A social security case behavior determination method, device, equipment and medium

Abstract

The present application relates to the field of Bayesian networks, and discloses a social security case criminal behavior determination method, device, equipment and medium, which can construct an evidence element system and evidence structure attribute according to the characteristics of a social security case, and construct the correlation between the evidence and the behavior in the social security case. Based on the evidence element system, the evidence structure attribute and the correlation, a Bayesian evidence reasoning network of the social security case is constructed. The prior condition probability table of the Bayesian evidence reasoning network is generated based on the coordination coefficient method and the weighted average method. The evidence information obtained by investigating the scene of the social security case is input into the Bayesian evidence reasoning network, so that the Bayesian evidence reasoning network determines the target behavior information and the corresponding occurrence probability according to the prior condition probability table and the evidence information. The present application can effectively reduce the consumption of human resources of case investigators and enhance the intelligent investigation level of social security cases.

Description

Technical Field

[0001] This invention relates to the field of Bayesian networks, and more particularly to a method, apparatus, device, and medium for determining criminal behavior in social security cases. Background Technology

[0002] With the development of science and technology, the application scope of intelligent auxiliary decision-making in criminal investigation and crime analysis is constantly expanding.

[0003] The investigation and management of social security cases is one of the important areas of social security research. Social security cases are highly harmful, occur frequently, and often have serious consequences and wide-ranging impacts.

[0004] The crime scene of a social security case often contains a variety of information that is helpful for investigation. Traditional investigation methods are mainly carried out by investigators based on experience and on-site information, which consumes a lot of human resources and has a low level of intelligence. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, and medium for determining criminal behavior in social security cases. It addresses the shortcomings of related technologies where investigations are mainly conducted by investigators based on experience and on-site information, resulting in high human resource consumption and low intelligence. The invention constructs a Bayesian inference network and determines criminal behavior in social security cases based on the Bayesian inference network, effectively reducing human resource consumption and enhancing the level of intelligent investigation of social security cases.

[0006] In a first aspect, the present invention provides a method for determining the criminal behavior in social security cases, comprising: Based on the characteristics of social security cases, we construct the evidence element system and evidence structure attributes for such cases, as well as the relationship between evidence and behavior in these cases. Based on the evidence element system, evidence structure attributes, and related relationships of the aforementioned social security cases, a Bayesian evidence reasoning network for the aforementioned social security cases is constructed. The prior conditional probability table of the Bayesian evidence reasoning network is generated based on the coordination coefficient method and the weighted average method. Evidence obtained through on-site investigation of the social security case is input into the Bayesian evidence reasoning network, so that the Bayesian evidence reasoning network can determine the target behavior information and the corresponding probability of occurrence based on the prior conditional probability table and the evidence information.

[0007] Optionally, the social security cases are collision-related cases; The evidence element system includes evidence objectives, factual layers, evidence categories, evidence forms, and evidence units in descending order of hierarchy. The evidence structure attributes include source attributes, morphological attributes, association attributes, functional attributes, and hierarchical attributes; The relationship between evidence and behavior includes the correspondence between behavior and evidence, the corroborative relationship between evidence, and the logical progression of the sequence of behaviors. The Bayesian evidence reasoning network for the social security case includes a behavioral hypothesis layer, a material exchange layer, and an evidence layer. Each of the behavioral hypothesis layer, the material exchange layer, and the evidence layer includes at least one node. The node in the behavioral hypothesis layer is the parent node of the node in the material exchange layer, and the node in the material exchange layer is the parent node of the node in the evidence layer. The prior conditional probability table includes the probability of an event occurring at the node given the probability of an event occurring at the parent node.

[0008] Optionally, generating the prior conditional probability table of the Bayesian evidence reasoning network based on the coordination coefficient method and weighted averaging includes: Collect initial probability estimates anonymously submitted by multiple experts; Each expert's anonymously submitted initial probability estimate is input into the created coordination coefficient model to determine the corresponding coordination coefficient, and then anonymously fed back to each expert so that each expert can revise the initial probability estimate based on the coordination coefficient. The process involves receiving a revised probability estimate from each anonymous expert feedback, inputting the revised probability estimate from each anonymous expert feedback into the coordination coefficient model to determine a new coordination coefficient, and iterating until a set condition is met to obtain the final probability estimate for each expert. A weighted average is calculated for the final probability estimates of each expert to obtain the prior conditional probability table of the Bayesian evidence reasoning network.

[0009] Optionally, the coordination coefficient model is: ; Where W is the coordination coefficient, ranging from 0 to 1, with higher values ​​indicating greater consistency; m is the number of experts; n is the number of indicators; and S is the sum of the squares of the differences between the average grade of each indicator and the overall average grade of all indicators. The value also ranges from 0 to 1, with higher values ​​indicating greater consistency.

