Crime analysis method and device and related product
By performing Booleanization on the case details and constructing hard and soft constraints, and using the MAXSAT solver to analyze the case elements and grade variables, the complex interactive relationships in sentencing were resolved, achieving clear sentencing reasoning and efficient and accurate analysis.
Patent Information
- Application Number
- CN202511123425.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-11
- Publication Date
- 2025-11-18
AI Technical Summary
Existing technologies are insufficient to effectively consider the complex interactions between case details in sentencing, resulting in a lack of clear reasoning and interpretability in the sentencing process.
By performing Booleanization on the case to be analyzed, hard and soft constraints are constructed. The MAXSAT solver is then used to analyze multiple case element variables and grade variables to obtain weights and values, and the analysis conclusions of the target case are determined.
It systematically represents the complex interactions between case element variables and grade variables, provides a complete formal reasoning process, improves the interpretability and transparency of case analysis, assists judicial personnel in accelerating analysis, and improves efficiency and accuracy.
Smart Images
Figure CN120975981A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of case analysis, in particular to a case analysis method and device and related products. BACKGROUND
[0002] In judicial practice, the basic point sentencing mode is usually used to realize sentencing of a case. Specifically, a basic sentence is first formulated based on the case situation, then a sentencing increase / decrease ratio is determined according to the formulated rules, and the determined increase / decrease ratios are simply superimposed to adjust the basic sentence to determine the sentencing result of the case. However, the basic point sentencing mode in the related technical solution is easy to ignore the complex interaction between the case situations, and it is difficult to provide a clear reasoning process.
[0003] Therefore, how to better analyze the case so as to consider the complex interaction between the case and improve the explainability of the case analysis is an important problem for those skilled in the art. SUMMARY
[0004] Based on the above problems, the present application provides a case analysis method, device and related products to better analyze the case, so as to consider the complex interaction between the case and improve the explainability of the case analysis.
[0005] The embodiments of the present application disclose the following technical solutions:
[0006] The first aspect of the present application provides a case analysis method. The case analysis method comprises:
[0007] Boolean processing is performed on a case to be analyzed to obtain a plurality of different case element variables and a plurality of different case level variables, wherein the case element variable represents case information existing in the case to be analyzed, and the case level variable represents a case analysis candidate conclusion corresponding to the case to be analyzed;
[0008] Based on the plurality of different case element variables and the plurality of different case level variables, a plurality of different hard constraints and a plurality of different soft constraints are constructed, wherein the hard constraint indicates that the case to be analyzed is related to the case level variable, and the soft constraint indicates the weight corresponding to the combination of the case element variable and the case level variable;
[0009] Based on the plurality of different case element variables and the plurality of different case level variables, the plurality of different hard constraints and the plurality of different soft constraints are analyzed and processed to obtain the weight and value corresponding to the case level variable when the plurality of different case level variables are respectively assigned to the true value, wherein the plurality of different hard constraints are all analyzed as the true value;
[0010] The analysis conclusion obtaining unit is configured to obtain a target case analysis conclusion corresponding to the case to be analyzed based on a case level variable corresponding to a weight sum value with a highest value.
[0011] The second aspect of the present application provides a case analysis device. The case analysis device comprises:
[0012] A Boolean processing unit is configured to perform Boolean processing on the case to be analyzed to obtain a plurality of different case element variables and a plurality of different case level variables, wherein the case element variables represent case information existing in the case to be analyzed, and the case level variables represent case analysis candidate conclusions corresponding to the case to be analyzed.
[0013] A constraint construction obtaining unit is configured to construct a plurality of different hard constraints and a plurality of different soft constraints based on the plurality of different case element variables and the plurality of different case level variables, wherein the hard constraints indicate that the case to be analyzed is related to the case level variables, and the soft constraints indicate weights corresponding to the combination of the case element variables and the case level variables.
[0014] A weight sum value obtaining unit is configured to analyze and process the plurality of different hard constraints and the plurality of different soft constraints based on the plurality of different case element variables and the plurality of different case level variables to obtain weight sum values corresponding to the plurality of different case level variables when the plurality of different case level variables are assigned to true values, wherein the plurality of different hard constraints are all analyzed as true values.
[0015] The analysis conclusion obtaining unit is configured to obtain a target case analysis conclusion corresponding to the case to be analyzed based on a case level variable corresponding to a weight sum value with a highest value.
[0016] The third aspect of the present application provides a computer device. The computer device comprises:
[0017] A memory having a computer program stored thereon;
[0018] A processor configured to execute the computer program in the memory to implement the steps of the case analysis method provided in the first aspect.
[0019] The fourth aspect of the present application provides a computer readable storage medium having a computer program stored thereon. The program is executed by a processor to implement the steps of the case analysis method provided in the first aspect.
[0020] Compared with the prior art, the present application has the following beneficial effects:
[0021] In the present application, first, the case to be analyzed is Booleanized to obtain a plurality of different case element variables and a plurality of different case level variables, and then based on the plurality of different case element variables and the plurality of different case level variables, a plurality of different hard constraints and a plurality of different soft constraints are constructed. Subsequently, based on the plurality of different case element variables and the plurality of different case level variables, the plurality of different hard constraints and the plurality of different soft constraints are analyzed and processed to obtain the weights and values corresponding to the plurality of different case level variables respectively assigned as true values. Finally, based on the case level variable corresponding to the weight and value with the highest value, the target case analysis conclusion corresponding to the case to be analyzed is obtained. It should be noted that the case element variable represents the case information existing in the case to be analyzed, the case level variable represents the case analysis candidate conclusion corresponding to the case to be analyzed, the hard constraint indicates that the case to be analyzed is related to the case level variable, the soft constraint indicates the weight corresponding to the combination of the case element variable and the case level variable, and the plurality of different hard constraints are analyzed as true values.
