Analysis and recommendation method and system for strategic game aided decision-making
Through multi-grained text semantic analysis and multi-source intelligence data association for game intelligence, combined with collaborative filtering recommendation and intelligent intelligence analysis, the complexity of decision-making in strategic games is solved, semantic understanding and intelligence recommendation are realized, and strategic deduction is improved. efficiency and accuracy.
Patent Information
- Application Number
- CN202310692679.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-12
- Publication Date
- 2024-12-06
AI Technical Summary
It is difficult for the prior art to quickly analyze and provide effective decision-making recommendations to support efficient decision-making in strategic games, especially in complex information and social domain simulations, which lack effective semantic understanding and intelligence recommendation capabilities.
Adopt multi-grained text semantic analysis and understanding, multi-source intelligence data association and recommendation, through multi-grained text semantic analysis technology and collaborative filtering recommendation, combined with intelligent intelligence and decision-making analysis modules, realize semantic understanding and intelligence recommendation of strategic situations. and situation analysis.
It realizes semantic understanding of unstructured text data, provides targeted intelligence recommendations and situation analysis, enhances the objectivity and quantification of strategic deduction, and meets the decision makers' needs for intelligent assistance.
Smart Images

Figure CN119089906A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of intelligent assistance, and specifically relates to an analysis and recommendation method and system for strategic game auxiliary decision-making. Background Art
[0002] At present, developed countries in the world attach great importance to strategic-level exercises and research demonstrations, and continue to introduce advanced technologies and deduction methods, such as big data, cloud computing, artificial intelligence, simulation, war game systems, etc., to promote strategic research with a combination of qualitative and quantitative confrontation deduction models. In addition, the emergence of ChatGPT, with its powerful semantic understanding ability, analytical judgment ability, and behavioral creation ability, provides strong support for the study of strategic issues.
[0003] New exercises require the support of new strategic training simulation systems, which greatly promotes the development of new strategic simulation models. These new simulation models expand the simulation objects and fields from the traditional combat physical domain to the information domain, cognitive domain and social domain, and also promote the innovation of simulation modeling methods. Guided by complexity science, new modeling ideas such as complex networks, CAS, artificial life and swarm intelligence have been applied and developed. Summary of the invention
[0004] The purpose of the present invention is to provide an analysis and recommendation method and system for strategic game decision-making assistance, which can quickly analyze the current strategic situation and provide decision makers with effective decision-making suggestions under the current situation, so as to formulate strategic measures more efficiently.
[0005] The technical solution adopted by the present invention is:
[0006] An analysis and recommendation method for strategic game decision-making assistance mainly includes three steps: multi-granular text semantic analysis and understanding for game intelligence, multi-source intelligence data association and recommendation, and countermeasure action auxiliary analysis and judgment; among which:
[0007] Multi-granularity text semantic analysis and understanding for game intelligence: semantic analysis and understanding of intelligence data, and semantic analysis and understanding results are obtained by semantic mining of texts of different granularities;
[0008] Multi-source intelligence data association and recommendation: Analyze and associate multi-source intelligence data, and make intelligent recommendations based on the results of semantic analysis and understanding;
[0009] Countermeasure action auxiliary analysis and judgment: Based on semantic analysis and understanding, intelligence association and recommendation, the advantages and disadvantages of the current game situation are judged to assist the game decision-making process.
[0010] Furthermore, the semantic mining of texts of different granularities utilizes multi-granularity text semantic analysis technology to perform semantic mining at three levels: words, sentences, and paragraphs, complete shallow semantic analysis and deep semantic analysis of the text, and construct a multi-granularity text semantic understanding framework for intelligence analysis.
[0011] Furthermore, the multi-source intelligence data includes: intelligence data in the strategic intelligence library, historical situation data in the deduction situation library, and historical decision information in the decision information library.