[0010] Optionally, the weighted average calculation of the final probability estimates for each expert to obtain the prior conditional probability table of the Bayesian evidence inference network includes: Each expert is assigned a corresponding weight based on the differences in their authority, experience, and / or area of ​​expertise. The final probability estimate of each expert is input into the created weighted average calculation model to obtain the prior conditional probability table of the Bayesian evidence reasoning network. The weighted average calculation model is as follows: ; It is a weighted composite score based on probability. is the score given by the i-th expert, and n is the total number of experts. It is the weight of the i-th expert.

[0011] Optionally, the step of inputting the evidence information obtained through on-site investigation of the social security case into the Bayesian evidence reasoning network, so that the Bayesian evidence reasoning network determines the corresponding target behavior information based on the prior conditional probability table and the evidence information, includes: The evidence information is input into the Bayesian evidence reasoning network, so that the Bayesian evidence reasoning network: determines the corresponding material exchange information based on the evidence information and the prior conditional probability table, and determines the target behavior information and the corresponding probability of occurrence based on the material exchange information and the prior conditional probability table.

[0012] Optionally, after determining the target behavior information and the corresponding probability of occurrence, the method further includes: Based on the target behavior information and the corresponding probability of occurrence, a matching exploration suggestion is searched among the multiple exploration suggestions created, and output to the technical personnel so that the technical personnel can investigate the relevant site according to the exploration suggestion and obtain new evidence information. The new evidence information is input into the Bayesian evidence reasoning network so that the Bayesian evidence reasoning network can determine new behavioral information and corresponding occurrence probabilities based on the prior conditional probability table and the new evidence information.

[0013] Secondly, the present invention provides a device for determining criminal behavior in social security cases, comprising: The first construction unit is used to construct the evidence element system and evidence structure attributes of the social security case based on the characteristics of the social security case, as well as to construct the relationship between evidence and behavior in the social security case. The second construction unit is used to construct a Bayesian evidence reasoning network for the social security case based on the evidence element system, evidence structure attributes, and the correlation relationships of the social security case. The generation unit is used to generate the prior conditional probability table of the Bayesian evidence reasoning network based on the coordination coefficient method and the weighted average method. The determining unit is used to input the evidence information obtained by conducting an on-site investigation of the social security case into the Bayesian evidence reasoning network, so that the Bayesian evidence reasoning network can determine the target behavior information and the corresponding probability of occurrence based on the prior conditional probability table and the evidence information.

[0014] Thirdly, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the method for determining criminal behavior in social security cases as described in the first aspect or any corresponding embodiment.

[0015] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the method for determining criminal behavior in social security cases as described in the first aspect or any corresponding embodiment thereof.

[0016] This invention provides a method, apparatus, equipment, and medium for determining criminal behavior in social security cases. Based on the characteristics of social security cases, it constructs an evidentiary element system and evidentiary structural attributes, as well as the correlation between evidence and behavior in social security cases. Based on the evidentiary element system, evidentiary structural attributes, and correlations of social security cases, a Bayesian evidence reasoning network for social security cases is constructed. A priori conditional probability table for the Bayesian evidence reasoning network is generated based on the coordination coefficient method and weighted average method. Evidence information obtained through on-site investigation of social security cases is input into the Bayesian evidence reasoning network, enabling it to determine target behavior information and corresponding probabilities of occurrence based on the priori conditional probability table and the evidence information. This invention can construct a Bayesian reasoning network based on the characteristics of social security cases, and achieve intelligent determination of criminal behavior in social security cases based on the Bayesian reasoning network and on-site information, effectively reducing human resource consumption and enhancing the intelligent investigation level of social security cases. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating a method for determining criminal behavior in social security cases, provided as an embodiment of the present invention; Figure 2A schematic diagram of an evidentiary element system for social security cases provided in an embodiment of the present invention; Figure 3 A schematic diagram of an evidence reasoning model for social security cases provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the first-stage model input for a collision case provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the first-stage behavior layer output of a collision case provided by an embodiment of the present invention; Figure 6 This is a schematic diagram illustrating the output of a first-stage investigation suggestion for a collision case, provided by an embodiment of the present invention. Figure 7 This is a schematic diagram of the second-stage model input for a collision case provided in an embodiment of the present invention; Figure 8 This is a schematic diagram of the second-stage behavior layer output of a collision case provided by an embodiment of the present invention; Figure 9 This is a schematic diagram of a device for determining criminal behavior in social security cases, provided in an embodiment of the present invention. Figure 10 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0020] In related technologies, the crime scene of social security cases often contains diverse evidentiary elements, and the relationships between these elements are complex. Traditional investigation and case-solving methods mainly rely on the experience of investigators, which is difficult to adapt to the needs of investigation and case-solving under the new circumstances.

