Alarm condition prediction system

Abstract

The invention belongs to a network user behavior information analysis system, and designs an alarm condition prediction system in order to improve the accuracy of social risk prediction in advance, which comprises a prediction object confirmation module, an information acquisition module, an information processing module, a model matching module, a risk early warning module, a risk verification module, a risk analysis module and a prediction object optimization module. By determining a prediction object, collecting related information of the prediction object, processing the information, establishing a risk model, automatically matching the risk model, carrying out risk early warning, carrying out risk verification in real time, providing a risk report and optimizing the risk model, pre-prediction of an accident can be realized, and the occurrence rate of the accident is reduced.

Description

Technical Field

[0001] The present invention belongs to the field of network information analysis, and particularly relates to a system for predicting police situations by analyzing network user behavior information. Background Art

[0002] All security incidents occur for a reason, and actions are taken based on behaviors. How to predict in advance based on reasons and behavioral manifestations to avoid or reduce the occurrence of incidents. Summary of the Invention

[0003] To solve the above technical problems of predicting social security incidents, the object of the present invention is:

[0004] Design a police situation prediction system that analyzes social status information and people's behaviors, focuses on high-risk groups, establishes a risk model, conducts risk early warning, and provides early warning support for management departments in advance.

[0005] To achieve the above object of the invention, the following technical solutions are adopted:

[0006] A police situation prediction system includes a prediction object confirmation module, an information collection module, an information processing module, a model matching module, a risk early warning module, a risk verification module, a risk analysis module, and a prediction object optimization module.

[0007] Prediction object confirmation module: First, select the prediction object, and predict possible risks in advance based on the relevant information of the prediction object. The prediction objects with risks can be subdivided into various units. Preferably, they are subdivided into individual units, population units, real estate units, vehicle units, event units, item units, and organization units.

[0008] Information collection module: According to the selected prediction object, collect relevant information of the prediction object from various system platforms.

[0009] Information processing module: Process the collected information. Preferably, it is divided into static information units, dynamic information units, positive information units, negative information units, and behavior prediction units.

[0010] Model matching module: According to the processed information, establish risk models of different categories, input the relevant processed information, and the risk models are automatically matched.

[0011] Risk early warning module: According to the results of model matching, classify and rank the risks for warning events, and give warnings to relevant managers. Preferably, it can be divided into normal state warning units, potential risk warning units, and serious risk warning units.

[0012] Risk verification module: Track and feedback cases with risk warnings in real time to verify the accuracy and deviation of risk warnings. Preferably, it can be divided into a normal status verification unit, a potential risk verification unit, and a serious risk verification unit.

[0013] Risk analysis module: Summarize a risk report based on the event analysis of risk warnings and risk verifications to understand the operation of the system's risk warning. Preferably, it is divided into a normal status report unit, a potential risk report unit, and a serious risk report unit.

[0014] Prediction object optimization module: Optimize the risk model according to the risk report, enabling the system to continuously adapt to changes in social development factors, improving the accuracy of risk model warnings, and enhancing the vitality of the system. Preferably, it is divided into a normal status optimization unit, a potential risk optimization unit, and a serious risk optimization unit.

[0015] Advantages of the present invention:

[0016] An alarm prediction system can achieve pre-accident prediction and reduce the accident incidence rate by determining the prediction object, collecting relevant information of the prediction object, processing the information, establishing a risk model, automatically matching the risk model, giving risk warnings, verifying risks in real time, providing a risk report, and optimizing the risk model. Description of the drawings

[0017] Figure 1 Schematic diagram of the object processing steps of an alarm prediction system.

[0018] Figure 2 Schematic diagram of the object processing content of an alarm prediction system. Detailed implementation manners

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0020] An alarm prediction system includes a prediction object confirmation module 1, an information collection module 2, an information processing module 3, a model matching module 4, a risk warning module 5, a risk verification module 6, a risk analysis module 7, and a prediction object optimization module 8.

[0021] Prediction object confirmation module: First, select the prediction object, and pre-judge the possible risks based on the relevant information of the prediction object. The prediction objects with risks can be further divided into various units.

[0022] Optionally, each sub - unit can be monitored individually, or multiple sub - units can be selected for simultaneous monitoring.