[0012] Furthermore, the intelligent recommendation of intelligence includes two forms: offline processing and online processing, and collaborative filtering recommendation is performed on the intelligence decision data; the collaborative filtering recommendation filters data resources at the deduction situation level from the perspective of the strategic deduction decision maker, so as to meet the decision-making needs of different decision makers at multiple levels, and uses the similarity between the historical decisions of the decision makers as the basis for judgment. If the pre-decision attributes and action methods are similar to the decision maker's behavior and are in line with the current strategic situation, they can be regarded as high-value recommended decisions or intelligence.
[0013] Furthermore, the collaborative filtering recommendation includes the following steps:
[0014] Collect user interest information, mark and score the decision or intelligence information retrieved by the user. The more times the user searches, the higher the evaluation score.
[0015] Perform nearest neighbor search based on users to find users with similar interests and hobbies to the user and calculate the similarity between users;
[0016] Based on the calculated similarity, different forms of recommendations are made according to different recommendation purposes, including Top-N recommendations and relationship recommendations.
[0017] Furthermore, the determination of the pros and cons of the current gaming situation to assist the gaming decision-making process includes:
[0018] It provides auxiliary calculation of equipment capabilities and auxiliary analysis of battlefield situations, performs constraint calculations on equipment speed, maneuverability, strike range, and reconnaissance range, and performs GIS-based visualization to assist in decision-making and enhance the objectivity and quantification of strategic deductions.
[0019] Furthermore, the determination of the pros and cons of the current gaming situation to assist the gaming decision-making process also includes:
[0020] Two modules are designed, namely the intelligent intelligence and decision-making analysis module and the intelligent intelligence service module; the intelligent intelligence service module is applied to the data layer and the processing layer. By managing the game situation database, the strategic intelligence database and the decision-making information database, the module accesses multi-source heterogeneous data such as the current strategic situation, game progress, the situation of the participating parties, and the strategic intentions of all parties. Through the game network environment, multi-machine collaboration and load uniformity under the multi-party game mechanism are realized; the intelligent intelligence and decision-making analysis module is located in the logical application layer, which can complete the whole process of data format conversion, data cleaning, mining and analysis, visualization display, report organization, writing and output, and realize the collaborative work among multiple functional components; the intelligent intelligence service module and the intelligent intelligence and decision-making analysis module complete the data transmission and conversion through the system integration interface, so as to analyze and judge the decisions and situations of all parties in the strategic game process.
[0021] An analysis and recommendation system for strategic game decision-making assistance, comprising:
[0022] The multi-granularity text semantic parsing and understanding module for game intelligence is used to perform semantic parsing and understanding on intelligence data. The semantic parsing and understanding results are obtained by performing semantic mining on texts of different granularities.
[0023] Multi-source intelligence data association and recommendation module, used to parse and associate multi-source intelligence data, and make intelligent recommendations based on the results of semantic analysis and understanding;
[0024] The countermeasure action auxiliary analysis and judgment module is used to judge the advantages and disadvantages of the current game situation based on semantic analysis and understanding, intelligence association and recommendation, so as to assist the game decision-making process.
[0025] As mentioned above, an analysis and recommendation method and system for strategic game decision support realizes the understanding and mining of unstructured semantics of strategic situations, intelligent recommendation of decisions and intelligence, and assessment of the advantages and disadvantages of action situations, etc., meeting the decision makers' needs for intelligent assistance in solving strategic problems.
[0026] The beneficial effects of the present invention are:
[0027] (1) The present invention realizes the semantic understanding of unstructured text data by machines.
[0028] (2) The present invention can analyze the current situation during the strategic deduction process, sort out intelligence information in a targeted manner, and make recommendations on intelligence and decision-making.
[0029] (3) The present invention can determine the impact of new situations on ourselves or others during the simulation process. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 Analysis and recommendation framework for strategic game decision support (module diagram);
[0031] Figure 2 A schematic diagram of multi-granular text semantic understanding for game intelligence;
[0032] Figure 3 Schematic diagram for neurons to perform multi-granular text semantic understanding;
[0033] Figure 4 It is a schematic diagram for the association and recommendation of multi-source intelligence data;
[0034] Figure 5 Schematic diagram to assist analysis and judgment of countermeasure actions. DETAILED DESCRIPTION
[0035] The following introduces an analysis and recommendation method for strategic game decision-making assistance provided by the present invention in conjunction with the accompanying drawings and embodiments.