[0021] Utilizing computer-aided models to assist in police hazard diagnosis and case analysis is a new research trend. This approach makes the analysis process traceable and transparent, providing better decision support for crime analysis experts and reducing wrongful convictions. Simultaneously, the semi-automation and automation of case analysis can improve efficiency and save on investigation costs. In conclusion, research on evidence reasoning methods in social security cases has significant theoretical and practical implications.

[0022] The following is combined with Figures 1-8 This invention describes a method for determining criminal behavior in social security cases.

[0023] like Figure 1 As shown in the figure, this embodiment proposes a first method for determining the criminal behavior in social security cases, which may include the following steps: S101. Based on the characteristics of social security cases, construct an evidentiary element system and evidentiary structural attributes for social security cases, as well as construct the relationship between evidence and behavior in social security cases.

[0024] Specifically, this embodiment can construct an evidentiary element system for social security cases. Evidence elements are the consequences of the criminal act in a case; they are formed by the evolution of objective elements under the drive of the act, and their forms of expression vary from case to case, but their evolutionary mechanisms and laws are the same. The evidentiary element system for social security cases is a hierarchical logical structure from abstract to concrete, from macro to micro. It ensures the comprehensiveness and directionality of evidence collection, organization, and analysis.

[0025] Optionally, public security cases are collision-related cases; The evidence element system includes evidence objectives, factual layers, evidence categories, evidence forms, and evidence units in descending order of hierarchy; the evidence structure attributes include source attributes, morphological attributes, relational attributes, functional attributes, and hierarchical attributes; the relationship between evidence and behavior includes the correspondence between behavior and evidence, the synergistic corroboration relationship between evidence, and the logical progression of the behavior sequence.

[0026] like Figure 2 As shown, the evidence element system can be divided into the following four levels: First level: Evidence objective.

[0027] This is the highest level of the system, defining the ultimate goal of all evidentiary work: to reconstruct the legal facts of the case. Based on this core, three main objectives of proof are derived: Subjective Facts: proving the identity, criminal responsibility, and relationships of the individuals involved. Behavioral Facts: proving the occurrence, process, means, and methods of the criminal act. Resulting Facts: proving the social harm caused by the criminal act.

[0028] Second level: Evidence category.

[0029] To achieve the above objectives, evidence can be categorized into three main types, each corresponding to a specific goal. Subjective evidence proves who was involved, including identity information, criminal records, and evidence of communication and intent in a joint crime. Behavioral evidence proves how the crime was committed. This is the core of the case, including evidence of premeditation, preparation, and execution. Outcome evidence proves the consequences, including records of assessments, evaluations, and investigations related to personal injury, property damage, and disruption of social order.

[0030] The third level: the form of evidence.

[0031] This level describes the specific form of evidence and serves as the carrier of evidence categories. It is mainly divided into: First-person evidence, evidence carried by human statements, such as witness testimonies, victim statements, and confessions and defenses of suspects. Second-person evidence, evidence carried by physical objects, such as tools used in the crime, stolen goods, and traces. Third-person documentary evidence, evidence carried by written records, such as contracts, invoices, and diaries. Fourth-person electronic evidence, evidence carried by digital information, such as communication records, emails, surveillance videos, and web logs. Fifth-person expert opinions, conclusions of examination and identification issued by professional institutions. Sixth-person investigation and inspection records, records made by investigators during the investigation and inspection of the scene, objects, and persons.

[0032] Fourth level: Evidence unit.

[0033] This is the most specific level in the system, referring to a single, smallest piece of evidence. For example, under the main evidence – witness testimony, a specific unit of evidence is the transcript of witness A's statement regarding the suspect's identity. Under the behavioral evidence – electronic data, a specific unit of evidence is the suspect's location information at the time of the incident retrieved from the server. Under the outcome evidence – expert opinions, a specific unit of evidence is the forensic injury assessment report issued by the judicial appraisal center.

[0034] Specifically, this embodiment can construct the evidentiary structure attributes of social security cases. The evidentiary structure of social security cases is a rigorous logical system aimed at comprehensively, objectively, and accurately reconstructing the facts. Its core attributes can be divided into five levels: **Source attribute:** Defines the original source of the evidence, such as physical evidence, documentary evidence, electronic data, witness testimony, etc., which is the foundation of the authenticity of the evidence. **Form attribute:** Describes the form in which the evidence exists, including physical form (such as physical objects) and data form (such as electronic records), determining the method of evidence collection and preservation. **Relationship attribute:** The logical connection between the evidence and the facts of the case, including direct evidence that directly proves key facts, and indirect evidence that needs to be combined with other evidence to form a chain of proof. **Functional attribute:** The specific role of the evidence in the proof system, such as subject evidence (proving the persons / units involved), behavioral evidence (proving criminal behavior), and result evidence (proving harmful consequences), etc. **Hierarchical attribute:** The hierarchical relationship of the evidence in constructing a complete proof system, from basic evidentiary materials to evidence groups proving individual facts to be proven, ultimately forming a complete evidentiary system that collectively points to a single conclusion.

