Test information compiling system and data mining method for accident risk value test

Abstract

The invention relates to a test information compiling system for accident risk value test and a data mining method, in the system, a file unit is used for independently recording single-dimensional behavior data stored by a safety management system; a first data mining analysis module mines out multi-dimensional behavior data from the behavior data of different dimensions stored in the archive unit in a classified manner; carrying out secondary classification on the mined multi-dimensional behavior data, and summarizing into a behavior data set; a second data mining analysis module independently screens out all behavior data related to the instruction target in a specific time interval from the behavior data set by taking the basic information of the instruction target selected by the decision maker as a reference; establishing a multi-dimensional behavior data model according to a preset weight configuration rule and the screened multi-dimensional behavior data, and generating a comprehensive evaluation report of the instruction target by using the multi-dimensional behavior data model; and the display terminal is used for inputting a weight configuration rule serving as a framework foundation of the multi-dimensional behavior data model by a decision maker.

Description

[0001] The original basis for this divisional application is patent application No. 202210733677.0, filed on June 23, 2022, entitled "A Method for Compiling Test Information for Accident Risk Value Testing". Technical Field

[0002] This invention relates to the field of group safety management system technology, and in particular to a test information compilation system and data mining method for accident risk value testing. Background Technology

[0003] In the information age, mobile internet, cloud computing, the Internet of Things, and big data technologies have been widely applied. The explosive growth of data and the expansion of its value will have a profound impact on the future development of the organization. Organization members use various targeted business tools in their daily work, training, life, and learning. However, the data from these tools is independent, and the content and format vary greatly. Over time, although a large amount of data accumulates, it does not enable a quick and comprehensive understanding of the information. Therefore, establishing an accurate big data platform for the organization's various grassroots operations is an urgent priority.

[0004] Patent document CN105357061A discloses an operation and maintenance monitoring and analysis system based on big data stream processing technology. The system includes: a monitoring terminal, used to obtain monitoring data in the client and send it to the storage terminal; a storage terminal, which stores multiple early warning processing rules, multiple data mining rules, and historical records; a caching terminal, which synchronizes the early warning processing rules, data mining rules, and historical records stored in the storage terminal to the caching terminal according to a preset time interval, and receives the monitoring data stream sent by the monitoring terminal; a first processing group, which performs early warning monitoring and alarm analysis based on the early warning rules, historical records, and monitoring data stream; and a second processing group, which performs data mining analysis based on the data mining rules, historical records, and monitoring data stream, and outputs monitoring statistical analysis based on the analysis results.

[0005] Patent document CN111914004A discloses an academic early warning method, system, and storage medium based on data mining algorithms. The method includes the following steps: cleaning campus big data to transform it into a uniform data format; classifying the cleaned campus big data and adding corresponding classification labels; performing principal component analysis on data with different classification labels to obtain multiple performance-related feature factors; inputting the obtained feature factors into a pre-trained performance prediction model to predict each student's performance; and marking a student with a performance below a warning threshold as subject to academic warning. This invention fully mines campus big data through data cleaning and principal component analysis, breaking down data silos and comprehensively analyzing each student's academic situation to enable teachers to provide personalized teaching guidance.

[0006] The aforementioned existing technologies can only analyze single-dimensional data sources and cannot analyze and manage multi-dimensional data that are independent of each other and have large differences in content and format. Therefore, there is a need for a multi-dimensional data collection scenario that can fit the group to help decision-makers make rational suggestions and decisions on the group's safety management, daily equipment and facility management, and comprehensive quality management of command objectives. In particular, it can help individuals or units to conduct targeted comprehensive assessments and training plans.

[0007] Furthermore, on the one hand, there are differences in understanding among those skilled in the art; on the other hand, the inventors studied a large number of documents and patents when making this invention, but due to space limitations, not all details and contents were listed in detail. However, this does not mean that the present invention does not possess the features of these prior art. On the contrary, the present invention already possesses all the features of the prior art, and the applicant reserves the right to add relevant prior art to the background art. Summary of the Invention

[0008] In response to the shortcomings of existing technologies, this invention can acquire comprehensive data from multiple third-party data acquisition systems of big data network platforms, and establish different analysis models according to actual application needs, fitting business scenarios and providing scientific basis and rational suggestions for decision-makers' decisions and daily work.

[0009] The present invention provides a method for compiling test information for accident risk value testing. The method includes the following steps: S1, creating a first test information database for instruction target testing, a second test information database for equipment testing, and a third test information database for environmental testing; S2, the test items in the second test information database are provided to the control unit of at least one piece of equipment related to training and life in a designated area to be evaluated for accident risk value in the form of data and signaling; S3, while the test items in the third test information database are provided to the control unit of at least one piece of equipment related to training and life in the designated area to be evaluated for accident risk value in the form of data and signaling, they are also provided in parallel to the control unit of at least one monitoring device related to training and life in the form of data and signaling.

[0010] According to a preferred embodiment, step S1 includes at least the following operations: S1.1 When creating the first test information database for testing command targets, the command targets in the specified area where the accident risk value needs to be evaluated are screened for the first time, and the test information whose test results conform to the normal distribution is selected to create the first test information database. S1.2. The test information in the first test information database is used to screen the command targets in the designated area whose accident risk values ​​need to be evaluated based on the equipment usage of the command targets during training. Test information whose test results conform to a normal distribution is selected to create a second test information database. S1.3. The test information in the second test information database is used to conduct a third screening of the command targets in the designated area whose accident risk value needs to be evaluated, based on whether the command targets can complete the training using the equipment under the specified environmental conditions. Test information whose test results conform to a normal distribution is selected to create a third test information database.

[0011] According to a preferred embodiment, the first test information database is created as follows: after analyzing the accident risk value, at least seven dimensions are set: ideology, safety skills, theory, physical training, daily management, teaching and self-study, and special training. A set of test information is set for each of the seven dimensions to form the first test information database.