[0023] Preferably, it is subdivided into personal unit, population unit, real - estate unit, vehicle unit, event unit, item unit, and organization unit.

[0024] Personal unit: mainly all information of an individual's identity file, including personal basic information (identity information, address, family relationship), educational information, work information, social information, past criminal records, etc.

[0025] Optionally, family relationship is a risk point for violent crimes. In society, many potential risk outbreak points are due to marital relationships, lover relationships, and economic relationships. Optionally, two newly - divorced people can be selected as prediction targets, analyze the reasons for divorce and predict their later acceptance willingness. Combining with their own situations, based on psychological and big - data analysis cases, analyze and predict the possible crime probability in the later stage and give early warnings.

[0026] Population unit: Classify the population. Different populations have different characteristics and different risk points. Optionally, monitor the population about to be demolished. Often, demolition leads to violent conflict events, and such populations can be included in the prediction targets.

[0027] Real - estate unit: As a heavy asset, a house can be an important clue for analysis and prediction.

[0028] Vehicle unit: As an asset and a means of transportation, many accidents are caused by vehicle collisions and implementations.

[0029] Event unit: Changes in events may cause new potential risks, such as new national government policies, company business changes, demolition events, social public opinions, etc.

[0030] Item unit: The special functions of items may cause some potential risks, such as dangerous goods, knives, dangerous raw materials, etc.

[0031] Organization unit: Some social organizations and groups may be high - risk outbreak areas and can be monitored and warned in a timely manner. For example, some organizations that handle membership cards are prone to running away with money, causing social unrest. However, if early warnings can be given and supervision can be carried out, the occurrence probability can be reduced.

[0032] Information collection module: According to the selected prediction targets, collect relevant information of the prediction targets from various system platforms.

[0033] The information collection is designed to access various system platforms, including public security systems, government affairs systems, social media platforms, search platforms, news platforms, police situation databases, financial systems, communication systems, online shopping platforms, historical data, public reporting platforms, transportation systems, enterprise information, and others.

[0034] The information collection design is connected to various system platforms. However, there are problems such as inconsistent data formats and different database types on various platforms. But in order to achieve a higher prediction accuracy rate, it is necessary to collect as much relevant information as possible and connect to other platforms.

[0035] The interaction design of the information collection module and other databases has an independent database docking unit, which realizes the interconnection and interoperability with various databases using existing device resources in the existing system, and also has the ability to dock with new databases that may appear in the future, effectively reducing the input cost and improving the utilization rate of system resources.

[0036] The database docking unit can be optionally configured with one or a combination of a database system docking unit, a data format conversion unit, and a database synchronization unit, and which unit to configure is determined according to actual requirements and applications.

[0037] Optionally, for the database system docking unit: If the database can be fully docked with the local database, then directly dock and exchange information, especially applicable to the docking between homologous and isomorphic database systems.

[0038] Optionally, for the data format conversion unit: If the database cannot be directly docked with the local database, but can be docked after format conversion, then first convert the data of the database and then complete the docking with the local database. Especially applicable to the docking between homologous heterogeneous and heterologous isomorphic database systems.

[0039] Optionally, for the database synchronization unit: If the access code database cannot be directly docked with the local database and cannot be docked after format conversion, then communicate with each other through the data synchronization unit. In a specific network scenario, when the database is not allowed to be docked with an external database, the database synchronization unit is used to synchronize data inside the external database environment at a specific time and under specific conditions, isolated from the local database; after synchronization is completed, then isolated from the external database and synchronized with the local database, avoiding direct communication between them.

[0040] The information processing module: processes the collected information. Preferably, it is divided into a static information unit, a dynamic information unit, a positive information unit, a negative information unit, and a behavior prediction unit.

[0041] The static information unit: refers to the static information of the prediction object, generally the characteristic information of the prediction object, such as the name, ID number, household register, educational experience, etc. of a person

[0042] The dynamic information unit: refers to the dynamic information of the prediction object, generally the information that is prone to change of the prediction object, such as the change of a person's location, travel information, work status, etc.