[0035] Specifically, this embodiment can construct the correlation between evidence and behavior in social security cases. This correlation is mainly reflected in three levels: First, the correspondence between behavior and evidence: each independent criminal act (such as preparation, execution, or harmful result) must be proven by corresponding evidence. For example, the act of purchasing a knife corresponds to physical evidence (the knife), transaction records (documentary evidence), and seller testimony; the act of posting threatening remarks corresponds to electronic data (chat logs, posts). Second, the synergistic corroboration relationship between evidence: the probative value of a single piece of evidence is limited; evidence from different sources and in different forms is needed to support each other and form a closed loop of evidence. For example, proving that A committed an attack at the crime scene requires eyewitness testimony (verbal evidence), on-site surveillance video (audiovisual materials), and the victim's biological traces found on A's clothing (physical evidence, expert opinion) to corroborate each other. Third, the logical progression of the behavioral sequence: the evidence must ultimately connect to form a complete behavioral process. From motive evidence (diaries, communication records) to preparatory evidence (purchasing tools, reconnaissance), then to execution evidence (on-site traces, surveillance footage), and finally to outcome evidence (injury and death assessment, property damage assessment), an irreversible logical chain is formed from cause to effect, thus fully reproducing the whole picture of the case.

[0036] S102. Based on the evidence element system, evidence structure attributes, and correlation relationships of social security cases, construct a Bayesian evidence reasoning network for social security cases.

[0037] Specifically, this embodiment can further construct a Bayesian inference network for evidence reasoning in social security cases based on the evidence-behavioral correlation network. A Bayesian network is a probabilistic graphical model that uses a directed acyclic graph to represent a set of variables and their conditional dependencies. In this embodiment, key behaviors and behavioral elements in the case can be used as nodes, evidence as observation nodes, and causal relationships can be represented by directed edges.

[0038] Optionally, the Bayesian evidence reasoning network for social security cases includes a behavioral hypothesis layer, a material exchange layer, and an evidence layer. Each of the behavioral hypothesis layer, the material exchange layer, and the evidence layer includes at least one node. The node in the behavioral hypothesis layer is the parent node of the node in the material exchange layer, and the node in the material exchange layer is the parent node of the node in the evidence layer.

[0039] like Figure 3 As shown, the inventors' research revealed that the behavioral hypothesis section is the starting point of the research, that is, the initial theoretical conception proposed for a certain behavior or phenomenon, which forms the basis for subsequent analysis. Next, the exchange of matter constitutes the core mechanism, describing how behavior is realized through the flow and transformation of matter or information, and is the key link connecting the hypothesis and the evidence. Finally, the evidence information represents empirical data obtained through observation or experimentation, used to verify whether the behavioral hypothesis is valid and whether its implementation mechanism is effective.

[0040] S103. Generate the prior conditional probability table of Bayesian evidence reasoning network based on the coordination coefficient method and weighted average method.

[0041] Specifically, this embodiment can determine the prior conditional probability table of the Bayesian evidence reasoning network.

[0042] Optionally, the prior conditional probability table includes the probability of an event occurring at a node given the probability of the event occurring at its parent node.

[0043] Optionally, step S103 includes: Collect initial probability estimates anonymously submitted by multiple experts; Each expert's anonymously submitted initial probability estimate is input into the created coordination coefficient model to determine the corresponding coordination coefficient, and then anonymously fed back to each expert so that each expert can revise the initial probability estimate based on the coordination coefficient. The algorithm receives the corrected probability estimate from each expert's anonymous feedback, inputs the corrected probability estimate from each expert's anonymous feedback into the coordination coefficient model to determine the new coordination coefficient, and iterates until the set conditions are met to obtain the final probability estimate for each expert. The final probability estimates of each expert are weighted and averaged to obtain the prior conditional probability table of the Bayesian evidence inference network.

[0044] Optionally, the coordination coefficient model is as follows: ; Where W is the coordination coefficient, ranging from 0 to 1, with higher values ​​indicating greater consistency; m is the number of experts; n is the number of indicators; and S is the sum of the squares of the differences between the average grade of each indicator and the overall average grade of all indicators. The value also ranges from 0 to 1, with higher values ​​indicating greater consistency.

[0045] Optionally, the final probability estimates for each expert are weighted and averaged to obtain the prior conditional probability table of the Bayesian evidence inference network, including: Each expert is assigned a corresponding weight based on their authority, experience, and / or area of ​​expertise. The final probability estimate of each expert is input into the created weighted average calculation model to obtain the prior conditional probability table of the Bayesian evidence reasoning network. The weighted average calculation model is as follows: ; It is a weighted composite score based on probability. is the score given by the i-th expert, and n is the total number of experts. It is the weight of the i-th expert.