[0022] It can be seen that, in the present application, the case to be analyzed is Booleanized into case element variables and case level variables, and based on the case element variables and the case level variables, hard constraints and soft constraints are constructed. Unlike the simple superposition and subtraction proportion in the related technical solutions, the present application can systematically represent the complex interaction between the case element variables and the case level variables in the case to be analyzed, avoiding possible analysis bias. Thereafter, the weights and values corresponding to the plurality of different case level variables respectively assigned as true values can be analyzed and obtained, and based on the case level variable with the highest weight and value, the target case analysis conclusion of the case to be analyzed is determined. In this way, through weight calculation, a complete formal reasoning process is provided, improving the explainability of case analysis, and also improving the transparency and credibility of case analysis. In this way, the present application can better realize the analysis of the case in the judicial system, and the case analysis conclusion obtained by analysis can assist the judicial personnel to speed up the analysis of the case, improving the efficiency and accuracy of case analysis. BRIEF DESCRIPTION OF DRAWINGS
[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments or the prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.
[0024] Figure 1 A flowchart of a case analysis method provided by an embodiment of the present application;
[0025] Figure 2A flow chart for constructing hard constraints and soft constraints in a case analysis method provided for an embodiment of the present application;
[0026] Figure 3 A flow chart for obtaining weights and values in a case analysis method provided for an embodiment of the present application;
[0027] Figure 4 A full flow chart for case analysis in a case analysis method provided for an embodiment of the present application;
[0028] Figure 5 A structural schematic diagram of a case analysis device provided for an embodiment of the present application. DETAILED DESCRIPTION
[0029] As described above, in judicial practice, the sentencing of a case is mainly implemented by means of artificial experience judgment (mainly relying on individual experience and subjective discretion for sentencing, lacking systematization and standardization), sentencing guide table (providing a basic reference according to the key situations of the case), basic point sentencing (setting a basic sentence, and then adding or reducing the sentence according to the light or heavy factors), case reference system (collecting historical case data, and providing a reference through simple case matching), primary expert system (implementing sentencing by means of an expert system based on a simple IF-THEN rule chain), and statistical analysis model (implementing sentencing prediction based on statistical rules of historical cases.
[0030] Specifically, in the related technical solutions, a basic sentence is first formulated based on the case situation, then the proportion of sentencing increase or decrease is determined according to the formulated rules, and the determined proportions are simply superimposed to adjust the basic sentence, so as to determine the sentencing result of the case. For example, the basic sentence of a theft case of A is 1 year and 6 months, and after adjusting the basic sentence, it can be 10 months. However, the basic point sentencing method in the related technical solutions is easy to ignore the complex interaction between the case situations, and it is difficult to provide a clear reasoning process. Therefore, how to better analyze the case so as to consider the complex interaction between the case situations and improve the explainability of the case analysis is the key problem for the person skilled in the art.
[0031] In view of the above problems, a solution is provided in the embodiments of the present application, and a case analysis method and device and related products are proposed, aiming to better analyze a case, so as to consider the complex interaction between cases and improve the explainability of case analysis. In the technical solution of the present application, the case to be analyzed is first subjected to Boolean processing to obtain a plurality of different case element variables and a plurality of different case level variables, and then based on the plurality of different case element variables and the plurality of different case level variables, a plurality of different hard constraints and a plurality of different soft constraints are constructed. Thereafter, based on the plurality of different case element variables and the plurality of different case level variables, the plurality of different hard constraints and the plurality of different soft constraints are analyzed and processed to obtain the weight and value corresponding to the case level variable when the plurality of different case level variables are respectively assigned a true value. Finally, based on the case level variable corresponding to the weight and value with the highest value, a target case analysis conclusion corresponding to the case to be analyzed is obtained. Among them, the case element variable represents the case information existing in the case to be analyzed, the case level variable represents the case analysis candidate conclusion corresponding to the case to be analyzed, the hard constraint indicates that the case to be analyzed is related to the case level variable, the soft constraint indicates the weight corresponding to the combination of the case element variable and the case level variable, and the plurality of different hard constraints are analyzed as true values.
[0032] It can be seen that, in the present application, the case to be analyzed is Booleanized into case element variables and case level variables, and hard constraints and soft constraints are constructed based on the case element variables and the case level variables. Unlike the simple superposition and subtraction ratio in the related technical solutions, the present application can systematically represent the complex interaction between the case element variables and the case level variables in the case to be analyzed, avoiding possible analysis bias. Thereafter, the weight and value corresponding to the case level variable when the plurality of different case level variables are respectively assigned a true value can be obtained by analysis, and based on the case level variable with the highest weight and value, the target case analysis conclusion of the case to be analyzed is determined. In this way, the complete formal reasoning process is provided through weight calculation, improving the explainability, transparency and credibility of case analysis. In this way, the present application can better analyze the case in the judicial system, and the case analysis conclusion obtained can assist the judicial personnel to accelerate the analysis of the case, improving the efficiency and accuracy of case analysis.
[0033] In order for those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.
[0034] Next, the technical terms that may appear in the present application are first explained.
[0035] MAXSAT: Maximum Boolean Satisfiability, which is an NP-hard combinatorial optimization problem.
[0036] Referring to Figure 1 , the figure is a flowchart of a case analysis method provided by an embodiment of the present application. As Figure 1 indicated, the case analysis method includes:
[0037] S101: Boolean processing is performed on the case to be analyzed to obtain a plurality of different case element variables and a plurality of different case level variables.
[0038] In this step, the case to be analyzed includes the case situation and the case analysis related to the case situation. Exemplarily, the case to be analyzed can be shown as follows (it should be noted that the case to be analyzed shown below in the present application is only a part of the example, and the technical solution of the present application can be applied to the analysis of various types of cases):
[0039] The basic case of A's theft case is: the defendant: A, male, 25 years old, junior high school education; case facts: on March 15, 2024, A sneaked into the underground garage of a certain community, pried open the doors of three cars, and stole the goods in the cars worth a total of 15000 yuan; case process: after the victim reported to the police, A voluntarily surrendered to the police station the next day; case trial: A confessed in court and was punished, and the victim's losses were fully compensated, and understanding was obtained. It is understood that Zhang is a first offender with no criminal record.
[0040] Applicable law analysis of A's theft case: Article 264 of the Criminal Law: Stealing public and private property in a large amount, shall be sentenced to not more than three years of fixed-term imprisonment, detention or control, and shall be punished or punished alone.