[0036] like Figure 1 As shown in the figure, an analysis and recommendation method for strategic game decision-making assistance is proposed, which mainly includes three steps: multi-granular text semantic analysis and understanding for game intelligence, multi-source intelligence data association and recommendation, and countermeasure action auxiliary analysis and judgment. In the multi-granular text semantic analysis and understanding step for game intelligence, a text processing, shallow semantic analysis, deep semantic analysis, and semantic understanding framework are established to perform semantic mining on texts of different granularities to obtain semantic analysis and understanding results. In the multi-source intelligence data association and recommendation step, by analyzing and associating the intelligence data in the strategic intelligence library, the historical situation data in the deduction situation library, and the historical decision information in the decision information library, combined with the semantic analysis and understanding results, the data is analyzed to realize the intelligent recommendation function of intelligence. Countermeasure action auxiliary analysis and judgment is to judge the advantages and disadvantages of the current game situation on the basis of semantic analysis and understanding, intelligence analysis, and data integration, so as to assist the game decision-making process.
[0037] 1. Multi-granular text semantic understanding for game intelligence
[0038] like Figure 2 As shown in the figure, it is a schematic diagram of the steps of multi-granular text semantic understanding for game intelligence. Figure 3The figure shows the principle diagram of multi-granular text semantic understanding based on neurons. In order to solve the obstacles to data compilation caused by mixed languages, non-standard languages, and diverse data forms in open source data, a text processing, shallow semantic analysis, deep semantic analysis, and semantic understanding technology framework are established by using intelligence semantic understanding and element extraction technology based on transfer learning. Constructing a processing, analysis, and semantic understanding framework for multi-source heterogeneous open source intelligence data is the premise and foundation for realizing the extraction of intelligence element information. Using multi-granular text semantic analysis technology, semantic mining is carried out at multiple levels such as words, sentences, and chapters to complete shallow to deep semantic analysis of texts and build a multi-granular text semantic understanding framework for intelligence analysis.
[0039] Among them, shallow semantic analysis is to complete word-granularity semantic analysis and sentence-granularity semantic analysis, including part-of-speech tagging, word sense disambiguation, and semantic role tagging; deep semantic analysis is paragraph-granularity semantic analysis, which combines the semantic information that has been given to the text at word-granularity and sentence-granularity to conduct overall text semantic mining and understand the text theme from a macro perspective.
[0040] 1) Word granularity semantic analysis technology
[0041] Semantic mining at word granularity is the basic work of text semantic mining. The processing unit of this granularity is words, which is not restricted by data structure and format. As long as it is text information, semantic mining can be performed at this granularity. The main semantic mining methods at word granularity include part-of-speech tagging and word sense disambiguation. Through semantic processing at word granularity, the first layer of semantic information can be added to text data, and the foundation for text semantic mining at the sentence and paragraph levels can be laid.
[0042] Word sense disambiguation is to identify the specific meaning of a word in a specific sentence when the word has multiple meanings. There are two methods: knowledge base-based and corpus-based. In the corpus-based word sense disambiguation method, it can be divided into supervised disambiguation and unsupervised disambiguation according to whether there is human intervention. Word sense disambiguation for extracting intelligence elements from open source data combines vector space models and clustering models to achieve intelligence word meaning coreference disambiguation and solve problems such as sparse and high-dimensional multi-source heterogeneous data.
[0043] 2) Sentence Granularity Semantic Analysis
[0044] Semantic analysis at sentence granularity is mainly semantic role labeling. This method is applicable to any information with syntactic structure. Because it has both semantic information and syntactic structure information at word granularity, it is more complete than text semantic mining at word granularity and is a key technology for text semantic mining.
[0045] Semantic role labeling is a shallow semantic analysis at the sentence level, labeling certain phrases in the sentence as arguments (semantic roles) of a given predicate, such as agent, patient, time and place, etc.