[0012] According to a preferred embodiment, the seven-dimensional test information content includes specific test information items and project files that classify and record the behavioral data of the instruction targets in completing the test items. The instruction targets and their project file data contained in the first test information database are used to screen instruction targets in a designated area where accident risks need to be assessed based on various behavioral data of the instruction targets within a certain time period, thereby screening out instruction targets that can complete the test items in the designated area where accident risk values ​​need to be assessed.

[0013] According to a preferred embodiment, the stimulus information is classified, sorted, and stored in the file unit according to the type and / or format of the stimulus information. The first test information database, the second test information database, and the third test information database are used to screen the stimulus information stored in the file unit multiple times according to different screening conditions, so as to select the instruction target that can complete the test project using specific equipment in a specified environment.

[0014] According to a preferred embodiment, the behavioral data of the instruction target is mined using a first data mining analysis module, and the decision-maker uses a second data mining analysis module to filter multi-dimensional behavioral data associated with the instruction target. A multi-dimensional behavioral data model of the instruction target is established using preset weight configuration rules and the filtered multi-dimensional behavioral data within at least a portion of the time interval. A comprehensive evaluation report of the instruction target is then generated using this multi-dimensional behavioral data model, which uses the preset weight configuration rules as its framework. Its advantage lies in that the safety management system, through secondary classification and summarization of behavioral data from different dimensions, allows decision-makers to utilize various low-correlation behavioral data of the instruction target within the group to evaluate the overall situation of the instruction target. This ensures that the comprehensive evaluation report generated by the safety management system has comprehensive data, the evaluation results are fully supported by data, and the accuracy of the evaluation is improved. This allows decision-makers to provide targeted guidance and planning for the instruction target (individual or unit of behavior) based on the comprehensive evaluation report, facilitating the rapid determination of a reasonable development direction for the instruction target.

[0015] According to a preferred embodiment, the second data mining and analysis module can compare the comprehensive evaluation report generated by establishing a multi-dimensional behavioral data model with the reference data model pre-generated by the security management system using test data, thereby assessing the behavioral security of the command target; the second data mining and analysis module generates predictive information of at least one behavioral data associated with the comparison result based on the above comparison or retrieves predictive information of at least one behavioral data associated with the comparison result from the database, and then displays the obtained predictive information using a display terminal, so that decision-makers can manage the behavioral security of the command target based on the predictive information.

[0016] According to a preferred embodiment, the multi-dimensional behavioral data mined by the first data mining and analysis module from the security management system is classified and recorded using multiple file units. Each file unit independently records single-dimensional behavioral data stored in the security management system, and the file unit updates its recorded behavioral data in real time. The file unit also associates its recorded behavioral data with the individual or unit that generated the behavioral data.

[0017] According to a preferred embodiment, the multi-dimensional behavioral data mining performed by the first data mining and analysis module involves secondary classification and aggregation of behavioral data recorded by the security management system using file units and categorized by dimension. The secondary classification and aggregation refers to reclassifying and summarizing the behavioral data recorded by different file units based on the different behavioral individuals or behavioral units associated with the behavioral data.

[0018] According to a preferred embodiment, the second data mining and analysis module uses multiple single basic information of the instruction target as the retrieval basis to filter out behavioral data associated with the instruction target from the behavioral data mined and secondary classified and aggregated by the first data mining and analysis module, and the second data mining and analysis module performs secondary filtering on the filtered behavioral data associated with the instruction target according to the time interval determined by the decision-maker.

[0019] According to a preferred embodiment, the preset weight configuration rule refers to the relationship between different dimensions of behavioral data and the comprehensive quality data of the instruction target, which the decision-maker pre-inputs into the security management system using a display terminal. Thus, the second data mining and analysis module establishes a multi-dimensional behavioral data model of the instruction target by filling the multi-dimensional behavioral data that it has selected and that are associated with the instruction target within at least a part of the time interval into the relationship.

[0020] According to a preferred embodiment, the decision-maker determines the weight of different dimensions of behavioral data based on the degree of correlation between the behavioral data of the command target within at least a certain time interval that can be monitored by the safety management system and the comprehensive quality data in the comprehensive assessment report to be generated. Thus, the decision-maker can formulate weight configuration rules that allow for differences in the weight values ​​corresponding to different dimensions of behavioral data.

[0021] According to a preferred embodiment, the predictive information of the behavioral data generated or retrieved by the second data mining and analysis module is the historical dataset information of the behavioral individuals or behavioral units recorded by the security management system or the pre-entered test dataset information. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the workflow of a preferred method for compiling test information for accident risk value testing proposed in this invention; Figure 2 This is a schematic diagram of the database creation steps for a preferred method of compiling test information for accident risk value testing proposed in this invention. Figure 3 This is a normal distribution diagram that is satisfied when screening the test information of a preferred accident risk value test test information compilation method proposed in this invention.

[0023] List of reference numerals 1: Archive unit; 2: First data mining and analysis module; 3: Second data mining and analysis module; 4: Display terminal; Q1: First test information database; Q2: Second test information database; Q3: Third test information database. Detailed Implementation

[0024] The following is a detailed explanation with reference to the accompanying drawings.

[0025] The accident risk value testing information compilation method provided by this invention can interface with other third-party data acquisition systems to obtain the group's safety record data (e.g., group personnel attendance, outings, returns, drills, daily routines, boundary warnings, and patrol records), thereby establishing the group's activity archives and analyzing the group's safety management issues. This invention can also comprehensively analyze the abnormal distribution locations, time periods, personnel, and affiliated organizations of various groups' safety data, and issue early warnings for abnormal situations. Furthermore, this invention can provide emergency process management for group decision-makers. Decision-makers can set emergency processes according to actual needs (processes include equipment (equipment) location, personnel location, administrator login and positioning, and material positioning), and combine this with the group's vehicle and material equipment and facility data to predict the emergency drill process and results.