[0041] Related judicial interpretation of A's theft case: the amount of theft is more than 10,000 yuan; confession can be punished lightly or lightly; confession can be handled leniently.
[0042] It should be noted that the case element variable represents the case information existing in the case to be analyzed, wherein the plurality of different case element variables obtained after Boolean processing include core element variables, a first element variable set and a second element variable set, the element level of the first element variable set is lower than that of the second element variable set, the element variable in the first element variable set will cause the level of the case level variable (case analysis conclusion) obtained by the final analysis to tend to decrease, and the element variable in the second element variable set will cause the level of the case level variable (case analysis conclusion) obtained by the final analysis to tend to increase.
[0043] Specifically, the core element variable can be understood as the core case information constituting the case to be analyzed, the element variable in the first element variable set can be understood as the positive case information constituting the case to be analyzed, and the element variable in the second element variable set can be understood as the negative case information constituting the case to be analyzed. Then, in combination with the case to be analyzed shown above, the core element variable can include: x1: whether the amount of theft is large (15000 yuan > 10000 yuan); the first element variable set can include: x2: whether it is a first offense (no criminal record), x3: whether it is a confession, x4: whether it admits guilt and punishment, x5: whether it is fully compensated, x6: whether it obtains the understanding of the victim, and x7: whether it has a repentance performance; and the second element variable set can include: x8: whether the means of crime is cruel (prying the door), x9: whether it infringes the interests of multiple victims, and x10: whether it commits crime at night.
[0044] The case level variable represents the case analysis candidate conclusion corresponding to the case to be analyzed. After Boolean processing, a plurality of different case level variables include a first level variable set and a second level variable set. The level grade of the first level variable set is lower than the level grade of the second level variable set. The level grade of the level variable in the first level variable set gradually increases, and the level grade of the level variable in the second level variable set gradually increases. That is, the level grades of the level variables in the plurality of different case level variables are all different.
[0045] Then, in combination with the case to be analyzed shown above, the plurality of different case level variables can include: y1: whether to apply for imprisonment, y2: whether to apply for 6 months or less of fixed-term imprisonment, y3: whether to apply for 6 months-1 year of fixed-term imprisonment, y4: whether to apply for 1-2 years of fixed-term imprisonment, and y5: whether to apply for 2-3 years of fixed-term imprisonment. The first level variable set can include: y1, y2 and y3, and the second level variable set can include: y2, y3, y4 and y5.
[0046] Thus, in the present application, the case to be analyzed is converted into a Boolean variable, establishing a formal expression basis for case analysis decision, so that the complex case becomes a variable assignment problem, providing explainability of case analysis and transparentizing the whole process of case analysis.
[0047] S102: Based on the plurality of different case element variables and the plurality of different case level variables, a plurality of different hard constraints and a plurality of different soft constraints are constructed.
[0048] In this step, the hard constraint indicates that the case to be analyzed is related to the case level variable, that is, the case level variable obtained by the final analysis is necessarily one of a plurality of different case level variables; the soft constraint indicates the corresponding weight when the case element variable is combined with the case level variable, that is, the soft constraint can indicate the optimization target and tendency in the case analysis process. In this way, in this application, the case element variable and the case level variable are formalized as a hard constraint (must follow legal provisions) and a soft constraint (analysis tendency guidance) to systematically represent the complex interaction between the case element variable and the case level variable in the case to be analyzed, which provides a unified expression framework, so that the case becomes an explicit optimization problem that can be systematically solved in the subsequent process.
[0049] As Figure 2 shown, Figure 2 a flowchart for constructing a hard constraint and a soft constraint in a case analysis method provided by an embodiment of the present application, Figure 2 introducing the process of constructing the hard constraint and the soft constraint in the present application, Figure 2 including steps S1021-S1022, steps S1021-S1022 are embodied as follows:
[0050] S1021: based on the core element variable, the first element variable set, the first level variable set and the second level variable set, a plurality of different hard constraints are constructed.
[0051] Specifically, in this application, a first hard constraint can be constructed based on the core element variable, the first level variable set and the second level variable set, wherein the first hard constraint indicates that the case to be analyzed is related to the case level variable. Exemplarily (the following example is based on the above-mentioned case to be analyzed), the first hard constraint can be embodied as: H1: x1∨y1∨y2∨y3∨y4∨y5, wherein denotes the relationship of non, and denotes the relationship of or. It can be understood that the first hard constraint in this application combines the core element variable with all case level variables, so that the case level variable obtained by the final analysis falls within the variable interval constituted by all case level variables.
[0052] In the present application, the second hard constraint and the third hard constraint can be constructed based on the first level variable set and the second level variable set, wherein the second hard constraint and the third hard constraint indicate that only one case level variable can be determined. For example, the second hard constraint can be represented as: H2: ¬(y1^y2)^¬(y1^y3)^¬(y1^y4)^¬(y1^y5)^¬(y2^y3)^¬(y2^y4)^¬(y2^y5)^¬(y3^y4)^¬(y3^y5)^¬(y4^y5), wherein ^ represents the relationship of AND. It can be understood that the second hard constraint in the present application expresses the mutual exclusion principle of the case level variable through the negation of the conjunctive formula, so as to ensure that there is only one case level variable obtained by the final analysis. The third hard constraint can be represented as: H3: y1^y2^y3^y4^y5. It can be understood that the third hard constraint in the present application can ensure that the case must be analyzed and a definite analysis conclusion is finally determined, and in combination with the second hard constraint, the two can jointly constitute the complete constraint of "must and can only analyze one case level variable".
[0053] In the present application, the fourth hard constraint can be constructed based on the core element variable, the first element variable set and the second level variable set, wherein the fourth hard constraint indicates that the first level variable set can be used as the analysis basis of the case to be analyzed. For example, the fourth hard constraint can be represented as: H4: ¬x1^y5^x3^x4. It can be understood that the fourth hard constraint in the present application embodies the specific case analysis rule, and when it is found in the case analysis process that the case to be analyzed is not suitable for the higher level variable, the lower level variable can be used as the analysis basis, which reflects the condition limitation of the lenient analysis of the case to be analyzed.