[0046] The main semantic role labeling methods are: semantic role labeling based on syntactic analysis, semantic role labeling based on feature vectors, and semantic role labeling based on machine learning. Semantic role labeling based on syntactic analysis is further divided into methods based on phrase structure, chunk analysis, and dependency syntax. In view of the complex data sources and high text noise of open source text, the traditional rule-based semantic role labeling method cannot solve the problem of non-standard social text semantic role labeling. Therefore, a semantic role labeling method based on a combination of dependency trees and machine learning algorithms is used to train semantic role labelers for specific fields and specific text types, laying the foundation for intelligence semantic element extraction and topic analysis.
[0047] 3) Chapter-level semantic analysis
[0048] Text semantic analysis at the chapter level overcomes the problem that traditional intelligence research methods are limited by formatted information, such as the problem of bibliographic information analysis, and can go deep into the text content level, combine the semantic information that has been given to the text at the word and sentence levels to conduct overall text semantic mining, and understand the text theme from a macro perspective. Due to the diverse sources and complex structures of open source data, the chapter text is represented by a text semantic space vector using the chapter semantic analysis technology, and the LDA topic model is used to obtain the text theme semantic information, which is convenient for subsequent text understanding and intelligence summary extraction.
[0049] Figure 2 Information extraction in the context of machine learning refers to a basic natural language processing (NLP) task that converts unstructured or semi-structured natural language text into structured features, mainly including term extraction, entity recognition, relationship combing, sentiment analysis, etc. Figure 2 The execution task in the process refers to completing the semantic analysis of natural speech text through information extraction, and then completing tasks such as semantic question answering, knowledge classification, automatic summarization, and semantic translation.
[0050] 2. Multi-source intelligence data association and recommendation
[0051] like Figure 4 As shown in Figure 2, multi-source intelligence data association and recommendation include two forms: offline processing and online processing, namely Figure 4"Offline recommendation service" and "Real-time data recommendation service" in. Offline processing builds a feature platform by processing data, provides raw materials for model training, and does some necessary data statistics; online processing provides raw materials for real-time recommendation and obtains recommendation results. In order to realize collaborative filtering recommendation of intelligence decision-making data, a feature platform is built based on the decision-making behavior, historical situation, and deduction situation of strategic deduction decision-makers, and user portraits are built for different strategic deduction decision-makers. The preference for decision types made by them in the face of different situations is statistically analyzed to complete the construction of a model for user characteristics, so as to recommend decision content that meets the current situation and conforms to the user's strategic direction.
[0052] Collaborative filtering recommendation filters data resources at the level of deduction situation from the perspective of strategic deduction decision-makers, thereby meeting the decision-making needs of different decision-makers at multiple levels. The similarity between the historical decisions of the decision-makers is used as the basis for judgment. If the pre-decision attributes and action methods are similar to the decision-makers' behavior and are in line with the current strategic situation, they can be regarded as high-value recommended decisions or intelligence. Collaborative filtering recommendation mainly includes three steps: collecting user information, nearest neighbor search, and generating recommendation results.
[0053] 1) Collect user information: Collect user interest information, mark and score the decision or intelligence information retrieved by the user. The more times the user searches, the higher the evaluation score.
[0054] 2) Nearest neighbor search: Find users with similar interests and hobbies based on users, that is, calculate the similarity between two users. For example: find n users with similar interests to A, and use their ratings of M as the prediction of A's rating of M; with the nearest neighbor set, the interests of the target user can be predicted.
[0055] 3) Generate recommendation results: According to the calculated similarity, different forms of recommendations are made according to different recommendation purposes. There are mainly Top-N recommendations and relationship recommendations. Top-N recommendations are generated for individual users and produce different results for each person. For example, by counting the nearest neighbor users of user A, those with high frequency of appearance and not present in the rating information of user A are selected as recommendation results. Relationship recommendations are to mine the association rules of the records of the nearest neighbor users.
[0056] Figure 4In the "scheduling service", "scheduling service" refers to the background data processing to build a feature platform for extracting data attribute characteristics; "statistical service" refers to the statistical probability of various decisions after the appearance of certain feature data; "data collection" refers to the collection of open source historical decision data and other data; "data buffering" refers to caching the calculated recommended data to accelerate system response; "relational database" is used to store open source decision data, user historical decision data and data characteristics and other data; "buffer database" is used to store intelligence decision data recommended for users.