[0043] Positive information unit: It is the analysis of positive information about the prediction object, processes and analyzes positive information, divides positive information into different categories, and divides different categories into different levels according to different degrees. For example, it analyzes the expressions of people through video pictures, and analyzes the emotional state of people according to expressions and actions. The smile category level classification: level 1 for smiling, level 2 for laughing out loud, and level 3 for laughing uncontrollably.

[0044] Negative information unit: It is the analysis of negative information about the prediction object, processes and analyzes negative information, divides negative information into different categories, and divides different categories into different levels according to different degrees. For example, it analyzes the expressions of people through video pictures, and analyzes the emotional state of people according to expressions and actions. The sadness category level classification: level 1 for shedding tears, level 2 for crying loudly, and level 3 for wailing.

[0045] Both the positive information unit and the negative information unit can analyze the words, phrases, semantics, and emotional states in the information such as the articles published by the prediction object, the content of exchanges with related personnel, and the information released on social media, identify the tendencies of the words, phrases, semantics, and emotional states, judge whether it is positive information or negative information, and identify the degree of its bias. The degree of bias is classified according to the severity.

[0046] Behavior prediction unit: According to the information of the prediction object that has been collected, it conducts reasoning and prediction according to objective laws, actual situations, historical data, and research theories, and uses the information of reasoning and prediction as the basic analysis information for early warning. Optionally, a warning object is a person or a vehicle, where it is now, where it starts from, where it may go, what routes there are, and what things it may do on each route and what the consequences are. During the prediction process, once the prediction object is determined, the previous familiar routes of the object can be queried. The probability that the object takes familiar routes during the movement process is relatively high, so as to narrow the search scope and reduce the input of energy, and quickly find the target.

[0047] Model matching module: According to the processed information, different categories of risk models are established, and each category of risk model divides different risk levels. Input the relevant processed information, and the risk model automatically matches.

[0048] Optionally, an emotional breakdown revenge model: For example, what is the reason for a person's sadness? According to the collected information, it is known that it is because of the breakdown of the marriage relationship. The types of marriage breakdown are divided into peaceful break-up, reluctant break-up, court judgment break-up, etc. For example, in the case of court judgment break-up, what is the acceptance degree of both parties? Combining the personal growth environment, education level, ideological state, online search behavior, and behavior of purchasing dangerous goods of both parties, and combining historical data and research results, comprehensively judge the possibility of possible post-event revenge events.

[0049] Risk early warning module: According to the results of model matching, classify and rank the risks of early warning events, and give early warnings to relevant managers. Preferably, it can be divided into normal state early warning unit, potential risk early warning unit, and severe risk early warning unit.

[0050] Normal state early warning unit: According to the results of model matching, classify the predicted risk level of the prediction object as the normal state, judge within the normal state range, and continuously monitor in the follow-up.

[0051] Potential risk early warning unit: According to the results of model matching, classify the predicted risk level of the prediction object as the potential risk state, judge that the prediction object may perform harmful behaviors, and focus on monitoring in the follow-up.

[0052] Severe risk early warning unit: According to the results of model matching, classify the predicted risk level of the prediction object as the severe risk state, judge that the prediction object is very likely to perform harmful behaviors, and should give timely warnings and take measures to prevent dangerous behaviors as soon as possible.

[0053] The predicted normal state events, potential risk state events, and severe risk state events may not match the actual situation during the subsequent development of the situation, and there may be mutual conversion events. For example, normal state events may be converted into potential risk state events or severe risk state events, potential risk state events may be converted into normal state events or severe risk state events, and severe risk state events may be converted into normal state events or potential risk state events.

[0054] Risk verification module: Track and feedback the cases of risk early warning in real time, and verify the accuracy and deviation of risk early warning. Preferably, it can be divided into normal state verification unit, potential risk verification unit, and severe risk verification unit.

[0055] Risk verification is to compare and analyze the actual occurrence of the predicted event with the predicted result for the prediction object within a certain time range, and verify the accuracy, deviation, and deviation situation of the prediction object.

[0056] Risk analysis module: Summarize a risk report based on the analysis of the events of risk early warning and risk verification, and use it to master the operation of the system risk early warning. Preferably, it is divided into normal state report unit, potential risk report unit, and severe risk report unit.

[0057] Normal state report unit: Include predicted normal state events and the analysis of the conversion of normal state events into other events.