[0054] Finally, in the present application, the first hard constraint, the second hard constraint, the third hard constraint and the fourth hard constraint are determined as a plurality of different hard constraints, and it needs to be noted that in the subsequent analysis process of the case, the plurality of different hard constraints need to be analyzed as true values, that is, the plurality of different hard constraints need to be satisfied. In this way, in the present application, the case to be analyzed satisfies the constructed hard constraint in the analysis process, that is, it is ensured that the analysis rule of the case is not violated, and it is also ensured that the case analysis is not wrong.
[0055] S1022: based on the first element variable set, the second element variable set, the first level variable set and the second level variable set, a plurality of different soft constraints are constructed.
[0056] Specifically, in the present application, first, the first initial constraint set and the second initial constraint set can be constructed based on the first element variable set and the first level variable set, and the third initial constraint set can be constructed based on the second element variable set and the second level variable set. Among them, the first initial constraint set and the second initial constraint set represent that the case to be analyzed can be analyzed leniently, and the third initial constraint set represents that the case to be analyzed can be analyzed severely, and the initial constraints in the first initial constraint set, the initial constraints in the second initial constraint set and the initial constraints in the third initial constraint set are all different.
[0057] After that, the first initial constraint set can be assigned a weight to obtain a first soft constraint set corresponding to the first initial constraint set, and the second initial constraint set can be assigned a weight to obtain a second soft constraint set corresponding to the second initial constraint set, that is, each effective initial constraint in the first initial constraint set and the second initial constraint set is assigned a weight. Among them, the weight corresponding to the soft constraint in the first soft constraint set and the weight corresponding to the soft constraint in the second soft constraint set are both positive values, and the weight corresponding to at least one soft constraint in the second soft constraint set is greater than the weight corresponding to the soft constraint in the first soft constraint set.
[0058] Exemplary (the following examples are based on the above-mentioned case to be analyzed), the soft constraints in the first soft constraint set can include S1-S7, wherein S1: ¬x3∨y1∨y2 [weight: 0.8], the S1 soft constraint in the present application associates the act of surrendering with detention or imprisonment for less than 6 months, and the weight is high, reflecting the tendency of analyzing the case to be analyzed leniently; S2: ¬x4∨y1∨y2∨y3 [weight: 0.6], the S2 soft constraint in the present application reflects the policy of analyzing the case to be analyzed leniently for confession and penalty, and the weight is lower than that of the S1 soft constraint, indicating that the importance is slightly lower than that of the S1 soft constraint; S3: ¬x5∨y1∨y2 [weight: 0.7], S4: ¬x6∨y1∨y2 [weight: 0.5], the S3 soft constraint and the S4 soft constraint in the present application respectively associate the full refund behavior and the behavior of obtaining the victim's understanding with light criminal analysis, and the weights are 0.7 and 0.5 respectively; S5: ¬x3∨¬x4∨¬x5∨y1 [weight: 0.9], the S5 soft constraint in the present application is a compound constraint, and the weight is high, reflecting the amplification effect of multiple mitigating circumstances superimposed; S6: ¬x2∨y1∨y2∨y3 [weight: 0.4], S7: ¬x7∨y1∨y2 [weight: 0.3], the weights corresponding to S6 soft constraint and S7 soft constraint are lower, indicating that their influence on case analysis is smaller.
[0059] Exemplarily, the soft constraints in the second soft constraint set can include S8-S9, wherein S8: ¬x4∨y1 [weight: 0.8], S9: ¬x3∨¬x5∨¬x6∨y1 [weight: 1.0], the S8 soft constraint and the S9 soft constraint in the present application embody the current judicial policy guidance, and indicate that the judicial policy plays a strong guiding role in the case analysis.
[0060] In the present application, the third initial constraint set can be assigned a weight, and a third soft constraint set corresponding to the third initial constraint set is obtained, that is, each effective initial constraint in the third initial constraint set is assigned a weight, and the weight corresponding to each soft constraint in the third soft constraint set is a negative value. Exemplarily, the soft constraints in the third soft constraint set can include S10-S12, wherein S10: ¬x8∨y3∨y4∨y5 [weight: -0.4], S11: ¬x9∨y3∨y4∨y5 [weight: -0.3], the S10 soft constraint and the S11 soft constraint in the present application indicate that when there are bad case circumstances, the weight is low, and the case level variable tends to be of a higher level; S10: ¬x10¬y2∨y3∨y4 [weight: -0.2], the S12 soft constraint in the present application has a higher weight than the S10 soft constraint and the S11 soft constraint, and indicates that the effect of the case analysis is weak. In this way, the soft constraints in the present application reflect the tendency and policy guidance of the case through different weights.
[0061] It should be noted that in the present application, the assignment of the weight can be adjusted in actual situations (such as adjustment of the current policy guidance), for example, the original weight can be adjusted to be lower or higher. In this way, the present application can ensure the robustness and adaptability of the case analysis, so that various complex case situations can be coped with.
[0062] Finally, the soft constraints in the first soft constraint set, the soft constraints in the second soft constraint set, and the soft constraints in the third soft constraint set can be determined as a plurality of different soft constraints, and it should be noted that in the subsequent case analysis process, the soft constraints can not all be true values, but the soft constraint with the highest weight and value needs to be found. In this way, in the present application, the hard constraints and the soft constraints are constructed to quantify the case to be analyzed into a calculable expression, the complex interaction relationship between the case circumstances is considered, and the relative importance of different case information in the case analysis process is simply and intuitively reflected through the weight assignment, thereby improving the explainability and transparency of the case analysis.
[0063] S103: based on the plurality of different case element variables and the plurality of different case level variables, analyzing and processing the plurality of different hard constraints and the plurality of different soft constraints, to obtain the weight and value corresponding to when the plurality of different case level variables are respectively assigned a true value.
[0064] In this step, the MAXSAT solver can automatically analyze and process a plurality of different hard constraints and a plurality of different soft constraints based on a plurality of different case element variables and a plurality of different case level variables to obtain the weights and values corresponding to the plurality of different case level variables being assigned true values, wherein the plurality of different hard constraints are all analyzed as true values in the analysis process. In this way, the weights and values of the case level variables when being assigned true values can be analyzed and solved.