[0057] In strategic deduction, this method can provide a rich data environment for participating experts and play a role in assisting decision-making. To this end, on the one hand, it is necessary to associate the assumed background and crisis situation of the exercise task with the constructed virtual game data environment; on the other hand, it is necessary to connect multi-source intelligence data, establish the association between real intelligence data and the virtual deduction background, realize the combination of virtual and real, enhance the availability, authenticity and richness of the recommended intelligence or decision, and provide greater guidance for solving strategic problems. These can be achieved using existing technologies.
[0058] To this end, in the construction of a system using this method, it is necessary to enrich the data environment support in the deduction process by connecting intelligence data from multiple sources such as message intelligence data, open source intelligence data, national conditions and military data. At the same time, semantic parsing, knowledge graphs, association queries, data visualization and other technologies (which can be implemented using existing technologies) are used to achieve the fusion and association between multi-source intelligence data and between multi-source intelligence data and deduction data.
[0059] 3. Countermeasures and actions to assist analysis and judgment
[0060] The processing flow of the countermeasure action auxiliary analysis and judgment steps is as follows Figure 5 As shown. In the strategic game simulation, it is crucial to analyze and judge the situation of the simulation process and the decisions of all parties. For decision-making measures with a high degree of structure and quantification, especially military and economic measures, certain rules and constraints can ensure the objectivity and rationality of the analysis of decision-making measures. For example, military operations need to consider factors such as the capabilities of equipment and troops and the time constraints of operations, while economic activities need to consider economic laws and the current economic strength of the country.
[0061] Based on the above analysis and judgment requirements, this method provides visualization tools for auxiliary calculation, capability analysis and other analysis and judgment, manages the equipment capability templates of the main stakeholders in the game in advance, presets their performance parameters, and provides functions for auxiliary calculation of equipment capabilities and auxiliary analysis of battlefield situations. It performs constraint calculations on the speed, mobility, strike range, reconnaissance range, etc. of the equipment, and performs visualization based on GIS, which plays an auxiliary decision-making role and enhances the objectivity and quantification of strategic games.
[0062] 3.1) Equipment capability auxiliary calculation
[0063] The equipment capability auxiliary calculation module mainly provides intelligent auxiliary analysis of equipment capabilities for all parties in the battlefield situation analysis process based on equipment attribute data and basic equipment data, so as to improve the mastery and utilization of equipment capabilities by all parties in the strategic deduction process. The system provides a variety of auxiliary calculation models, including measurement calculation, speed calculation, maneuverability calculation, strike range calculation, etc.
[0064] (1) Measurement and calculation. The basic information of the equipment is measured and calculated, including length, width, radius, area, floor space, etc.
[0065] (2) Speed calculation: The system calculates the estimated speed of the equipment based on the actual movement data of the equipment on the battlefield.
[0066] (3) Mobility calculation: The system calculates the specific mobility of the equipment per unit time / demonstration time based on the equipment performance parameters input by all parties during use, combined with the basic equipment database and battlefield situation.
[0067] (4) Calculation of strike range. When the battlefield situation contains equipment with strike range attributes, the system can analyze and calculate the strike range of the current equipment based on the basic equipment data and the equipment attributes set by each party.
[0068] 3.2) Auxiliary analysis of battlefield situation
[0069] The battlefield situation auxiliary analysis module is mainly used during the strategic deduction process. The director and participants can assist in the prediction and dynamic control of the rationality of the equipment settings, strike capability, and maneuver range in the current situation based on the basic equipment database data, equipment performance parameters, and battlefield situation data. All parties can make a reasonable calculation and analysis of the current battlefield situation based on the results of the auxiliary prediction.