[0026] Example 1 This embodiment provides a method for compiling test information for accident risk value testing, including an archive unit 1, a first data mining and analysis module 2, a second data mining and analysis module 3, a display terminal 4, and several data processing modules capable of supporting a safety management system for group safety management. The modules mentioned in this embodiment can refer to hardware, software, or a combination of data processors capable of performing their related steps. A method step corresponding to a certain module can also be broken down into multiple method steps and executed separately by multiple modules. Where there is no conflict or contradiction, the whole and / or part of the preferred embodiments of other embodiments can be used as supplements to this embodiment.

[0027] according to Figure 1-3 One specific implementation method shown is a test information compilation method for accident risk value testing, which includes the following steps: S1. Create a first test information database Q1 for instruction target testing, a second test information database Q2 for device testing, and a third test information database Q3 for environmental testing; S1.1 When creating the first test information database Q1 for testing command targets, the command targets in the specified area where the accident risk value needs to be evaluated are screened for the first time, and the test information whose test results conform to the normal distribution is selected to create the first test information database Q1. S1.2. The test information in the first test information database is used to screen the command targets in the designated area where the accident risk value needs to be evaluated based on the equipment usage of the command targets during training. Test information whose test results conform to a normal distribution is selected to create the second test information database Q2. S1.3. The test information in the second test information database is used to conduct a third screening of the command targets in the designated area whose accident risk value needs to be evaluated, based on whether the command targets can complete the training using the equipment under the specified environmental conditions. Test information whose test results conform to a normal distribution is selected to create the third test information database Q3.

[0028] S2, the test items of the second test information database Q2 are provided to the control unit of at least one training life-related device in a designated area where the accident risk value is to be evaluated in the form of data and signaling.

[0029] The test items of S3 and the third test information database Q3, when provided to the control unit of at least one work equipment related to training life in the designated area where the accident risk value is to be evaluated in the form of data and signaling, are also provided in parallel to the control unit of at least one monitoring device related to training life in the form of data and signaling.

[0030] Preferably, the archive unit 1 classifies and stores the behavioral data of different dimensions of groups acquired by the security management system from other third-party data acquisition systems according to the different dimensions of the behavioral data, and stores them in different locations or storage sub-units. The first data mining and analysis module 2 can mine behavioral data belonging to different groups, individuals, or equipment and facilities from the behavioral data of different dimensions stored in the archive unit 1. The first data mining and analysis module 2 can summarize the mined behavioral data into a behavioral dataset by performing secondary classification on the mined behavioral data of different dimensions. The second data mining and analysis module 3 can use the basic information of the instruction target selected by the decision-maker as a benchmark to separately filter all behavioral data related to the instruction target within a specific time interval from the behavioral dataset summarized by the first data mining and analysis module 2. The second data mining and analysis module 3 can use the filtered behavioral data to establish a multi-dimensional behavioral data model that can be used to evaluate the comprehensive situation of the instruction target within a specific time interval. Preferably, multi-dimensional behavioral data refers to behavioral data of different dimensions. The decision-maker can input the weight configuration rules as the framework basis of the multi-dimensional behavioral data model through the display terminal 4. Weighting rules characterize the correlation between behavioral data from different dimensions and the overall situation of the instruction objective. This allows for the determination of the weight of different behavioral data in generating a comprehensive evaluation report of the instruction objective based on their impact on the objective. This invention, through secondary classification and summarization of behavioral data from different dimensions, facilitates decision-makers' use of various behavioral data with low correlation within the group to evaluate the overall situation of the instruction objective. This results in comprehensive evaluation reports of the instruction objective generated by the safety management system, ensuring data comprehensiveness and providing sufficient data support for the evaluation results. This improves the accuracy of the evaluation and allows decision-makers to provide targeted guidance and planning for the instruction objective (individual or unit) based on the comprehensive evaluation report, promoting the rapid determination of a reasonable development direction for the instruction objective. Preferably, the instruction objective can be an individual or unit, where an individual can be any person, instructor, or student within the group.

[0031] Preferably, the first data mining and analysis module 2 mines multi-dimensional behavioral data stored in the security management system through automatic monitoring or data scanning. Preferably, the multi-dimensional behavioral data mined by the first data mining and analysis module 2 from the security management system is categorized and stored by the security management system using multiple file units 1. When the security management system obtains behavioral data from other third-party data acquisition systems, it categorizes and stores the data according to the different dimensions of the behavioral data, thus forming multi-dimensional behavioral data. That is, the security management system categorizes and stores the behavioral data based on the different sources (representing the terminal acquisition devices of the third-party data acquisition systems). For example, a file unit 1 may store usage records of all personnel using the same device or the same training ground at different times. Preferably, different dimensions of behavioral data mean that due to different terminal acquisition devices or training devices generating the behavioral data, there is no data correlation between different behavioral data other than a temporal relationship. For example, usage data collected by the training ground usage record system and sleep status data collected by the sleep monitoring system installed in the dormitory do not have a direct connection. Preferably, any file unit 1 stores only the behavioral data collected by one terminal acquisition device received by the security management system. The file unit 1 updates its recorded behavioral data in real time. Preferably, the terminal acquisition device for acquiring behavioral data collects data in real time and continuously, thus the data transmitted to the security management system is also constantly increasing. Therefore, the archive unit 1 needs to continuously update or add the behavioral data it stores. The behavioral data stored in the archive unit 1 may be recorded by multiple individuals or units using the same training device. Therefore, the archive unit 1 also associates the behavioral data generated by different instruction targets using the training device at different times / time periods with the basic information of the instruction target (behavioral individual or behavioral unit). Thus, the first data mining and analysis module 2 mines the behavioral data generated by different terminal acquisition devices and associated with the instruction target based on the basic information of the instruction target. Preferably, the multi-dimensional behavioral data mining performed by the first data mining and analysis module 2 is a secondary classification and aggregation of the behavioral data recorded by the security management system using the archive unit 1 in a dimensional division manner. Preferably, the secondary classification and aggregation refers to reclassifying and summarizing the behavioral data recorded by different archive units 1 according to the different behavioral individuals or behavioral units associated with the behavioral data. The behavioral data recorded in the archive unit 1 is classified and stored according to the different terminal acquisition devices. To facilitate the extraction of behavioral data for the specified instruction target by the second data mining and analysis module 3, the first data mining and analysis module 2 performs secondary classification processing on the behavioral data extracted from the archive unit 1, so that behavioral data of different dimensions are grouped into different subsets of behavioral individuals or behavioral units according to the different behavioral individuals or behavioral units associated with the behavioral data.Preferably, secondary classification and aggregation refers to further splitting and summarizing the already classified data into different datasets according to different classification methods. Preferably, the datasets summarized in the secondary aggregation can be arranged according to the order of collection time.