[0058] Potential risk report unit: Include predicted potential risk state events and the analysis of the conversion of potential risk state events into other cases.

[0059] Serious Risk Reporting Unit: It includes predicting serious risk status events and analyzing the transformation of serious risk status events into other cases.

[0060] Prediction Object Optimization Module: According to the risk report, optimize the risk model to make the system continuously adapt to the changes in social development factors, improve the accuracy of the risk model warning, and enhance the vitality of the system. Preferably, it is divided into a normal state optimization unit, a potential risk optimization unit, and a serious risk optimization unit.

[0061] Optimize the risk model based on the analysis results of the transformation of normal state events, potential risk state events, and serious risk state events into other events, and adjust the information and parameter settings of the risk model by combining big data statistical analysis and relevant research reports to improve the prediction accuracy.

[0062] Optionally, a warning case of a dangerous case caused by demolition belongs to a typical case: Generally, for the parties involved in the demolition case, they first express dissatisfaction, do not sign the demolition agreement, feedback their dissatisfaction to the relevant departments, and only retaliate against society with their harmful behaviors to express dissatisfaction if there is no solution.

[0063] Optionally, the system takes the demolition population as the prediction object, accesses the public security system, government affairs system, social media platform, search platform, communication system, online shopping platform, public reporting platform, transportation system, and enterprise information, collects personal information of this group, processes the information, and establishes a demolition risk model.

[0064] Optionally, classify the risk model levels.

[0065] Preferably, if the party has dissatisfaction, it is marked as a first-level risk;

[0066] Preferably, if the demolition agreement is not signed, it is marked as a second-level risk;

[0067] Preferably, if dissatisfaction is feedback to a single relevant department, it is marked as a third-level risk;

[0068] Preferably, if dissatisfaction is feedback to multiple relevant departments, it is marked as a fourth-level risk;

[0069] Preferably, if a single department fails to solve the problem for a long time, it is marked as a fifth-level risk; if multiple departments fail to solve the problem, the risk level gradually increases by one level.

[0070] Preferably, unconventional behaviors such as purchasing dangerous goods such as alcohol and controlled knives are marked as a sixth-level risk, and the greater the harm of purchasing dangerous goods, the higher the risk level;

[0071] Preferably, engaging in social work and assuming relevant social responsibilities, such as a bus driver, is marked as a seventh-level risk;

[0072] Preferably, in general, before committing a crime, the perpetrator will say goodbye to their relatives, friends or the general public. There is a possibility that they will express their inner emotions on social media and platforms, showing feelings of parting or insights. Combining communication and Internet-related behavior analysis, those with tendencies of misanthropy or violence are marked as level-eight risks.

[0073] Optionally, events with a risk level of four or below are considered normal state events. If the parties are dissatisfied and seek solutions from relevant departments, at this time, the parties have hope and there is a possibility of solving the problem. Generally, at this stage, the parties do not have the tendency to harm society. However, continuous monitoring is required. If the situation deteriorates, for example, if the relevant department delays in solving the problem for a long time and the parties have no way to solve it, have no hope but are still dissatisfied, reaching the level-five risk level, it can be transformed into a potential risk state event, and there may be harmful behaviors. When the early warning system detects that the parties have unconventional harmful behaviors such as purchasing alcohol and controlled knives, reaching a risk level of six or above, it can be transformed into a serious risk state event, and immediate early warning and relevant measures should be taken. The relevant departments should handle it as soon as possible.

[0074] Preferably, track the development of events according to the early warning situation, and adjust the early warning factors of the risk model through big data statistical analysis, scientific research and psychological research results to improve the accuracy of prediction.

[0075] Optionally, police cases caused by economic asset disputes are also a typical case type. Taking economic asset disputes as the prediction object, when the parties feel the risk of being harmed, they can apply for risk early warning monitoring. However, when the other party has not committed a crime and does not meet the conditions for judicial personal protection, the potential risk objects can be monitored, relevant information can be collected and processed, an economic asset dispute model can be established and matched. When it is found that the relevant personnel purchase dangerous goods or have abnormal emotional tendencies, they are classified according to the severity level, and risk early warning is carried out for the relevant personnel. The potential victims who may be invaded should be notified in a timely and proactive manner to pay attention to prevention, and psychological counseling and legal publicity should be carried out for the objects who may implement the infringement to reduce the possibility of committing crimes. Risk verification and risk analysis should be carried out according to the actual situation development, and the prediction object should be optimized according to the analysis results to improve the prediction system and the accuracy of prediction.