[0070] For difficult-to-structure content such as politics and public opinion, there is a need for unstructured decision-making action analysis in the strategic game process. This method designs intelligent analysis of decision measures. For the decision measures provided by all parties, it provides functions such as intelligent analysis, model calculation, and simulation plotting output. It can convert specific categories of decision measures provided by all parties into plotting situation maps integrated with the digital earth. Specifically, it includes four functional modules: structured processing of measures, manual / automatic auxiliary refinement, model simulation, and auxiliary plotting.
[0071] (1) Structural processing of measures. The structural processing module of measures matches the text measures input by all parties with the basic equipment database and measure library according to the parsing template, automatically extracts the structured data therein, and processes the extracted structured data into equipment labels and measure labels for subsequent use.
[0072] (2) Manual / automatic assisted refinement. The manual / automatic assisted refinement module refines the various structural data labeled in the previous step with human participation, and concretizes the actual measures and geographical locations corresponding to the structural data, and establishes preliminary simulation mapping association data.
[0073] (3) Model simulation. The model simulation module calculates the simulation model based on the simulation mapping related data after auxiliary refinement, combined with the basic database, geographic information database, and various simulation models, and transmits the calculated simulation model to the situation mapping subsystem.
[0074] (4) Auxiliary mapping. The auxiliary mapping module mainly provides the real mapping function of the simulation mapping model. According to the various mapping attributes, geographic information attributes, and measure information contained in the simulation model, the system selects the appropriate geographical location, perspective, mapping symbols, and measures in the geographic situation mapping module for situation mapping.
[0075] According to the above requirements, two modules are designed, namely the intelligent intelligence and decision analysis module and the intelligent intelligence service module. The strategic decision and intelligence intelligent analysis architecture covers three levels: data layer, processing layer and logical application layer. The intelligent intelligence service module is applied to the data layer and processing layer. Through the management of the deduction situation database, strategic intelligence database and decision information database, this module accesses multi-source heterogeneous data such as the current strategic situation, deduction process, situation of the participating parties, and strategic intentions of all parties. Through the deduction network environment, it realizes multi-machine collaboration and load uniformity under the multi-party deduction mechanism; the intelligent intelligence and decision analysis module is mainly the logical application layer of the system, which consists of 5 functional components. It can complete the whole process of intelligence analysis and research such as data format conversion, data cleaning, mining analysis, visualization display, report organization writing and output, and realize the collaborative work between multiple functional components. The intelligent intelligence service module and the intelligent intelligence and decision analysis module complete the data transmission and conversion through the system integration interface, so as to analyze and judge the decisions and situations of all parties in the strategic game process.
[0076] Another embodiment of the present invention provides an analysis and recommendation system for strategic game decision support, which includes:
[0077] The multi-granularity text semantic parsing and understanding module for game intelligence is used to perform semantic parsing and understanding on intelligence data. The semantic parsing and understanding results are obtained by performing semantic mining on texts of different granularities.
[0078] Multi-source intelligence data association and recommendation module, used to parse and associate multi-source intelligence data, and make intelligent recommendations based on the results of semantic analysis and understanding;
[0079] The countermeasure action auxiliary analysis and judgment module is used to judge the advantages and disadvantages of the current game situation based on semantic analysis and understanding, intelligence association and recommendation, so as to assist the game decision-making process.
[0080] The specific implementation process of each module refers to the above description of the method of the present invention.
[0081] Another embodiment of the present invention provides a computer device (computer, server, smart phone, etc.), which 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 each step in the method of the present invention.
[0082] Another embodiment of the present invention provides a computer-readable storage medium (such as ROM / RAM, magnetic disk, optical disk), wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a computer, the steps of the method of the present invention are implemented.
[0083] The present invention does not elaborate on some of the well-known technologies that belong to those skilled in the art.
[0084] The specific embodiments of the present invention disclosed above are intended to help understand the content of the present invention and implement it accordingly. It can be understood by those skilled in the art that various replacements, changes and modifications are possible without departing from the spirit and scope of the present invention. The present invention should not be limited to the contents disclosed in the embodiments of this specification, and the scope of protection of the present invention shall be subject to the scope defined in the claims.