[0032] Preferably, the second data mining and analysis module 3 responds to the decision-maker's instructions by filtering multi-dimensional behavioral data associated with the instruction target. The second data mining and analysis module 3 can retrieve various behavioral data of different dimensions associated with the instruction target from the behavioral data collected by the second classification of the first data mining and analysis module 2 based on the basic information of the instruction target. Preferably, the second data mining and analysis module 3 can also establish a multi-dimensional behavioral data model of the instruction target based on preset weight configuration rules and the filtered multi-dimensional behavioral data within a time interval, thereby generating a comprehensive evaluation report of the instruction target using the multi-dimensional behavioral data model. Preferably, the decision-maker determines the weight of different dimensions of behavioral data based on the correlation between the different dimensions of behavioral data recorded by the safety management system within a time interval and the comprehensive quality data of the comprehensive evaluation report to be generated, thereby enabling the decision-maker to formulate weight configuration rules that differentiate the weight values ​​corresponding to different dimensions of behavioral data. The second data mining and analysis module 3 further filters the retrieved behavioral data by limiting the collection time of the behavioral data, thereby obtaining all behavioral data of the instruction target within a time interval. For example, the training time and performance of an individual using all equipment or venues within a week. Preferably, the length of the time interval is selectively determined according to the decision-maker's needs. Preferably, the individual actor can be any member of the group, and the unit of action can be the entire group or a small collective.

[0033] Preferably, the second data mining and analysis module 3 can compare the comprehensive evaluation report generated by establishing a multi-dimensional behavioral data model with the reference data model pre-generated by the safety management system using test data, thereby achieving an assessment of the behavioral safety of the instruction target. Preferably, the test data of the safety management system can be a complete historical data set of other behavioral individuals or units, reference data loaded from other systems, or standard reference data obtained through big data analysis methods. Preferably, the reference data model refers to a data model established using test data and the same weight configuration rules, and the comprehensive evaluation data and comprehensive evaluation report output by this data model are the comprehensive evaluation reference data and comprehensive evaluation reference report. Preferably, decision-makers can assess the behavioral safety of the instruction target by comparing the comprehensive evaluation report generated from the behavioral data of the instruction target with the comprehensive evaluation reference report. Preferably, the assessment of behavioral safety can refer to judging whether the daily work, training, life, learning, and physical condition of the instruction target are reasonable. Preferably, the second data mining and analysis module 3 retrieves prediction information of at least one behavioral data associated with the comparison results from the database based on the comparison results, and then displays the obtained prediction information using the display terminal 4, so that the decision-maker can perform behavioral safety management of the instruction target based on the prediction information.

[0034] Preferably, the predictive information of the behavioral data retrieved by the second data mining and analysis module 3 is extracted by the safety management system from historical datasets of recorded behavioral individuals or units, or from pre-entered test datasets. Preferably, the historical dataset can be a complete historical dataset of other behavioral individuals or units. Preferably, the test dataset is a summary of the aforementioned test data. Preferably, the second data mining and analysis module 3 can compare the comprehensive evaluation report generated by establishing a multi-dimensional behavioral data model with the reference data model pre-generated by the safety management system using the test data to assess the behavioral safety of the target. Preferably, the predictive information refers to the analysis and prediction of the real-time status and future behavioral safety of the target using historical behavioral data that corresponds to the behavioral data of the target in the historical dataset. This facilitates decision-makers in providing targeted guidance and planning for the target (behavioral individual or unit) based on the comprehensive evaluation report, promoting the rapid determination of a reasonable development direction for the target.

[0035] The second data mining and analysis module 3 uses multiple single basic information about the instruction target as a retrieval basis to filter out behavioral data associated with the instruction target from the behavioral data processed by the second classification and aggregation of the first data mining and analysis module 2. Furthermore, the second data mining and analysis module 3 performs a second filtering of the filtered behavioral data associated with the instruction target based on the time interval determined by the decision-maker. Preferably, the single basic information can be the instruction target's name, number, or other basic information with unique characteristics. Preferably, the preset weight configuration rule refers to the relationship between the different dimensions of behavioral data of the security management system and the comprehensive quality data of the instruction target, which the decision-maker pre-inputs using the display terminal 4. The second data mining and analysis module 3 then establishes a multi-dimensional behavioral data model of the instruction target by filling in the multi-dimensional behavioral data associated with the instruction target within a certain time interval that it has filtered out. The multi-dimensional behavioral data model is constructed by combining the preset weight configuration rule with the multi-dimensional behavioral data, and it directly generates a comprehensive evaluation report based on the preset weight configuration rule and the multi-dimensional behavioral data.