[0065] As shown in Figure 3 , a flowchart of obtaining weights and values in a case analysis method provided by an embodiment of the present application is shown in Figure 3 , and the process of automatically solving and obtaining weights and values based on the MAXSAT solver in the present application is introduced in Figure 3 , which includes steps S1031-S1035, and the steps S1031-S1035 are embodied as follows: Figure 3
[0066] S1031: Assign and process the plurality of different case element variables based on the case to be analyzed to obtain a plurality of target values.
[0067] In this step, each valid case element variable can be assigned a corresponding target value, that is, one target value corresponds to one case element variable, wherein the target values of the case element variables can be the same or different, and the target values include true or false, true is true and false is false.
[0068] For example (the following example is based on the above-mentioned case to be analyzed, and the example is only a part of the example), the present application can assign a plurality of different case element variables (x1-x10) to true values according to the actual situation of the case to be analyzed. It can be understood that if the case element variable corresponds to the case information in the case to be analyzed, the case element variable is assigned to true value, and if the case element variable does not correspond to the case information in the case to be analyzed, the case element variable is assigned to false value.
[0069] After that, the present application can perform the operation of "analyzing and processing a plurality of different hard constraints and a plurality of different soft constraints based on a plurality of different case element variables and a plurality of different case level variables being assigned target values to obtain the weights and values corresponding to the plurality of different case level variables being assigned true values", which includes the following steps S1032-S1035. In this way, the case element variables are assigned according to the actual situation of the case in the present application, so that the case analysis can be based on the actual situation of the case, and the explainability of the case analysis can be met to a certain extent.
[0070] S1032: For each case level variable, the case level variable is assigned a true value.
[0071] In this step, each case level variable is a valid case level variable. It can be understood that in this application, the case level variables are assigned true values one by one, so that the weight and value corresponding to each case level variable assigned a true value are obtained in the subsequent process.
[0072] After that, the application can perform the operation of "analyzing and processing a plurality of different hard constraints and a plurality of different soft constraints based on a plurality of different case element variables assigned a target value and the case level variable assigned a true value, to obtain the weight and value corresponding to the case level variable assigned a true value", which includes the following steps S1033-S1035. In this way, by assigning case level variables one by one in this application, the case level variable that satisfies the highest weight and value is solved in the subsequent process, providing a clear case analysis process.
[0073] S1033: Based on a plurality of different case element variables assigned a target value and the case level variable assigned a true value, the plurality of different hard constraints are calculated and processed to obtain a plurality of different hard constraints calculated as true.
[0074] In this step, the plurality of different case element variables assigned a target value and the case level variable assigned a true value can be brought into a plurality of different hard constraints for solving, so that the plurality of different hard constraints are all solved as true.
[0075] Exemplary (the following example is based on the above example, and this example is only a part of the example), the case element variables x1-case element variables x10 are all assigned true, the case level variable y1 is assigned true, and the case element variables x1-case element variables x10 assigned true and the case level variable y1 assigned true are brought into the hard constraints H1-H4 (the first hard constraint, the second hard constraint, the third hard constraint and the fourth hard constraint), so that the hard constraints H1-H4 are all solved as true.
[0076] S1034: When the plurality of different hard constraints calculated as true are obtained, based on a plurality of different case element variables assigned a target value and the case level variable assigned a true value, the plurality of different soft constraints are calculated and processed to obtain the soft constraint in the plurality of different soft constraints calculated as true.
[0077] In this step, when the plurality of different hard constraints are all solved as true values, i.e., the plurality of different hard constraints are all satisfied, the plurality of different case element variables assigned with the target values and the case level variables assigned with the true values are brought into the plurality of different soft constraints (the first soft constraint set, the second soft constraint set and the third soft constraint set) to be solved to determine the soft constraints in the plurality of different soft constraints solved as true values.
[0078] For example, the case element variables x1-case element variable x10 assigned with the true values and the case level variable y1 assigned with the true value are brought into the soft constraints S1-S12 (the first soft constraint set, the second soft constraint set and the third soft constraint set) to obtain the soft constraints S1-S9 solved as true values, wherein the soft constraints S10-S11 are solved as false values.
[0079] S1035: based on the soft constraints in the plurality of different soft constraints calculated as true values, obtaining the weights corresponding to the soft constraints in the plurality of different soft constraints calculated as true values; and based on the weights corresponding to the soft constraints in the plurality of different soft constraints calculated as true values, obtaining the weight and value corresponding to the case level variable assigned with the true value.
[0080] In this step, based on the soft constraints calculated as true values, the weights corresponding to the soft constraints calculated as true values can be determined, which are the weights assigned in the above step S1022. Then, the weights corresponding to the soft constraints calculated as true values are summed up to obtain the weight and value corresponding to each case level variable assigned with the true value. It can be understood that when each case level variable is assigned with the true value, the remaining case level variables are not assigned, i.e., do not participate in the subsequent solving operation.
[0081] For example, based on the soft constraint S1 solved as true value, the weight [0.8] corresponding to the soft constraint S1 is determined; based on the soft constraint S2 solved as true value, the weight [0.6] corresponding to the soft constraint S2 is determined; based on the soft constraint S3 solved as true value, the weight [0.7] corresponding to the soft constraint S3 is determined; based on the soft constraint S4 solved as true value, the weight [0.5] corresponding to the soft constraint S4 is determined; based on the soft constraint S5 solved as true value, the weight [0.9] corresponding to the soft constraint S5 is determined; based on the soft constraint S6 solved as true value, the weight [0.4] corresponding to the soft constraint S6 is determined; based on the soft constraint S7 solved as true value, the weight [0.3] corresponding to the soft constraint S7 is determined; based on the soft constraint S8 solved as true value, the weight [0.8] corresponding to the soft constraint S8 is determined; based on the soft constraint S9 solved as true value, the weight [1.0] corresponding to the soft constraint S9 is determined.