Claims
1. An analysis and recommendation method for strategic game decision support, characterized in that: The following steps are involved: Perform semantic analysis and understanding of intelligence data, and obtain semantic analysis and understanding results by performing semantic mining on texts of different granularities; Analyze and correlate multi-source intelligence data, and make intelligent recommendations based on the results of semantic analysis and understanding; Based on the understanding of semantic analysis, association and recommendation of intelligence, the advantages and disadvantages of the current game situation are judged to assist the game decision-making process.
2. The method according to claim 1, characterized in that The semantic mining of texts of different granularities utilizes multi-granularity text semantic analysis technology to perform semantic mining at three levels: words, sentences, and paragraphs, complete shallow semantic analysis and deep semantic analysis of the text, and construct a multi-granularity text semantic understanding framework for intelligence analysis.
3. The method according to claim 1, characterized in that The multi-source intelligence data includes: intelligence data in the strategic intelligence library, historical situation data in the deduction situation library, and historical decision information in the decision information library.
4. The method according to claim 1, characterized in that The intelligent recommendation of intelligence includes two forms: offline processing and online processing, and collaborative filtering recommendation is performed on the intelligence decision data; the collaborative filtering recommendation filters data resources at the deduction situation level from the perspective of the strategic deduction decision maker, so as to meet the decision-making needs of different decision makers at multiple levels, and uses the similarity between the historical decisions of the decision makers as the basis for judgment. If the pre-decision attributes and action methods are similar to the decision maker's behavior and are in line with the current strategic situation, it can be regarded as a high-value recommended decision or intelligence.
5. The method according to claim 4, characterized in that The collaborative filtering recommendation comprises the following steps: Collect user interest information, mark and score the decision or intelligence information retrieved by the user. The more times the user searches, the higher the evaluation score. Perform nearest neighbor search based on users to find users with similar interests and hobbies to the user and calculate the similarity between users; Based on the calculated similarity, different forms of recommendations are made according to different recommendation purposes, including Top-N recommendations and relationship recommendations.
6. The method according to claim 1, characterized in that The above-mentioned analysis and judgment of the pros and cons of the current gaming situation to assist the gaming decision-making process includes: It provides auxiliary calculation of equipment capabilities and auxiliary analysis of battlefield situations, performs constraint calculations on equipment speed, maneuverability, strike range, and reconnaissance range, and performs GIS-based visualization to assist in decision-making and enhance the objectivity and quantification of strategic deductions.
7. The method according to claim 6, characterized in that The said judging the pros and cons of the current gaming situation to assist the gaming decision-making process also includes: Two modules are designed, namely the intelligent intelligence and decision-making analysis module and the intelligent intelligence service module; the intelligent intelligence service module is applied to the data layer and the processing layer. By managing the game situation database, the strategic intelligence database and the decision-making information database, the module accesses multi-source heterogeneous data such as the current strategic situation, game progress, the situation of the participating parties, and the strategic intentions of all parties. Through the game network environment, multi-machine collaboration and load uniformity under the multi-party game mechanism are realized; the intelligent intelligence and decision-making analysis module is located in the logical application layer, which can complete the whole process of data format conversion, data cleaning, mining and analysis, visualization display, report organization, writing and output, and realize the collaborative work among multiple functional components; the intelligent intelligence service module and the intelligent intelligence and decision-making analysis module complete the data transmission and conversion through the system integration interface, so as to analyze and judge the decisions and situations of all parties in the strategic game process.
8. An analysis and recommendation system for strategic game decision support, characterized in that: include: The multi-granularity text semantic parsing and understanding module for game intelligence is used to perform semantic parsing and understanding on intelligence data. The semantic parsing and understanding results are obtained by performing semantic mining on texts of different granularities. Multi-source intelligence data association and recommendation module, used to parse and associate multi-source intelligence data, and make intelligent recommendations based on the results of semantic analysis and understanding; The countermeasure action auxiliary analysis and judgment module is used to judge the advantages and disadvantages of the current game situation based on semantic analysis and understanding, intelligence association and recommendation, so as to assist the game decision-making process.
9. A computer device, 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.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a computer, the method according to any one of claims 1 to 7 is implemented.