[0036] Example 2 This embodiment is a further improvement on embodiment 1, and repeated content will not be described again.

[0037] Preferably, the test information compilation method for accident risk value testing can also consist of a data comprehensive display module, a capability support module, and a business-specific application module. The data comprehensive display module can classify and display different indicators and dimensions of data from numerous business systems within a group, helping the target to observe and analyze data from different perspectives and focus on data trends and patterns. The data comprehensive display module can also display the calculation and analysis results of different models and allow for cross-sectional viewing based on multiple dimensional conditions. Furthermore, the data comprehensive display module also establishes an early warning mechanism based on data mining by establishing early warning indices or models to monitor and analyze data related to them. When real-time data reaches predetermined limits or an anomaly occurs, the safety management system will automatically issue an alarm. Simultaneously, the safety management system supports setting limits such as alarm thresholds and confidence thresholds. Preferably, the safety management system supports multiple data filtering methods, including field queries, map selection, and point selection. Preferably, the capability support module can process data by reading comprehensive data and summarizing the processed data. The capability support module can also perform data interaction by selectively developing different types of data ports and providing real-time feedback of calculation results. Preferably, decision-makers can also customize the parameter weights of the application models established for each business involved in the security management system, so that each model has growth factors and can be corrected based on the model output results.

[0038] Preferably, the business-themed application modules include at least the following: comprehensive flight personnel capability data application, comprehensive personnel quality data application, support quality data application, equipment failure and potential hazard data application, flight environmental factor data application, flight safety risk data application, group safety management data application, and comprehensive assessment data application. Preferably, the comprehensive flight personnel capability data application comprehensively analyzes various data from instructors and trainees to derive individual strengths and weaknesses, provide a comprehensive assessment, and periodically offer development directions and gap-filling analysis results. It automatically generates scores for individuals, units, and the entire group; the comprehensive flight personnel capability data application can also automatically generate scores for individuals, units, and the entire group.

[0039] Preferably, the personnel comprehensive quality data application includes an examination system with customizable test questions and weighted scoring; it assesses instructors and trainees separately, automatically generating individual and unit scores. Based on this examination, the system analyzes trainees' development direction and shortcomings by combining comprehensive data and historical data on their ideology, skills, safety skills, theory, physical training, daily management, self-study, and specialized training. Preferably, the quality assurance data application analyzes single and multi-dimensional aspects of equipment (attendance rate, availability rate, maintenance quality, failure rate, etc.) to reflect the overall status of the equipment (condition, historical data, etc.), and provides real-time alerts to instructors and trainees on whether work training can proceed in a way that integrates with work training scenarios. Preferably, the equipment fault hazard data application compares flight parameters, pilot feedback issues, and quality assurance data, comprehensively comparing data from similar equipment to analyze common problems. Preferably, the flight environment factor data application models and analyzes meteorological information, geographical information, terrain information, electromagnetic interference, flight activities, and site conditions before flight, providing relevant prompts to pilots after they select the area, time, and altitude for the day's flight. Preferably, the flight safety risk data application is integrated with the flight safety assessment plan implementation system and copies flight safety dynamic monitoring data to conduct a comprehensive analysis of the flight battalion's flight cycle, generate a safety assessment report, and provide detailed analysis for each item. Preferably, the group safety management data application performs real-time analysis of various data related to group safety management, identifies problems in specific aspects of group safety and trends in concentrated phases, provides early warnings for high-frequency and sudden issues, and comprehensively grasps the changes in group instability factors. Preferably, the comprehensive assessment data application analyzes and evaluates the overall combat capability of the entire brigade through the above data analysis, providing better support for decision-making in areas such as personnel training, combat capability adjustment, management of various instability factors, and mission execution.

[0040] Preferably, the behavioral data mining and retrieval operations of the first data mining analysis module 2 and the second data mining analysis module 3 are based on the text features of the retrieval request or retrieval target. They use a retrieval engine and machine learning models to obtain retrieval results based on text features. Simultaneously, they combine text feature retrieval methods with knowledge graph-based knowledge association, knowledge reasoning based on business objectives, and computation based on graph computing, machine learning, and other methods to obtain multi-dimensional intelligent retrieval of personnel. Preferably, the retrieval dimensions include basic information, skill information, qualifications, technical skills, and professional inclinations. Preferably, the retrieval operation can also perform reasoning and display based on personnel basic attributes, related events (training, examinations, simulations, etc.), and related equipment (aircraft being piloted, training equipment, etc.), further presenting personnel retrieval information in a deeper, more multi-dimensional, and intuitive way.

[0041] Preferably, the safety management system, based on a big data platform, establishes a personnel data tagging system through a holographic multi-dimensional profile tagging system for developers. It integrates and calculates information such as basic information, ideology, technology, safety skills, theory, physical training, daily management, teaching and self-study, and training of various personnel to construct a holographic knowledge graph of personnel information. Based on this, combined with artificial intelligence and natural language processing technology, it achieves automatic evaluation and tagging of personnel's professional qualities, technical capabilities, safety skills, training results, and teaching quality, forming a 360-degree holographic profile of personnel. This supports personnel selection and management analysis in training, examination, and simulation exercises, improving training effectiveness and providing data support for long-term training planning. The safety management system also establishes a closed loop covering the tagging system construction, tag management, personnel profiling, tag application, and effectiveness evaluation tag construction required for holographic profile assessment. This empowers teaching and training management through the precise implementation of the tagging system in various scenarios. Preferably, the safety management system also implements a lifecycle management process for tag creation, review, release, evaluation, deactivation, and optimization, achieving full lifecycle management of personnel tags and ensuring their high efficiency, stability, practicality, and effectiveness. This invention enables periodic evaluation and version management of tags through tag lifecycle management, ensuring the adaptability of tags to various scenario analyses and supporting the rapid implementation of various tag operation organizational structures (strong matrix, balanced matrix, weak matrix).

