (4), is the average frequency of C place name entities, It represents the difference between the place name entity frequency of the c-th place name entity and the average place name entity frequency of C place name entities. The larger the ratio, the more frequent the c-th place name entity appears in defense intelligence, and the greater the importance of the c-th place name entity in defense intelligence. is the relative difference between the location of the cth place name entity appearing for the qth time in defense intelligence and the preset key location information. The smaller the ratio, The larger the value of , the smaller the difference between the location of the c-th place name entity in the defense intelligence and the preset key location information is, and the greater the importance of the c-th place name entity in the defense intelligence. The preset key location information is usually set to the beginning or end position of the defense intelligence. When the place name entity appears at the beginning or end position of the defense intelligence, it means that the place name entity is more important in the defense intelligence. The average operation is performed based on the number of times the c-th place name entity appears, indicating the importance of the c-th place name entity in defense intelligence determined by the location where the place name entity appears. The weighted sum of the importance of the place-name entity in defense intelligence determined by the frequency of its occurrence and the importance of the place-name entity in defense intelligence determined by the location of its occurrence represents the integrity of the cth place-name entity in defense intelligence. It can be used to determine the overall importance of the most important geographical entities in defense intelligence. In other words, the geographical entity with the highest overall importance can be determined as the most important geographical entity.

[0091] According to one embodiment of the present invention, in formula (5), is the average frequency of D organizational name entities, It represents the difference between the organizational name entity frequency of the d-th organizational name entity and the average organizational name entity frequency of D organizational name entities. The larger the ratio, the more frequent the d-th organizational name entity appears in defense intelligence, and the higher the importance of the d-th organizational name entity in defense intelligence. The relative difference between the position of the dth organization name entity in the defense intelligence and the preset key position information is as follows: The larger the value of , the smaller the difference between the yth appearance of the d-th organization name entity in the defense intelligence and the preset key position information, and the higher the importance of the d-th organization name entity in the defense intelligence. The preset key position information is usually set to the beginning or end position of the defense intelligence. When the organization name entity appears at the beginning or end position of the defense intelligence, it means that the organization name entity is relatively important in the defense intelligence. The average operation is performed based on the number of times the organization name entity appears, indicating the importance of the organization name entity in defense intelligence determined by the position of the organization name entity. The weighted sum of the importance of the organization name entity in defense intelligence determined by the frequency of its appearance and the importance of the organization name entity in defense intelligence determined by the position of its appearance represents the overall importance of the dth organization name entity in defense intelligence. It can be used to determine the overall importance of the most important organization name entity in defense intelligence. In other words, the organization name entity with the highest overall importance can be determined as the most important organization name entity.

[0092] In this way, the first key named entity recognition result, the second key named entity recognition result and the third key named entity recognition result can be determined based on the frequency of personal name entity words, the frequency of place name entity words, the frequency of organizational name entity words, the position of personal name entity, the position of place name entity and the position of organizational name entity. During the calculation process, the influence of the frequency of occurrence of named entities and the position of occurrence of named entities on the importance of named entities in defense intelligence is fully considered, providing a data basis for the subsequent classification management of defense intelligence.

[0093] According to an embodiment of the present invention, in step S107, the defense intelligence is classified and managed according to the defense intelligence importance score and the key named entity recognition result.

[0094] According to one embodiment of the present invention, step S107 includes:

[0095] Determining, based on the first key named entity recognition result, a person name entity corresponding to the first key named entity recognition result;

[0096] Determining, based on the second key named entity recognition result, a place name entity corresponding to the second key named entity recognition result;

[0097] Determining, based on the third key named entity recognition result, an organization name entity corresponding to the third key named entity recognition result;

[0098] Defense intelligence is classified and managed according to the person name entity corresponding to the first key named entity recognition result, the place name entity corresponding to the second key named entity recognition result, the organization name entity corresponding to the third key named entity recognition result, and the defense intelligence importance score.

[0099] For example, based on the first key named entity recognition result, the second key named entity recognition result and the third key named entity recognition result of the a-th defense intelligence, determine the important person name entity, important place name entity and important organization name entity corresponding to the a-th defense intelligence, and classify the a-th defense intelligence into the categories of the above-mentioned important person name entity, important place name entity and important organization name entity. Furthermore, multiple defense intelligence in each named entity category are arranged in descending order according to the defense intelligence importance score of each defense intelligence, and each defense intelligence is hierarchically managed according to the determined descending sequence.