Claims

1. An alarm situation prediction system, characterized in that It includes a prediction object confirmation module, an information collection module, an information processing module, a model matching module, and a risk warning module.

2. The police situation prediction system according to claim 1, wherein The prediction object confirmation module first selects the prediction object, anticipates possible risks in advance based on the relevant information of the prediction object, and the prediction objects with risks can be subdivided into various units.

3. A police situation prediction system according to claim 1, characterized in that The information collection module collects the relevant information of the prediction object from various system platforms according to the selected prediction object.

4. A police situation prediction system according to claim 1, characterized in that The information processing module processes the collected information.

5. A police situation prediction system according to claim 1, characterized in that The model matching module establishes risk models of different categories according to the processed information, inputs the relevant processed information, and the risk models are automatically matched.

6. The police situation prediction system according to claim 1, characterized in that The risk warning module classifies and ranks the risks for warning events according to the results of model matching, and warns the relevant managers.

7. An alarm situation prediction system according to claim 1, characterized in that, Further set up a risk verification module to track and feedback the cases of risk warnings in real time to verify the accuracy and deviation of risk warnings.

8. A police situation prediction system according to claim 1, characterized in that, Further set up a risk analysis module to summarize a risk report based on the analysis of risk warnings and risk verification events, which is used to master the operation of the system's risk warning.

9. The police situation prediction system according to claim 1, wherein Further set up a prediction object optimization module: according to the risk report, optimize the risk model parameters and monitoring data, so that the system can continuously adapt to the changes of social development factors, improve the accuracy of risk model warnings, and improve the vitality of the system.

10. A police situation prediction system according to claim 5, characterized in that The model matching module includes an emotional breakdown revenge model and a demolition risk model. An emotional breakdown revenge model judges a person's sadness, and learns from the collected information that it is due to the breakdown of the marriage relationship. The types of marriage breakdown are divided into peaceful break-up, reluctant break-up, court judgment break-up, etc. The acceptance of the court judgment by both parties, combined with the personal growth environment, education level, ideological state, online search behavior, and behavior of purchasing dangerous goods of both parties, combined with historical data and research results, comprehensively judges the possibility of possible revenge events afterwards. A demolition risk model: If the parties have dissatisfaction, it is marked as a first-level risk; if the demolition agreement has not been signed, it is marked as a second-level risk. Feedbacking dissatisfaction to a single relevant department is marked as a third-level risk. Feedbacking dissatisfaction to multiple relevant departments is marked as a fourth-level risk. If a single department fails to solve the problem for a long time, it is marked as a fifth-level risk. If multiple departments fail to solve the problem, the risk level gradually increases by one risk level. Unconventional behaviors such as purchasing dangerous goods such as alcohol and controlled knives are marked as a sixth-level risk, and the greater the harm of purchasing dangerous goods, the higher the risk level increases; being engaged in social work and assuming relevant social responsibilities is marked as a seventh-level risk; having a tendency to be world-weary or violent is marked as an eighth-level risk. Regarding events with a risk level of four or below as normal state events, the parties are dissatisfied and seek solutions from relevant departments. At this time, the parties have hope and there is a possibility of solving the problem. Generally, at this stage, the parties do not have the tendency to harm society. However, continuous monitoring is required. If the situation deteriorates, for example, if the relevant department delays in solving the problem for a long time, the parties have no way to solve the problem, have no hope but are dissatisfied, reaching the fifth-level risk level, it can be transformed into a potential risk state event, and there may be harmful behaviors. When the early warning system detects that the parties have engaged in unconventional harmful behaviors such as purchasing alcohol and controlled knives, and the risk reaches level six or above, it can be transformed into a serious risk state event, which requires immediate warning and the adoption of relevant measures. The relevant departments should handle it as soon as possible; Track the development of the event according to the early warning situation, and adjust the early warning factors of the risk model through big data statistical analysis, scientific research, and psychological research results to improve the accuracy of prediction.

Images