[0042] Preferably, the present invention enables dynamic management of label indicators. Specifically, dynamic management of label indicators includes management using traditional statistical analysis models and management using big data analysis models. Traditional statistical analysis models construct label indicators from multiple dimensions and summarize them into statistical result labels; big data analysis models extract various data features that describe the overall picture of personnel and categorize and organize the data features to form big data analysis labels. The present invention achieves dynamic management of label indicators by defining label rules (statistical analysis or calling big data models). Preferably, the security management system of the present invention can also construct group profiles of personnel, displaying the group label status of examiners, instructors, etc., enabling rapid identification of the group characteristics of such personnel, assisting professional departments in formulating differentiated management strategies for different groups of such personnel, and guiding the development of appropriate teaching, training, and logistical support plans. Individual profiles of personnel are constructed based on information such as the qualifications, technical abilities, and professional inclinations of examiners or trainees, constructing individual profiles of personnel from the perspective of personnel management, and comprehensively displaying personnel information from specific hierarchical evaluation indicators.

[0043] Preferably, the main data sources for the comprehensive evaluation report of this invention include instructor data and student data. Instructor data includes flight theory scores, senior instructor evaluations, student performance, number of emergency responses, and flight frequency and quality (obtained from examiner scores and flight parameters). Student data includes flight theory scores, test scores, instructor evaluations, and flight performance. This invention can perform trend analysis on the units to which students and instructors belong (such as teaching groups, companies, and battalions), facilitating decision-makers' understanding of student performance changes and providing data support for selecting the best teaching methods, intensity, frequency, and combinations. The safety management system of this invention also provides data interfaces that can provide auxiliary decision-making information to other programs.

[0044] Preferably, comprehensive quality data refers to the summary of information on each person's ideology, skills, safety skills, theoretical knowledge, physical training, daily management, teaching and self-study, and training. This invention can conduct single-directional and multi-directional comprehensive analyses for each person, thereby reflecting each pilot's strengths and weaknesses and providing a basis for the development plan of each pilot's comprehensive abilities. Simultaneously, this invention also identifies the strengths and weaknesses of an organization through historical data analysis and multi-dimensional analysis at the unit level.

[0045] Preferably, the data recorded in file unit 1 can also come from each person's ideological information, technical information, safety skills information, theoretical information, physical training information, daily routine information, teaching and self-study information, training information, attendance records, and data from the group's comprehensive management system. Preferably, the safety management system provides the function of setting the weights of various parameters, enabling the safety management system to analyze the different attributes of personnel or organizations according to the weight configuration given by the decision-maker, thereby allowing terminal 4 to display the following content: It presents the changing trends of personnel in the same professional field at different times and with alerts for sudden changes; It showcases the different abilities of individuals at the same time, and provides insights into their strengths and weaknesses; Present the organization's average performance in various aspects during the same period, and highlight its strengths and weaknesses; It presents the changing trends of an organization in the same area of ​​work over different periods and provides alerts for sudden changes.

[0046] Example 3 This embodiment is a further improvement on embodiment 1, and repeated content will not be described again.

[0047] The safety management system organizes, categorizes, and stores personnel data, equipment data, and environmental data collected by different terminal acquisition devices. Preferably, personnel data is analyzed to make a preliminary assessment of personnel's real-time status, and based on the assessment results and real-time personnel status monitoring data, a judgment is made as to whether the personnel possess the ability to perform tasks or whether they can undergo further training. Preferably, during personnel data analysis, equipment data is also analyzed to determine the actual usage status of the equipment, thereby ensuring that the equipment meets the personnel's needs while the personnel possess the corresponding capabilities. Furthermore, when suitable personnel (instruction targets) are required to perform specific tasks or training, the environment of the task or training area also needs to be analyzed to further screen personnel within the group.

[0048] Preferably, personnel data, equipment data, and environmental data can be used as three screening criteria to select personnel from the group, thereby identifying the best personnel capable of completing specific tasks or assessing the success rate of different personnel in performing specific tasks. Preferably, personnel data can include basic information, ideology, skills, safety skills, theoretical knowledge, physical training, daily management, teaching and self-study, and training information for various personnel. Preferably, equipment data can include equipment category, model, purchase date, purchase method, scrapping deadline, validity period, user unit, user, custodian, storage location, and remarks. Preferably, remarks can be used to record equipment maintenance and abnormal operation information. Preferably, environmental data can include weather forecasts, geographical information, site conditions, obstacle distribution, electromagnetic interference, and flight activities in the flight area. When group personnel need to conduct flight training, a comprehensive collection of information on the flight area, including weather forecasts, geographical information, site conditions, obstacle distribution, electromagnetic interference, and flight activities, along with a comprehensive analysis of the flight environment, identifies environmental factors that may affect flight execution and safety. This provides a basis for developing targeted preventative measures, implementation plans, and schedules, thereby improving training quality and ensuring flight safety. Furthermore, by analyzing environmental data and personnel and equipment data to assess the recent training status of personnel and the operational status of equipment, a risk assessment can be conducted to determine the suitability of different personnel for flight missions, ultimately selecting appropriate personnel for the missions.

[0049] Preferably, this embodiment also provides a safety scoring processing method, including: determining the historical ability indicators, historical physiological indicators, historical psychological indicators, real-time psychological indicators, and real-time physiological indicators of the personnel to be evaluated based on the personnel data, and further obtaining a personnel score. Determining the historical performance indicators and current performance indicators of the equipment to be evaluated based on the equipment data, and further obtaining a equipment score. Determining the task difficulty score associated with the personnel and equipment based on the task data, the historical ability indicators of the personnel to be evaluated, and the equipment data. Determining the environment score associated with the task and equipment based on the environmental data, the equipment data, and the task data. Based on preset scoring rules, determining the safety score of the personnel to be evaluated performing the task under the environmental conditions of the task, according to the personnel score, task difficulty score, and environment score. Specifically, the personnel to be evaluated are allowed to perform the task only when the safety score is higher than a preset safety threshold to ensure that the task can be safely performed by the personnel to be evaluated in the task execution environment.