[0100] According to the defense intelligence classification management method based on natural language processing of an embodiment of the present invention, defense intelligence keywords and named entities in defense intelligence can be extracted based on natural language processing technology. Furthermore, the importance of defense intelligence can be accurately analyzed based on defense intelligence keywords, and key named entities in defense intelligence can be determined. Then, based on the importance of defense intelligence and key named entities in defense intelligence, defense intelligence can be classified and managed, thereby improving the accuracy of defense intelligence classification management. When determining the stability coefficient of sample defense intelligence terms, the stability coefficient of the sample defense intelligence terms can be determined based on the word frequency of the sample defense intelligence terms. During the calculation process, the importance and stability of the sample defense intelligence terms in defense intelligence can be evaluated based on the word frequency of the sample defense intelligence terms in each paragraph of the defense intelligence, thereby improving the scientificity and accuracy of the stability coefficient of the sample defense intelligence terms. When determining the importance score of defense intelligence, the importance score of defense intelligence can be determined based on defense intelligence keywords, defense intelligence keyword frequency, known important intelligence keywords, and the importance of known important intelligence keywords. During the calculation process, the impact of multiple aspects on the importance of defense intelligence, such as the similarity between defense intelligence keywords and known important intelligence keywords, the importance of known important intelligence keywords, the frequency of defense intelligence keywords, and the number of defense intelligence keywords, is fully considered, thereby improving the comprehensiveness and accuracy of the defense intelligence importance score. When determining the key named entity recognition results, the first key named entity recognition result, the second key named entity recognition result, and the third key named entity recognition result can be determined based on the frequency of person name entities, the frequency of place name entities, the frequency of organization name entities, the location of person name entities, the location of place name entities, and the location of organization name entities. During the calculation process, the impact of the frequency of named entity appearance and the location of named entity appearance on the importance of named entities in defense intelligence is fully considered, providing a data foundation for the subsequent classification and management of defense intelligence.

[0101] Figure 2 A block diagram of a defense intelligence classification management system based on natural language processing according to an embodiment of the present invention is exemplarily shown. The system includes:

[0102] A keyword module is used to determine defense intelligence keywords based on defense intelligence;

[0103] A word frequency module, used to determine the frequency of the defense intelligence keyword appearing in the defense intelligence;

[0104] A known keyword module is used to obtain known important intelligence keywords and determine the importance of known important intelligence keywords;

[0105] an importance scoring module, configured to determine an importance score of defense intelligence based on the defense intelligence keyword, the frequency of the defense intelligence keyword, the known important intelligence keyword, and the importance of the known important intelligence keyword;

[0106] A named entity module, used to obtain named entities of defense intelligence, wherein the named entities include: person name entities, place name entities and organization name entities;

[0107] An entity recognition module, configured to determine a key named entity recognition result based on the named entity;

[0108] A classification management module is used to classify and manage defense intelligence based on the defense intelligence importance score and the key named entity recognition result.

[0109] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0110] Those skilled in the art will appreciate that the embodiments of the present invention described above and shown in the accompanying drawings are intended to be illustrative only and are not intended to limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functional and structural principles of the present invention have been demonstrated and illustrated in the embodiments. Any variations or modifications may be made to the embodiments of the present invention without departing from the principles described.

Claims

1. A defense intelligence classification management method based on natural language processing, characterized in that: include: Determine defense intelligence keywords based on defense intelligence; Determining the frequency of the defense intelligence keyword appearing in the defense intelligence; Obtain known important intelligence keywords and determine the importance of known important intelligence keywords; Determining a defense intelligence importance score according to the defense intelligence keyword, the defense intelligence keyword frequency, the known important intelligence keyword, and the importance of the known important intelligence keyword; Acquire named entities of defense intelligence, wherein the named entities include: person name entities, place name entities, and organization name entities; Determine a key named entity recognition result according to the named entity; Classify and manage the defense intelligence according to the defense intelligence importance score and the key named entity recognition result; Based on defense intelligence, determine defense intelligence keywords, including: Segment the defense intelligence and determine all defense intelligence terms; Get known stop words; Determining sample defense intelligence terms based on all defense intelligence terms and the known stop words; Determining the word frequency of the sample defense intelligence terms; Determining the stability coefficient of the sample defense intelligence terms according to the word frequency of the sample defense intelligence terms; Determining defense intelligence keywords according to the stability coefficient of the sample defense intelligence words; Determining a key named entity recognition result according to the named entity includes: Determining the frequency of occurrence of the personal name entity in defense intelligence; Determining the frequency of occurrence of the place name entity in defense intelligence; Determining the frequency of occurrence of the organizational name entity in defense intelligence; Determine the location of the named entity in the defense intelligence; Determine the location of the geographical entity in the defense intelligence; Determine the location of the organization name entity in the defense intelligence; Determine a first key named entity recognition result, a second key named entity recognition result, and a third key named entity recognition result according to the person name entity word frequency, the place name entity word frequency, the organization name entity word frequency, the person name entity position, the place name entity position, and the organization name entity position; Determining a first key named entity recognition result, a second key named entity recognition result, and a third key named entity recognition result according to the person name entity word frequency, the place name entity word frequency, the organization name entity word frequency, the person name entity position, the place name entity position, and the organization name entity position, including: According to the formula , , , Determine the first key named entity recognition result , the second key named entity recognition results And the third key named entity recognition results ,in, and is the preset weight, max is the maximum value function, is the frequency of the bth person name entity, is the position of the e-th occurrence of the b-th person name entity in defense intelligence, is the place name entity frequency of the cth place name entity, is the location of the cth place name entity in the defense intelligence for the qth time it appears. is the organizational name entity frequency of the d-th organizational name entity, is the position of the yth occurrence of the dth organization name entity in the defense intelligence, is the preset key location information, B is the number of personal name entities in defense intelligence, E is the number of times the b-th personal name entity appears in defense intelligence, C is the number of place name entities in defense intelligence, Q is the number of times the c-th place name entity appears in defense intelligence, D is the number of organization name entities in defense intelligence, Y is the number of times the d-th organization name entity appears in defense intelligence, b≤B, e≤E, c≤C, q≤Q, d≤D, y≤Y, b, B, e, E, c, C, q, Q, d, D, y and Y are all positive integers.