[0082] At this time, the weight and value corresponding to the case level variable y1 being assigned a true value can be obtained as 6.0. Further, through the above process, it can be obtained that the weight and value corresponding to the case level variable y2 being assigned a true value is 5.1, the weight and value corresponding to the case level variable y3 being assigned a true value is 4.3, the weight and value corresponding to the case level variable y4 being assigned a true value is -0.9, and the weight and value corresponding to the case level variable y5 being assigned a true value is -0.7.
[0083] Thus, in the present application, a clear two-level constraint system of "hard constraint-soft constraint" is constructed, and a MAXSAT solver is used for solving. In the solving process, the case element variables and the case level variables are assigned values to solve the hard constraints and the soft constraints. This provides a complete formal reasoning process and clear quantitative basis for case analysis, so that each step of case analysis has clear logical derivation and weight calculation support. This strong interpretability improves the transparency and credibility of case analysis.
[0084] S104: Based on the case level variable corresponding to the weight and value with the highest sum, a target case analysis conclusion corresponding to the case to be analyzed is obtained.
[0085] In this step, the weight and values corresponding to multiple different case level variables being assigned a true value can be screened, the weight and value with the highest sum is obtained from the multiple different weight and values, and the case level variable corresponding to the weight and value with the highest sum is determined. Thereafter, the case level variable being assigned a true value can be determined as the target case analysis conclusion corresponding to the case to be analyzed.
[0086] Exemplarily (the following example is obtained based on the above example, and the example is only a part of the example), the case level variable y1 can be determined as the case level variable corresponding to the weight and value with the highest sum, and the case level variable y1 is whether to apply for a detention sentence. When the case level variable y1 is assigned a true value, it can be obtained that the detention sentence is applied. Then, the target case analysis conclusion corresponding to the case to be analyzed can be that the detention sentence is applied.
[0087] After that, under the condition that the target case analysis conclusion corresponding to the case to be analyzed is obtained, the judicial personnel can determine the final case analysis result corresponding to the case to be analyzed according to the actual case situation, such as 3 months of detention or 4 months of detention, etc. Thus, in the present application, the MAXSAT technology is used in the case analysis scene, which not only considers the complex interaction between cases, but also improves the explainability of case analysis, to a certain extent, assists the judicial personnel to speed up the analysis of the case, and improves the efficiency and accuracy of case analysis.
[0088] As shown in FIG. 8, the case analysis system 800 can include a case analysis module 810, a case analysis result module 820, and a case analysis result output module 830. Figure 4 The case analysis module 810 can be configured to analyze the case to be analyzed based on the case element variables and the case level variables.Figure 4 A full-process diagram of case analysis in a case analysis method provided by an embodiment of the present application is shown in Figure 4 In the method, the case to be analyzed can be first Booleanized to obtain a plurality of different case element variables and a plurality of different case level variables. Then, based on the plurality of different case element variables and the plurality of different case level variables, a plurality of different hard constraints and a plurality of different soft constraints can be constructed, wherein the plurality of different soft constraints include a first soft constraint set, a second soft constraint set and a third soft constraint set. Finally, the hard constraints and the soft constraints can be analyzed and processed under the condition that the case element variables and the case level variables are assigned values, to obtain a plurality of different case level variables respectively assigned to true values, and a corresponding target case analysis conclusion of the case to be analyzed based on the case level variable corresponding to the highest weight and value.
[0089] In summary, in the embodiment, the case to be analyzed is Booleanized into case element variables and case level variables, and hard constraints and soft constraints are constructed based on the case element variables and the case level variables. Different from the simple superposition and subtraction proportion in the related technical solution, the present application can systematically represent the complex interaction between the case element variables and the case level variables in the case to be analyzed through Booleanization processing and logical formula expression, so that the subsequent MAXSAT can globally optimize to find the optimal solution, rather than mechanically calculating, thereby avoiding the analysis deviation caused by ignoring the complex interaction between the cases in the related technical solution. Then, the corresponding weight and value of the plurality of different case level variables respectively assigned to true values can be analyzed and obtained, and the target case analysis conclusion of the case to be analyzed can be determined based on the case level variable with the highest weight and value. In this way, the present application provides a complete formal reasoning process through weight calculation, improves the explainability of case analysis, and also improves the transparency and credibility of case analysis. Thus, the present application can better realize the analysis of the case in the judicial system, and the case analysis conclusion obtained by the analysis can assist the judicial personnel to accelerate the analysis of the case, thereby improving the efficiency and accuracy of case analysis.
[0090] Based on the case analysis method provided in the foregoing embodiments, the present application also correspondingly provides a case analysis device. Figure 5 A structural schematic diagram of a case analysis device provided by an embodiment of the present application is shown in Figure 5 As shown in the figure, the case analysis device includes:
[0091] A Booleanization processing unit 501 is configured to perform Booleanization processing on a case to be analyzed to obtain a plurality of different case element variables and a plurality of different case level variables. The case element variables represent case information existing in the case to be analyzed, and the case level variables represent a case analysis candidate conclusion corresponding to the case to be analyzed.
[0092] The constraint construction obtaining unit 502 is configured to construct a plurality of different hard constraints and a plurality of different soft constraints based on the plurality of different case element variables and the plurality of different case level variables, wherein the hard constraint indicates that the case to be analyzed is related to the case level variable, and the soft constraint indicates a corresponding weight when the case element variable is combined with the case level variable;
[0093] The weight and value obtaining unit 503 is configured to analyze and process the plurality of different hard constraints and the plurality of different soft constraints based on the plurality of different case element variables and the plurality of different case level variables, and obtain a corresponding weight and value when the plurality of different case level variables are respectively assigned a true value, wherein the plurality of different hard constraints are all analyzed as true values.
[0094] The analysis conclusion obtaining unit 504 is configured to obtain a target case analysis conclusion corresponding to the case to be analyzed based on a case level variable corresponding to a weight and value with a highest value.
[0095] In an implementable embodiment, the weight and value obtaining unit 503 includes:
[0096] The element variable assignment unit is configured to assign the plurality of different case element variables based on the case to be analyzed, and obtain a plurality of target values, wherein one target value corresponds to one case element variable, and the target value includes a true value or a false value.