[0046] In practical applications, this embodiment requires determining the network parameters in the logistic Bayesian inference network structure, i.e., the prior conditional probability table of the Bayesian inference network. Since real-world investigative case data is difficult to obtain, expert experience is relied upon to acquire the probability parameters. The Delphi method is applied, allowing multiple experts to anonymously submit their probability estimates. The results are then aggregated and anonymously fed back to all experts. After observing the distribution of opinions from the group, experts can revise their views and proceed to the next round of estimation. After several iterations, a convergent result with higher consensus is reached. Finally, the final results from multiple experts are weighted (based on expert authority) to derive the final probability parameter values.

[0047] When experts differ in their authority, experience, or area of ​​expertise, assigning different weights to different experts yields more scientific results.

[0048] Specifically, this embodiment can further calculate the conditional probabilities of child nodes in a Bayesian network. For any variable in a Bayesian Network (BN), and belonging to the probability space { The set of} The elements correspond one-to-one. n It is an integer greater than 1. The probability space of a Bayesian network consists of three parts. The first part is the space... , It is the structure or diagram of a Bayesian network, in which These are nodes (or variables) in a Bayesian network. It is the set of directed edges between variables in a network. The variables and directed edges together form a directed acyclic graph (DAG). The second part is the space. ,yes The space of all possible states The third part is space. It is all related to and Related The probability distribution.

[0049] definition It is the set of all possible Bayesian networks (BN). .in, It is an adjoint parameter A directed acyclic graph, .parameter , is a node A conditional probability table (CPT) is a table that lists the nodes in tabular form. Each state is related to its parent node The conditional probability of a state, i.e., P( When a node has no parent node, its conditional probability table is directly derived from the prior probability distribution P(…). Given. Any node The conditional probability table can be initialized based on the user's prior knowledge or learned through a training set. For the node space... Sampling is One observation at each node in the database. For node space of d Compilation of sub-sampling. When all values ​​of all variables are known, it is considered that... There are no missing values. This assumes that each sample is taken individually. They are independent of each other and have the same unknown distribution.

[0050] Bayesian network analysis is a mathematical model based on applying Bayesian rules to the obtained data. When variables are given... The value of the child node Bayes' theorem can be used to calculate variables. The posterior probability distribution is calculated using formula (1): .

[0051] in, Prior probability yes state ( The known probability distribution of (). Likelihood function , including with The instantiation conditional probability of contiguous sub-variables. When all instantiated variables are known... Therefore, there is The marginalization of observed variables, This explains the relationship between instantiated variables and... The relationship between all possible states is given by formula (2): .

[0052] in, It is a variable All p The first child node j The instantiation and value of each variable. The posterior probability is given by Representation, also known as The marginal probability is expressed as its confidence level, representing the probability of its occurrence given the evidence. Finally, can be It is deduced from the above formula (1).

[0053] S104. Input the evidence information obtained through on-site investigation of social security cases into the Bayesian evidence reasoning network so that the Bayesian evidence reasoning network can determine the target behavior information and the corresponding probability of occurrence based on the prior conditional probability table and the evidence information.

[0054] Optionally, step S104 includes: Evidence information is input into the Bayesian evidence reasoning network so that the Bayesian evidence reasoning network can: determine the corresponding material exchange information based on the evidence information and the prior conditional probability table, and determine the target behavior information and the corresponding probability of occurrence based on the material exchange information and the prior conditional probability table.

[0055] Optionally, in other methods for determining criminal behavior in social security cases proposed in this embodiment, after determining the target behavior information and the corresponding probability of occurrence, the method further includes: Based on the target behavior information and the corresponding probability of occurrence, a matching exploration suggestion is searched among the multiple exploration suggestions created, and output to the technical personnel so that they can investigate the relevant site according to the exploration suggestions and obtain new evidence information. New evidence information is input into the Bayesian evidence reasoning network so that the network can determine new behavioral information and corresponding probabilities of occurrence based on the prior conditional probability table and the new evidence information.

[0056] Specifically, after constructing the Bayesian inference network, this embodiment conducts application analysis on the inference model. A case is used as input. The basic facts of the case are: a bus accident occurred in a city, resulting in 4 deaths and 5 injuries.

[0057] Phase 1: Site inspection and investigation.

[0058] The specific evidence obtained from the investigation includes: Motorcycle 1 that was hit: The scratches were determined to have been caused by a collision with the bus; Motorcycle 2 that was hit: The scratches were determined to have been caused by a collision with the bus; Motorcycle 3: The scratches were determined to have been caused by a collision with the bus. The car that was hit: the scratches were determined to have been caused by a collision with the bus; The black sedan that was hit: the scratches were determined to have been caused by a collision with the bus. Bloodstains on the right side door of the driver's cab, the left side window of the steering wheel, and the front door armrest: DNA profile of suspect 1 was detected; The deceased were Chen, Lian, Lin, Wang, and Xu. The extracted bus video footage provides a relatively complete picture of the suspect driving his car into the pedestrians.