2. The defense intelligence classification management method based on natural language processing according to claim 1 is characterized in that: Determining the stability coefficient of the sample defense intelligence terms according to the word frequency of the sample defense intelligence terms includes: According to the formula , Determine the stability coefficient of the i-th sample defense intelligence term ,in, is the frequency of the i-th sample defense intelligence word appearing in the j-th paragraph of the defense intelligence, m is the number of paragraphs in the defense intelligence, j≤m, j and m are both positive integers.

3. The defense intelligence classification management method based on natural language processing according to claim 2 is characterized in that: According to the stability coefficient of the sample defense intelligence words, defense intelligence keywords are determined, including: If the stability coefficient of the sample defense intelligence term is greater than or equal to the set threshold, the sample defense intelligence term is determined to be a defense intelligence keyword; If the stability coefficient of the sample defense intelligence term is less than the set threshold, it is determined that the sample defense intelligence term is not a defense intelligence keyword.

4. The defense intelligence classification management method based on natural language processing according to claim 1 is characterized in that: Determining the importance score of the defense intelligence according to the defense intelligence keyword, the frequency of the defense intelligence keyword, the known important intelligence keyword, and the importance of the known important intelligence keyword includes: Determining a keyword semantic vector according to the defense intelligence keyword; Determine a known keyword semantic vector based on the known important intelligence keyword; According to the formula , Determine the defense intelligence importance score of the a-th defense intelligence , where if is a conditional function, is the preset similarity threshold, max is the maximum value function, is the frequency of the k-th defense intelligence keyword in the j-th paragraph of defense intelligence, is the importance of the rth known important intelligence keyword, is the keyword semantic vector of the k-th defense intelligence keyword, for The transposed vector of is the known keyword semantic vector of the rth known important intelligence keyword, K is the number of defense intelligence keywords, R is the number of known important intelligence keywords, m is the number of paragraphs in the defense intelligence, j≤m, k≤K, r≤R, j, m, k, K, r and R are all positive integers.

5. The defense intelligence classification management method based on natural language processing according to claim 1 is characterized in that: According to the importance score of the defense intelligence and the key named entity recognition result, the defense intelligence is classified and managed, including: Determine, according to the first key named entity recognition result, a person name entity corresponding to the first key named entity recognition result; Determine, according to the second key named entity recognition result, a place name entity corresponding to the second key named entity recognition result; According to the third key named entity recognition result, determining the organization name entity corresponding to the third key named entity recognition result; Defense intelligence is classified and managed according to the person name entity corresponding to the first key named entity recognition result, the place name entity corresponding to the second key named entity recognition result, the organization name entity corresponding to the third key named entity recognition result and the defense intelligence importance score.

6. A defense intelligence classification management system based on natural language processing for executing the method according to any one of claims 1 to 5, characterized in that: include: A keyword module is used to determine defense intelligence keywords based on defense intelligence; A word frequency module, used to determine the word frequency of the defense intelligence keyword appearing in the defense intelligence; A known keyword module is used to obtain known important intelligence keywords and determine the importance of known important intelligence keywords; An importance scoring module, used to determine the importance score of the defense intelligence according to the defense intelligence keyword, the frequency of the defense intelligence keyword, the known important intelligence keyword and the importance of the known important intelligence keyword; A named entity module, used to obtain named entities of defense intelligence, wherein the named entities include: person name entities, place name entities and organization name entities; An entity recognition module, used to determine a key named entity recognition result based on the named entity; A classification management module is used to classify and manage defense intelligence according to the defense intelligence importance score and the key named entity recognition result.

Citation Information

Patent Citations

  • Sales information classification management system and searching method thereof

    CN105243568A

  • Threat intelligence theme detection method and device and computer storage medium

    CN111581355A