[0097] The constraint analysis processing unit is configured to analyze and process the plurality of different hard constraints and the plurality of different soft constraints based on the plurality of different case element variables assigned the target values and the plurality of different case level variables, and obtain a corresponding weight and value when the plurality of different case level variables are respectively assigned a true value.
[0098] In an implementable embodiment, the constraint analysis processing unit includes:
[0099] The level variable assignment unit is configured to assign each case level variable to a true value.
[0100] The hard-soft constraint analysis unit is configured to analyze and process the plurality of different hard constraints and the plurality of different soft constraints based on the plurality of different case element variables assigned the target values and the case level variable assigned the true value, and obtain a corresponding weight and value when the case level variable is assigned the true value.
[0101] In an implementable embodiment, the hard-soft constraint analysis unit is specifically configured to:
[0102] Based on the multiple different case element variables assigned with the target values and the case level variable assigned with the true value, the multiple different hard constraints are calculated to obtain multiple different hard constraints calculated as true values;
[0103] When the multiple different hard constraints calculated as true values are obtained, based on the multiple different case element variables assigned with the target values and the case level variable assigned with the true value, the multiple different soft constraints are calculated to obtain the soft constraints calculated as true values in the multiple different soft constraints;
[0104] Based on the soft constraints calculated as true values in the multiple different soft constraints, the weights corresponding to the soft constraints calculated as true values in the multiple different soft constraints are obtained; and based on the weights corresponding to the soft constraints calculated as true values in the multiple different soft constraints, the weight and value corresponding to the case level variable assigned with the true value are obtained.
[0105] In an implementable embodiment, the multiple different case element variables include a core element variable, a first element variable set and a second element variable set, wherein the element level of the first element variable set is lower than the element level of the second element variable set; and the multiple different case level variables include a first level variable set and a second level variable set, wherein the level level of the first level variable set is lower than the level level of the second level variable set.
[0106] The constraint construction obtaining unit 502 includes:
[0107] The hard constraint construction unit is configured to construct and obtain multiple different hard constraints based on the core element variable, the first element variable set, the first level variable set and the second level variable set.
[0108] The soft constraint construction unit is configured to construct and obtain multiple different soft constraints based on the first element variable set, the second element variable set, the first level variable set and the second level variable set.
[0109] In an implementable embodiment, the soft constraint construction unit is specifically configured to:
[0110] construct and obtain a first initial constraint set and a second initial constraint set based on the first element variable set and the first level variable set; and construct and obtain a third initial constraint set based on the second element variable set and the second level variable set.
[0111] weighting the first initial constraint set, the second initial constraint set and the third initial constraint set to obtain a first soft constraint set corresponding to the first initial constraint set, a second soft constraint set corresponding to the second initial constraint set and a third soft constraint set corresponding to the third initial constraint set, wherein the weights corresponding to the soft constraints in the first soft constraint set and the second soft constraint set are positive values, the weight corresponding to at least one soft constraint in the second soft constraint set is greater than the weights corresponding to the soft constraints in the first soft constraint set, and the weights corresponding to the soft constraints in the third soft constraint set are negative values;
[0112] determining the soft constraints in the first soft constraint set, the soft constraints in the second soft constraint set and the soft constraints in the third soft constraint set as a plurality of different soft constraints.
[0113] In an implementable embodiment, the hard constraint construction unit is specifically configured to:
[0114] constructing a first hard constraint based on the core element variable, the first level variable set and the second level variable set, wherein the first hard constraint indicates that the case to be analyzed is related to the case level variable;
[0115] constructing a second hard constraint and a third hard constraint based on the first level variable set and the second level variable set, wherein the second hard constraint and the third hard constraint indicate that only one case level variable is determined;
[0116] constructing a fourth hard constraint based on the core element variable, the first element variable set and the second level variable set, wherein the fourth hard constraint indicates that the first level variable set is allowed to be used as an analysis basis for the case to be analyzed;
[0117] determining the first hard constraint, the second hard constraint, the third hard constraint and the fourth hard constraint as a plurality of different hard constraints.
[0118] The case analysis device provided by the embodiments has the same beneficial effects as the case analysis method provided by the above embodiments, and can better analyze the case, so as to consider the complex interaction between cases and improve the explainability of case analysis, which will not be described herein.
[0119] The present application also provides a computer device. The computer device comprises:
[0120] a memory having a computer program stored thereon.
[0121] a processor configured to execute the computer program in the memory to implement some or all steps of the case analysis method introduced in the foregoing embodiments.
[0122] The application further provides a computer readable storage medium, which stores a computer program. The program is executed by a processor to implement some or all steps of the case analysis method introduced in the foregoing embodiments.
[0123] It should be noted that the "first" and "second" in the names of the "first" and "second" mentioned in the embodiments of the application are only used for name identification, and do not represent the first and second in order.
[0124] It should be further noted that each of the embodiments in the specification is described in a progressive manner, and the same and similar parts between the embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments. In particular, since the device and equipment embodiments are basically similar to the method embodiments, they are described more simply, and the relevant parts can be referred to the part of the description of the method embodiments. The device and equipment embodiments described above are only illustrative, and the units described as separate components can be or can not be physically separated, and the components indicated as units can be or can not be physical units, that is, they can be located in one place or distributed on multiple network units. According to the actual needs, some or all of the modules can be selected to achieve the purpose of the embodiments. Those skilled in the art can understand and implement it without creative labor.
[0125] The above is only one specific embodiment of the application, but the protection scope of the application is not limited to this. Any skilled person in the art can easily think of changes or replacements within the technical range disclosed in the application, which should be covered in the protection scope of the application. Therefore, the protection scope of the application should be subject to the protection scope of the claims.
