Analysis method based on combination of police affair data containerization parallel throughput and AI model

Abstract

The invention discloses an analysis method based on combination of police service data containerization parallel throughput and an AI model, and the method achieves the receiving, collection and extraction of massive police service source data through a data container, and carries out the cleaning, conversion, decoupling, deduplication, grid conversion and verification, and forms standard structure metadata. The standard structure metadata is distributed to an actual combat layer, an evaluation layer and an AI model training area in parallel through a distribution program and a feature extraction program, and the purpose that source data exerts the maximum value is achieved; the AI model training area is composed of a feature library, a sample library, algorithm factors, a data set, a model architecture and computing resources; and the actual combat area comprises model racking, a model result set and business application. According to the invention, assembling, feature extraction, training and result output operations are carried out on the police service source data needing to be analyzed through an automation technology, the efficiency of the police service source data is realized to the greatest extent, second-level analysis of the data is realized through a data containerization parallel input and output technology, and the requirement of police service actual combat enabling is met.

Description

Technical Field

[0001] The present invention belongs to the field of police data analysis, and in particular to an analysis method based on the combination of police data containerization parallel throughput and AI model. Background Art

[0002] By receiving, collecting and extracting police source data, assembling the source data into standardized metadata through ETL, and inputting it into the actual combat area, evaluation area and AI model training area through parallel distribution and extraction, the results of the "three areas" are interwoven and used to achieve the goal of maximizing the efficiency of police data.

[0003] Based on the needs of police big data to assist in actual combat, the development and deployment method of AI model adaptation for actual combat applications is adopted. Police source data is collected externally, aggregated into data containers for unified management, unified strategy distribution, and a unified data feature extraction, model training, and result set application platform is built to establish an integrated actual combat operation mechanism from police data generation to assembly, extraction, distribution, training, and structured output.

[0004] In addition, through self-configured models, it supports the customization of model result sets based on actual combat needs, and combines the result set circulation application mechanism to achieve the full-process closed-loop practical management needs of real-time monitoring, notification and early warning, rapid disposal, tracking and tracing, and transmit the result set application backflow data to the model training area for recycling, iterate the model algorithm sample library, and improve the model accuracy rate.

[0005] Police data analysis model (analysis model types are shown in Table 1) 1. Abnormal landing point Analyze the registration status of the police contact person’s location 2. Analysis of Minors’ Mental Health Analyze the mental health indicators of minors and output the thresholds that are likely to cause malicious incidents 3. Warning to relevant persons Risk assessment of persons of concern and output of early warning records 4. Risk Assessment of Civil to Criminal Cases Comprehensive risk assessment of civil-to-criminal incidents, output of risk thresholds 5. Analysis of abnormal behavior of minors Analyze the behavior data of minors, provide labeled scenario applications, and output warnings, comprehensive evaluation and reminders Table 1 1. Data island problem: Data islands exist between different departments or systems, which makes it difficult to share and integrate data, affecting the coordination and efficiency of police work.

[0006] 2. Varying data quality: The quality of police data varies and may contain errors, omissions or incompleteness, which reduces the reliability and validity of the data.

[0007] 3. Information overload: Massive amounts of police data may lead to information overload, making it difficult for police officers to obtain useful information, affecting work efficiency.

[0008] 4. Privacy and security challenges: Police data involves personal privacy and sensitive information. How to balance efficiency and privacy security in data use and sharing is a challenge.

[0009] 5. Backward technology: Some police agencies may lack advanced technical equipment and professionals, resulting in relatively backward data management and analysis capabilities. Summary of the invention

[0010] In order to solve the defects and shortcomings existing in the above-mentioned prior art, the present invention provides an analysis method based on the combination of data containerization parallel throughput and AI model, which can realize operations such as monitoring, analysis, diagnosis and calculation of data through automation technology, improve the efficiency of data utilization, and can predict potential risk points of social governance by analyzing historical data and behavior patterns, and give early warning reminders to avoid the occurrence of disaster events; it can automatically adjust the model algorithm configuration and optimize the result set output performance by analyzing the model result set and the triggering event situation, thereby improving the robustness and efficiency of the system.

[0011] The technical solution adopted by the present invention to solve the problem is: An analysis method based on the combination of police data containerized parallel throughput and AI models, which enables massive police source data to be received, collected, extracted, cleaned, converted, decoupled, deduplicated, converted, and verified through data containers to form standard structure metadata; and the standard structure metadata is distributed in parallel to the practical layer, evaluation layer, and AI model inspection pool through distribution programs and feature extraction programs to achieve the goal of maximizing the value of source data.

[0012] Preferably, the police source data: an object-oriented data source, is determined by actual combat needs and AI model algorithm factors and feature libraries.

[0013] Preferably, the data container obtains police source data by multiple means of receiving, collecting and extracting, and forms standard structure metadata after ETL processing.

[0014] Preferably, the standard structure metadata: outputs the data in parallel through a distribution program and an AI model feature extraction program.

[0015] Preferably, the practical area includes model shelves, model result sets and business applications, and the business application scenarios can be configured as needed.

[0016] Preferably, the practical area: receives data from data containers in parallel, combines AI models to output model results in seconds, forms a practical result set, and realizes a fully closed-loop circulation of supporting business applications; for models not needed for practical use, temporary sealing operations can be performed to support secondary listing at a later time.