[0050] The personnel data includes basic information, ideology, skills, safety skills, theoretical knowledge, physical training, historical assessment scores, historical ideological evaluations, historical psychological evaluations, historical task completion status, simulated theoretical exams, simulated practical exams, and current physical fitness assessments. This personnel data is managed in a closed-loop dynamic tagging system, establishing a precise data tagging system through tag creation, review, publication, evaluation, and deactivation. The tagging system is dynamically managed through indicators, including: constructing tag indicators across multiple dimensions and summarizing them into statistical result tags; using big data analysis models to extract various data features describing the overall picture of personnel, and classifying and organizing these features based on analysis objectives to form big data analysis tags. Dynamic management of tag indicators is achieved by defining tag rules (statistical analysis or calling big data models), the frequency of updates, and the scheduling of update tasks. Group profiles of personnel are constructed to quickly identify the characteristics of different groups, assisting professional departments in developing differentiated management strategies for different groups of personnel. Individual profiles are constructed based on personnel qualifications, technical skills, and professional inclinations, providing a comprehensive view of personnel information from the perspective of personnel management and through specific hierarchical evaluation indicators.

[0051] The data to be evaluated includes the planned number of equipment in use, the number of times equipment is used, the number of times equipment fails, the planned number of personnel in use, the number of times personnel is used, historical maintenance indicators, factory performance records, spare parts replacement records, and spare parts performance records. The equipment performance data is correlated with at least the following parameters: temperature, humidity, wind speed, and oxygen content in the air.

[0052] Preferably, environmental data includes regional weather forecasts, geographic information, site conditions, obstacle distribution, electromagnetic interference, and flight activities. For various types of environmental data, a Geographic Information System (GIS) is used to present risk factors in the form of regional coloring, intuitively indicating the risk level of the area. Furthermore, for a single task, the weights can be adjusted based on the scores given by the personnel being assessed, so that the same environmental factor presents different risk levels for personnel with different characteristics; and the risk level will change over time and as environmental factors change.

[0053] Preferably, after each task is completed, a task completion score is determined based on the task's completeness, and the preset scoring rules are adjusted based on the comparison and matching between the task completion score and the safety score. For example, when the difference between the task completion score and the safety score exceeds a preset accuracy threshold, an alert is issued, and relevant management personnel are notified to adjust and verify the preset rules. Personnel and equipment data are updated based on the task completion score. Task completion status includes completion time, action accuracy, route accuracy, the actual number of hazards during the task, and the hazard value of each hazard.

[0054] Preferably, the safety management system uses real-time equipment data (aircraft usage data and maintenance data) and real-time environmental data to conduct targeted analysis and assessment of safety risk points, thereby identifying potential safety hazards. It also focuses on conducting special assessments of flight safety risks during special periods and major missions, which helps to reduce safety hazards.

[0055] Preferably, flight safety analysis refers to screening for potential hazards in a flight mission using comprehensive flight record data, equipment maintenance and usage record data, comprehensive numerical data of flight personnel, and data related to the flight environment, thereby ensuring the flight safety of flight personnel. Preferably, the safety management system also has a pre-simulation function for risk assessment of designated flight activities. Specifically, based on personnel data, equipment data, and existing flight environment data, the safety management system provides information on potential risks that may arise when personnel perform designated flight projects under hypothetical conditions, thereby simulating and predicting the risks of flight activities that groups of personnel need to perform during major projects and important periods.

[0056] Preferably, in use, decision-makers can select the aircraft type, pilot, flight date, flight area, flight route, and high-difficulty subjects within the safety management system. The safety management system will then provide a risk assessment value for the pilot's flight at the specified time, which can be used as a reference by planning personnel. Preferably, the flight safety risk factor assessment conducted by the safety management system can utilize different analytical methods depending on the selected reference data, and the reference value of the final output analysis results will also differ, allowing decision-makers to conduct different flight safety risk factor assessments based on the different analytical results. Preferably, the analytical methods can be categorized based on the different reference data, including horizontal comparison, trend analysis, lineage mining, and factor change simulation assessment. Specifically, horizontal comparison refers to using flight environment factor data, pilot comprehensive quality data, equipment support data, and flight parameter record data to trace the causes of flight risks encountered by pilots during flight, and analyze the reasons for specific flight risks encountered by a particular pilot, thereby generating the probability of various flight risks encountered by that pilot during flight. Specifically, trend analysis refers to using flight environment factor data, pilot comprehensive quality data, equipment support status data, and flight parameter record data to plot flight safety risk coefficient trend charts, thereby analyzing and predicting the direction of change in aircraft safety risk factors. Specifically, lineage analysis refers to using flight environment factor data, pilot comprehensive quality data, equipment support status data, flight parameter record data, the implementation status of safety assessment plans, sub-item safety assessment values, and overall assessment scores to analyze the distribution of flight safety risk factors for pilots, thereby assisting in the development of pilot skills training programs. Specifically, factor change simulation assessment refers to using flight environment factor data, pilot comprehensive quality data, equipment support status data, flight parameter record data, the implementation status of safety assessment plans, sub-item safety assessment values, and overall assessment scores to validate the established model, thereby predicting pilot development and simulating and predicting risks for specified flight activities, thus providing early warnings of potential flight risks and providing pilots with training directions conducive to capacity building.