[0059] For the model input in the first stage, please refer to Figure 4 The behavioral layer results inferred by the Bayesian inference network can be found in [reference needed]. Figure 5 It should be noted that -1 in the location column indicates that the scenario may occur, but the location is currently unknown. The exploration suggestions provided by the first-stage model are listed below. Figure 6 middle.

[0060] Phase Two: Other Information.

[0061] Through the investigation of the relevant crime scenes (corresponding to the investigative suggestions output by the model in the previous stage, which are highlighted with an underline), the evidentiary information obtained in this stage includes: Suspect 1 did not conduct any reconnaissance or scouting activities, did not bring his own vehicle for the crime, did not retrieve the vehicle on-site, did not undergo any skills training, and did not handle the scene; Suspect 1 hijacked a bus and rammed it.

[0062] The model input information at this stage is as follows: Figure 7 As shown, the behavioral-level reasoning results given by the analysis are as follows: Figure 8 As shown.

[0063] By comparing the investigative suggestions given by the Bayesian inference network in the first stage with the evidence information obtained in the second stage, it can be found that the inference model can provide effective investigative suggestions. By comparing the inference results of the final model with the actual case situation, it can be concluded that the crime results reconstructed by the model inference are relatively accurate and have a basis in fact.

[0064] The method for determining criminal behavior in social security cases proposed in this embodiment can construct an evidentiary element system and evidentiary structural attributes for social security cases, as well as establish the correlation between evidence and behavior in social security cases, based on the characteristics of such cases. Based on the evidentiary element system, evidentiary structural attributes, and correlations, a Bayesian evidence reasoning network for social security cases is constructed. A priori conditional probability table for the Bayesian evidence reasoning network is generated using the coordination coefficient method and weighted averaging. Evidence information obtained through on-site investigation of social security cases is input into the Bayesian evidence reasoning network, enabling it to determine the target behavior information and its corresponding probability of occurrence based on the priori conditional probability table and the evidentiary information. This embodiment can construct a Bayesian reasoning network based on the characteristics of social security cases, and achieve intelligent determination of criminal behavior in social security cases based on the Bayesian reasoning network and on-site information, effectively reducing human resource consumption and enhancing the intelligent investigation level of social security cases.

[0065] like Figure 9 As shown in the figure, this embodiment proposes a device for determining criminal behavior in social security cases, which may include: The first construction unit 101 is used to construct the evidence element system and evidence structure attributes of social security cases based on the characteristics of social security cases, as well as to construct the relationship between evidence and behavior in social security cases. The second construction unit 102 is used to construct a Bayesian evidence reasoning network for social security cases based on the evidence element system, evidence structure attributes, and correlation relationships of social security cases. The generation unit 103 is used to generate a prior conditional probability table for the Bayesian evidence reasoning network based on the coordination coefficient method and the weighted average method. The determination unit 104 is used to input the evidence information obtained by conducting on-site investigations of social security cases into the Bayesian evidence reasoning network, so that the Bayesian evidence reasoning network can determine the target behavior information and the corresponding probability of occurrence based on the prior conditional probability table and the evidence information.

[0066] It should be noted that the processing procedures of the first building unit 101, the second building unit 102, the generation unit 103, and the determining unit 104, and their beneficial effects, can be referred to respectively. Figure 1 Steps S101 to S104 in the process will not be described again.

[0067] Optionally, public security cases are collision-related cases; The evidence element system includes, in descending order of hierarchy, evidence objectives, factual layers, evidence categories, evidence forms, and evidence units; The structural attributes of evidence include source attributes, morphological attributes, relational attributes, functional attributes, and hierarchical attributes; The relationship between evidence and behavior includes the correspondence between behavior and evidence, the corroborative relationship between evidence, and the logical progression of the sequence of behaviors. The Bayesian evidence reasoning network for social security cases includes a behavioral hypothesis layer, a material exchange layer, and an evidence layer. Each of the behavioral hypothesis layer, the material exchange layer, and the evidence layer includes at least one node. The node in the behavioral hypothesis layer is the parent node of the node in the material exchange layer, and the node in the material exchange layer is the parent node of the node in the evidence layer. The prior conditional probability table includes the probability of an event occurring at a node given the probability of the event occurring at its parent node.