[0017] Preferably, the evaluation area includes model application, test result set and model evaluation. Models that pass the evaluation results can be put on the practical area for practical application.

[0018] Preferably, the evaluation area: receives data from the data container in parallel, combines the AI ​​model to output the test results, and performs model evaluation. After the evaluation is passed, it can be put on the practical area for practical application. Models that fail the evaluation are rolled back to the AI ​​model training area for secondary training.

[0019] Preferably, the AI ​​model training area includes a model feature library, algorithm factors, a sample library, a data set, and basic supporting model structures and computing resources, so as to achieve the extraction of standard structure metadata and model library output in seconds.

[0020] The present invention can realize the operations such as assembly, feature extraction, training, and output of the police source data that need to be analyzed through automation technology, so as to maximize the efficiency of the police source data, and realize data analysis in seconds through data containerization and parallel input / output technology, thus meeting the needs of practical police empowerment. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0022] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0023] like Figure 1 As shown in the figure, an analysis method based on the combination of police data containerized parallel throughput and AI model is used. Massive police source data is received, collected, extracted, cleaned, converted, decoupled, deduplicated, converted, and verified through data containers to form standard structure metadata; and the standard structure metadata is distributed in parallel to the actual combat layer, evaluation layer, and AI model training area through distribution programs and feature extraction programs to achieve the goal of maximizing the value of source data.

[0024] Among them, police source data is an object-oriented data source, which is determined by actual combat needs and AI model algorithm factors and feature libraries; Data container: obtains police source data through multiple means such as receiving, collecting, and extracting, and forms standard structure metadata after ETL processing; Standard structure metadata: Data is output in parallel through the distribution program and AI model feature extraction program.

[0025] Practical area: Receive data from data containers in parallel, combine with AI models to output model results in seconds, form a practical result set, and support business applications to achieve a fully closed-loop circulation; for models not needed for practical use, they can be temporarily sealed to support secondary listing later; Evaluation area: Receives data from data containers in parallel, combines AI models to output test results, and conducts model evaluation. Models that pass the evaluation can be put on the shelves in the practical area for practical application. Models that fail the evaluation are rolled back to the AI ​​model training area for secondary training.

[0026] AI model training area: includes model feature library, algorithm factors, sample library, data set and basic supporting model structure and computing resources, which can extract standard structure metadata and model library output in seconds.

[0027] The present invention is not limited to the above-mentioned specific implementation modes. A person skilled in the art can implement the present invention in various other specific implementation modes based on the contents disclosed in the present invention. Therefore, any design that adopts the design structure and concept of the present invention and makes some simple changes or modifications shall fall within the scope of protection of the present invention.

Claims

1. An analysis method based on the combination of police data containerization parallel throughput and AI model, characterized by: Massive police source data is received, collected, and extracted through data containers, and cleaned, converted, decoupled, deduplicated, converted, and verified to form standard structure metadata; the standard structure metadata is distributed in parallel to the practical layer, evaluation layer, and AI model inspection pool through distribution programs and feature extraction programs to achieve the goal of maximizing the value of source data.

2. According to claim 1, an analysis method based on the combination of police data containerization parallel throughput and AI model is characterized by: The police source data: an object-oriented data source, determined by actual combat needs and AI model algorithm factors and feature libraries.

3. According to claim 1, an analysis method based on the combination of police data containerization parallel throughput and AI model is characterized by: The data container obtains police source data through multiple means of receiving, collecting, and extracting, and forms standard structure metadata after ETL processing.

4. According to claim 1, the analysis method based on the combination of police data containerization parallel throughput and AI model is characterized by: The standard structure metadata: outputs data in parallel through a distribution program and an AI model feature extraction program.

5. According to claim 1, the analysis method based on the combination of police data containerization parallel throughput and AI model is characterized by: The practical area includes model listing, model result sets and business applications, and the business application scenarios can be configured as needed.

6. The analysis method based on the combination of police data containerization parallel throughput and AI model according to claim 5 is characterized by: The practical area: receives data from data containers in parallel, combines AI models to output model results in seconds, forms a result set that can be put into practical use, and realizes a fully closed-loop circulation through supporting business applications; for models that are not needed for practical use, temporary sealing operations can be performed to support secondary listing at a later time.

7. The analysis method based on the combination of police data containerization parallel throughput and AI model according to claim 1 is characterized by: The evaluation area includes model application, test result set and model evaluation. Models that pass the evaluation results can be put on the practical area for practical application.

8. The analysis method based on the combination of police data containerization parallel throughput and AI model according to claim 7 is characterized by: The evaluation area: receives data from the data container in parallel, combines the AI ​​model to output the test results, and performs model evaluation. After the evaluation passes, the model can be put on the shelf in the practical area for practical application. Models that fail the evaluation are rolled back to the AI ​​model training area for secondary training.

9. The analysis method based on the combination of police data containerization parallel throughput and AI model according to claim 1 is characterized by: The AI ​​model training area includes a model feature library, algorithm factors, a sample library, a data set, and basic supporting model structures and computing resources, which can extract standard structure metadata and model library output in seconds.

Images