[0057] It should be noted that the specific embodiments described above are exemplary. Those skilled in the art can devise various solutions inspired by the disclosure of this invention, and these solutions all fall within the scope of this invention and its protection. Those skilled in the art should understand that this specification and its accompanying drawings are illustrative and not intended to limit the scope of the claims. The scope of protection of this invention is defined by the claims and their equivalents. Throughout the text, features introduced by "preferred" are merely optional and should not be construed as mandatory. Therefore, the applicant reserves the right to abandon or delete relevant preferred features at any time.

Claims

1. A test information compilation system for accident risk value testing, characterized in that, The system comprises: An archive unit (1) for independently recording single-dimension behavior data stored by a safety management system and updating the recorded behavior data in real time; and for associating the recorded behavior data with a behavior individual or a behavior unit generating the behavior data; A first data mining analysis module (2) for mining multi-dimension behavior data from different-dimension behavior data classified and stored by the archive unit (1); for secondarily classifying the mined multi-dimension behavior data; and for summarizing the behavior data into a behavior data set; A second data mining analysis module (3) for individually screening all behavior data related to an instruction target within a specific time interval from the behavior data set summarized by the first data mining analysis module (2) based on basic information of the instruction target selected by a decision maker; for establishing a multi-dimension behavior data model of the instruction target according to a preset weight configuration rule and the multi-dimension behavior data screened within the time interval; and for generating a comprehensive evaluation report of the instruction target by the second data mining analysis module (3) using the multi-dimension behavior data model. A display terminal (4) for inputting, by the decision maker, the weight configuration rule as a framework basis of the multi-dimension behavior data model.

2. The test information preparation system according to claim 1, characterized by, The first data mining analysis module (2) mines the multi-dimension behavior data stored by the safety management system using the archive unit (1) through automatic monitoring or data scanning.

3. The test information preparation system according to claim 1 or 2, characterized by, The first data mining analysis module (2) secondarily classifies the mined multi-dimension behavior data in the following process: The classified data is secondarily split according to different classification methods and summarized into different data sets, so that the behavior data of different dimensions is collected in different sub-sets of behavior individuals or behavior units according to the different behavior individuals or behavior units associated with the behavior data.

4. The test information preparation system according to any one of claims 1 to 3, characterized by, The second data mining analysis module (3) further screens the retrieved behavior data by limiting the collection time of the behavior data, thereby obtaining all behavior data of the instruction target within a time interval.

5. The test information preparation system according to any one of claims 1 to 4, characterized by, The second data mining analysis module (3) compares the comprehensive evaluation report generated by establishing the multi-dimension behavior data model with a reference data model generated in advance by the safety management system using test data, thereby evaluating the behavior safety of the instruction target. The second data mining analysis module (3) generates or retrieves prediction information of at least one behavior data associated with the comparison result according to the comparison, thereby displaying the obtained prediction information by the display terminal (4) so that the decision maker manages the behavior safety of the instruction target according to the prediction information.

6. The test information preparation system according to any one of claims 1 to 5, characterized by, The second data mining analysis module (3) secondarily screens the behavior data associated with the instruction target according to a time interval determined by the decision maker, and then establishes the multi-dimension behavior data model of the instruction target by filling the multi-dimension behavior data associated with the instruction target within the time interval into a relational expression.

7. The test information preparation system according to any one of claims 1 to 6, characterized by, When the first test information database (Q1) for instruction target test is created, the first screening is performed on the instruction targets in the specified area to be evaluated for the accident risk value, and the first data mining analysis module (2) selects the test information with test results conforming to the normal distribution from the stimulation information stored in the archive unit (1), and then creates the first test information database (Q1); The first data mining analysis module (2) performs the second screening on the instruction targets in the specified area to be evaluated for the accident risk value according to the equipment usage of the instruction targets during training, selects the test information with test results conforming to the normal distribution from the test information of the first test information database (Q1), and creates the second test information database (Q2) in this way; The first data mining analysis module (2) performs the third screening on the instruction targets in the specified area to be evaluated for the accident risk value according to whether the instruction targets can use the equipment to complete the training under the specified environmental conditions, selects the test information with test results conforming to the normal distribution from the test information of the second test information database (Q2), and creates the third test information database (Q3) in this way.

8. The test information preparation system according to any one of claims 1 to 7, characterized by, The first data mining analysis module (2) first analyzes the accident risk value, then sets at least seven dimensions of thought, safety skill, theory, physical training, daily management, teaching self-study and special training, and then sets a group of test information for each of the seven dimensions, and finally forms the first test information database (Q1).

9. The test information preparation system according to Claim 8, wherein The test information content of the seven dimensions includes specific items of test information and item archives for classifying and recording the behavior data of the instruction targets for completing the test items, and the instruction targets and item archive data included in the first test information database (Q1) are based on the multiple behavior data of the instruction targets within a certain period of time to screen the instruction targets in the specified area to be evaluated for the accident risk value, so as to screen out the instruction targets capable of completing the test items in the specified area to be evaluated for the accident risk value.

10. A data mining method of a test information preparation system of an accident risk value test, characterized by, The method comprises: The decision maker inputs the weight configuration rule as the framework basis of the multi-dimensional behavior data model; The single-dimensional behavior data stored in the safety management system is recorded independently, and the recorded behavior data is updated in real time; the recorded behavior data is also associated with the behavior individual or behavior unit generating the behavior data; Multi-dimensional behavior data is mined from the classified and stored behavior data of different dimensions; the mined multi-dimensional behavior data is classified again, and is aggregated into a behavior data set; The basis information of the instruction target selected by the decision maker is taken as a reference to separately screen all the behavior data involved by the instruction target in a specific time interval from the behavior data set summarized by the first data mining analysis module (2); and a multi-dimensional behavior data model of the instruction target is established according to the preset weight configuration rule and the screened multi-dimensional behavior data in the time interval, so that the second data mining analysis module (3) generates a comprehensive evaluation report of the instruction target by using the multi-dimensional behavior data model.

Images