[0068] Optionally, the generating unit 103 is also used for: Collect initial probability estimates anonymously submitted by multiple experts; Each expert's anonymously submitted initial probability estimate is input into the created coordination coefficient model to determine the corresponding coordination coefficient, and then anonymously fed back to each expert so that each expert can revise the initial probability estimate based on the coordination coefficient. The algorithm receives the corrected probability estimate from each expert's anonymous feedback, inputs the corrected probability estimate from each expert's anonymous feedback into the coordination coefficient model to determine the new coordination coefficient, and iterates until the set conditions are met to obtain the final probability estimate for each expert. The final probability estimates of each expert are weighted and averaged to obtain the prior conditional probability table of the Bayesian evidence inference network.

[0069] Optionally, the coordination coefficient model is as follows: ; Where W is the coordination coefficient, ranging from 0 to 1, with higher values ​​indicating greater consistency; m is the number of experts; n is the number of indicators; and S is the sum of the squares of the differences between the average grade of each indicator and the overall average grade of all indicators. The value also ranges from 0 to 1, with higher values ​​indicating greater consistency.

[0070] Optionally, the generating unit 103 is also used for: Each expert is assigned a corresponding weight based on their authority, experience, and / or area of ​​expertise. The final probability estimate of each expert is input into the created weighted average calculation model to obtain the prior conditional probability table of the Bayesian evidence reasoning network. The weighted average calculation model is as follows: ; It is a weighted composite score based on probability. is the score given by the i-th expert, and n is the total number of experts. It is the weight of the i-th expert.

[0071] Optionally, the determining unit 104 is also used for: Evidence information is input into the Bayesian evidence reasoning network so that the Bayesian evidence reasoning network can: determine the corresponding material exchange information based on the evidence information and the prior conditional probability table, and determine the target behavior information and the corresponding probability of occurrence based on the material exchange information and the prior conditional probability table.

[0072] Optionally, the above-mentioned device further includes: The input unit, after determining the target behavior information and its corresponding probability of occurrence, searches for a matching exploration suggestion among multiple created exploration suggestions based on the target behavior information and its corresponding probability of occurrence, and outputs it to the technical personnel so that the technical personnel can investigate the relevant site according to the exploration suggestion and obtain new evidence information; the new evidence information is then input into the Bayesian evidence reasoning network so that the Bayesian evidence reasoning network can determine the new behavior information and its corresponding probability of occurrence based on the prior conditional probability table and the new evidence information.

[0073] The device for determining criminal behavior in social security cases proposed in this embodiment can construct an evidentiary element system and evidentiary structural attributes for social security cases, as well as establish the correlation between evidence and behavior in social security cases, based on the characteristics of such cases. Based on the evidentiary element system, evidentiary structural attributes, and correlations of social security cases, a Bayesian evidence reasoning network for social security cases is constructed. A priori conditional probability table for the Bayesian evidence reasoning network is generated using the coordination coefficient method and weighted averaging. Evidence information obtained through on-site investigation of social security cases is input into the Bayesian evidence reasoning network, enabling it to determine the target behavior information and its corresponding probability of occurrence based on the priori conditional probability table and the evidence information. This embodiment can construct a Bayesian reasoning network based on the characteristics of social security cases, and intelligently determine criminal behavior in social security cases based on the Bayesian reasoning network and on-site information, effectively reducing human resource consumption and enhancing the intelligent investigation level of social security cases.

[0074] In this embodiment, the device for determining criminal behavior in social security cases is presented in the form of a functional unit. Here, a unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0075] This invention also provides a computer device having the above-described features. Figure 9 The device shown is for determining the criminal behavior in social security cases.

[0076] Please see Figure 10 The present invention provides a schematic diagram of the structure of a computer device according to an optional embodiment. The computer device includes one or more processors 10, a memory 20, and interfaces for connecting the various components, including high-speed interfaces and low-speed interfaces. The various components are interconnected via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, multiple processors and / or multiple buses can be used with multiple memories, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 10 Take a processor 10 as an example.

[0077] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0078] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0079] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function. The data storage area may store data created based on the use of the computer device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, which can be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0080] Memory 20 may include volatile memory, such as random access memory. Memory may also include non-volatile memory, such as flash memory, hard disk, or solid-state drive. Memory 20 may also include combinations of the above types of memory.

[0081] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0082] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0083] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for determining the criminal behavior in social security cases, characterized in that, include: Based on the characteristics of social security cases, we construct the evidence element system and evidence structure attributes for such cases, as well as the relationship between evidence and behavior in these cases. Based on the evidence element system, evidence structure attributes, and related relationships of the aforementioned social security cases, a Bayesian evidence reasoning network for the aforementioned social security cases is constructed. The prior conditional probability table of the Bayesian evidence reasoning network is generated based on the coordination coefficient method and the weighted average method. Evidence information obtained through on-site investigation of the social security case is input into the Bayesian evidence reasoning network, so that the Bayesian evidence reasoning network can determine the target behavior information and the corresponding probability of occurrence based on the prior conditional probability table and the evidence information.

2. The method according to claim 1, characterized in that, The social security cases mentioned are collision-related cases; The evidence element system includes evidence objectives, factual layers, evidence categories, evidence forms, and evidence units in descending order of hierarchy. The evidence structure attributes include source attributes, morphological attributes, association attributes, functional attributes, and hierarchical attributes; The relationship between evidence and behavior includes the correspondence between behavior and evidence, the corroborative relationship between evidence, and the logical progression of the sequence of behaviors. The Bayesian evidence reasoning network for the social security case includes a behavioral hypothesis layer, a material exchange layer, and an evidence layer. Each of the behavioral hypothesis layer, the material exchange layer, and the evidence layer includes at least one node. The node in the behavioral hypothesis layer is the parent node of the node in the material exchange layer, and the node in the material exchange layer is the parent node of the node in the evidence layer. The prior conditional probability table includes the probability of an event occurring at the node given the probability of an event occurring at the parent node.

3. The method according to claim 2, characterized in that, The generation of the prior conditional probability table for the Bayesian evidence reasoning network based on the coordination coefficient method and weighted averaging includes: Collect initial probability estimates anonymously submitted by multiple experts; Each expert's anonymously submitted initial probability estimate is input into the created coordination coefficient model to determine the corresponding coordination coefficient, and then anonymously fed back to each expert so that each expert can revise the initial probability estimate based on the coordination coefficient. The process involves receiving a revised probability estimate from each anonymous expert feedback, inputting the revised probability estimate from each anonymous expert feedback into the coordination coefficient model to determine a new coordination coefficient, and iterating until a set condition is met to obtain the final probability estimate for each expert. A weighted average is calculated for the final probability estimates of each expert to obtain the prior conditional probability table of the Bayesian evidence reasoning network.

4. The method according to claim 3, characterized in that, The coordination coefficient model is as follows: ； Where W is the coordination coefficient, which ranges from 0 to 1. The closer it is to 1, the higher the consistency. m is the number of experts, n is the number of indicators, and S is the sum of the squares of the differences between the average grade of each indicator and the total average grade of all indicators. It ranges from 0 to 1. The closer it is to 1, the higher the consistency.

5. The method according to claim 3, characterized in that, The weighted average calculation of the final probability estimates for each expert to obtain the prior conditional probability table of the Bayesian evidence inference network includes: Each expert is assigned a corresponding weight based on the differences in their authority, experience, and / or area of ​​expertise. The final probability estimate of each expert is input into the created weighted average calculation model to obtain the prior conditional probability table of the Bayesian evidence reasoning network. The weighted average calculation model is as follows: ； It is a weighted composite score based on probability. is the score given by the i-th expert, and n is the total number of experts. It is the weight of the i-th expert.

6. The method according to claim 2, characterized in that, The step of inputting evidence information obtained through on-site investigation of the social security case into the Bayesian evidence reasoning network, so that the Bayesian evidence reasoning network can determine the corresponding target behavior information based on the prior conditional probability table and the evidence information, includes: The evidence information is input into the Bayesian evidence reasoning network, so that the Bayesian evidence reasoning network: determines the corresponding material exchange information based on the evidence information and the prior conditional probability table, and determines the target behavior information and the corresponding probability of occurrence based on the material exchange information and the prior conditional probability table.

7. The method according to claim 6, characterized in that, After determining the target behavior information and the corresponding probability of occurrence, the method further includes: Based on the target behavior information and the corresponding probability of occurrence, a matching exploration suggestion is searched among the multiple exploration suggestions created, and output to the technical personnel so that the technical personnel can investigate the relevant site according to the exploration suggestion and obtain new evidence information. The new evidence information is input into the Bayesian evidence reasoning network so that the Bayesian evidence reasoning network can determine new behavioral information and corresponding occurrence probabilities based on the prior conditional probability table and the new evidence information.

8. A device for determining criminal behavior in social security cases, characterized in that, include: The first construction unit is used to construct the evidence element system and evidence structure attributes of the social security case based on the characteristics of the social security case, as well as to construct the relationship between evidence and behavior in the social security case. The second construction unit is used to construct a Bayesian evidence reasoning network for the social security case based on the evidence element system, evidence structure attributes, and the correlation relationships of the social security case. The generation unit is used to generate the prior conditional probability table of the Bayesian evidence reasoning network based on the coordination coefficient method and the weighted average method. The determining unit is used to input the evidence information obtained by conducting an on-site investigation of the social security case into the Bayesian evidence reasoning network, so that the Bayesian evidence reasoning network can determine the target behavior information and the corresponding probability of occurrence based on the prior conditional probability table and the evidence information.

9. A computer device, characterized in that, include: A memory and a processor are interconnected, the memory stores computer instructions, and the processor executes the computer instructions to perform the method for determining the criminal behavior in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the method for determining the criminal behavior in social security cases as described in any one of claims 1 to 7.

Images