Claims
1. A case analysis method, characterized in that, include: Booleanization is performed on the case to be analyzed to obtain multiple different case element variables and multiple different case level variables. The case element variables represent the case information present in the case to be analyzed, and the case level variables represent the candidate conclusions of the case analysis corresponding to the case to be analyzed. Based on the multiple different case element variables and the multiple different case level variables, multiple different hard constraints and multiple different soft constraints are constructed, wherein the hard constraints indicate that the case to be analyzed is related to the case level variable, and the soft constraints indicate the weights corresponding to the combination of the case element variables and the case level variables; Based on the multiple different case element variables and the multiple different case level variables, the multiple different hard constraints and the multiple different soft constraints are analyzed and processed to obtain the weights and values corresponding to the multiple different case level variables when they are assigned true values, wherein the multiple different hard constraints are all analyzed as true values; Based on the case level variable corresponding to the highest weighted sum, the analysis conclusion of the target case corresponding to the case to be analyzed is obtained.
2. The method according to claim 1, characterized in that, The process involves analyzing and processing the various hard constraints and soft constraints based on the multiple different case element variables and the multiple different case severity level variables to obtain the weights and values corresponding to each of the multiple different case severity level variables when they are assigned true values, including: Based on the case to be analyzed, the multiple different case element variables are assigned values to obtain multiple target values, one of which corresponds to one case element variable. The target value includes a true value or a false value. Based on multiple different case element variables and multiple different case level variables that are assigned target values, the multiple different hard constraints and multiple different soft constraints are analyzed and processed to obtain the weights and values corresponding to the multiple different case level variables when they are assigned true values.
3. The method according to claim 2, characterized in that, The process involves analyzing and processing multiple different case element variables and multiple different case severity level variables, based on the assigned target values, to obtain the weights and values corresponding to the multiple different case severity level variables when each is assigned a true value, including: For each case level variable, assign the case level variable a true value; Based on multiple different case element variables assigned target values and the case level variable assigned true values, the multiple different hard constraints and multiple different soft constraints are analyzed and processed to obtain the weight and value corresponding to the case level variable when it is assigned a true value.
4. The method according to claim 3, characterized in that, The process involves analyzing and processing multiple different case element variables assigned target values and a case severity level variable assigned true values, along with multiple different hard constraints and multiple different soft constraints, to obtain the weights and values corresponding to the case severity level variable when it is assigned a true value. This includes: Based on multiple different case element variables assigned target values and the case grade variable assigned true values, the multiple different hard constraints are calculated and processed to obtain multiple different hard constraints that are calculated to be true values. When obtaining multiple different hard constraints that are calculated to be true, the multiple different soft constraints are calculated based on multiple different case element variables that are assigned target values and the case grade variable that is assigned to be true values, and the soft constraints that are calculated to be true among the multiple different soft constraints are obtained. Based on the soft constraints that are calculated to be true among the multiple different soft constraints, obtain the weights corresponding to the soft constraints that are calculated to be true among the multiple different soft constraints; and based on the weights corresponding to the soft constraints that are calculated to be true among the multiple different soft constraints, obtain the weights and values corresponding to the case level variable when it is assigned a true value.
5. The method according to claim 1, characterized in that, The multiple different case element variables include core element variables, a first element variable set, and a second element variable set, wherein the element level of the first element variable set is lower than the element level of the second element variable set; the multiple different case grade variables include a first grade variable set and a second grade variable set, wherein the grade level of the first grade variable set is lower than the grade level of the second grade variable set. Based on the multiple different case element variables and the multiple different case level variables, multiple different hard constraints and multiple different soft constraints are constructed, including: Based on the core element variables, the first element variable set, the first level variable set, and the second level variable set, multiple different hard constraints are constructed and obtained; Based on the first set of element variables, the second set of element variables, the first set of level variables, and the second set of level variables, multiple different soft constraints are constructed.
6. The method according to claim 5, characterized in that, The process involves constructing multiple different soft constraints based on the first set of element variables, the second set of element variables, the first set of level variables, and the second set of level variables, including: Based on the first set of element variables and the first set of grade variables, a first initial constraint set and a second initial constraint set are constructed; and based on the second set of element variables and the second set of grade variables, a third initial constraint set is constructed. Weights are assigned to the first initial constraint set, the second initial constraint set, and the third initial constraint set to obtain a first soft constraint set corresponding to the first initial constraint set, a second soft constraint set corresponding to the second initial constraint set, and a third soft constraint set corresponding to the third initial constraint set. The weights of the soft constraints in the first soft constraint set and the second soft constraint set are all positive. The weight of at least one soft constraint in the second soft constraint set is greater than the weight of the soft constraint in the first soft constraint set. The weights of the soft constraints in the third soft constraint set are all negative. The first set of soft constraints, the second set of soft constraints, and the third set of soft constraints are identified as multiple different soft constraints.
7. The method according to claim 5, characterized in that, Based on the core element variables, the first element variable set, the first level variable set, and the second level variable set, multiple different hard constraints are constructed, including: Based on the core element variables, the first set of variables and the second set of variables, a first hard constraint is constructed, wherein the first hard constraint indicates that the case to be analyzed is related to the case level variables; Based on the first set of level variables and the second set of level variables, a second hard constraint and a third hard constraint are constructed, wherein the second hard constraint and the third hard constraint indicate that only one case level variable is determined; A fourth hard constraint is constructed based on the core element variables, the first set of element variables, and the second set of level variables, wherein the fourth hard constraint indicates that the first set of level variables can be used as the basis for analysis of the case to be analyzed. The first hard constraint, the second hard constraint, the third hard constraint, and the fourth hard constraint are determined as multiple different hard constraints.
8. A case analysis device, characterized in that, include: The Booleanization unit is used to perform Booleanization on the case to be analyzed, thereby obtaining multiple different case element variables and multiple different case level variables. The case element variables represent the case information present in the case to be analyzed, and the case level variables represent the candidate conclusions of the case analysis corresponding to the case to be analyzed. The constraint construction and acquisition unit is used to construct and acquire multiple different hard constraints and multiple different soft constraints based on the multiple different case element variables and the multiple different case grade variables. The hard constraints indicate that the case to be analyzed is related to the case grade variable, and the soft constraints indicate the weights corresponding to the combination of the case element variables and the case grade variables. The weight and value acquisition unit is used to analyze and process the multiple different hard constraints and multiple different soft constraints based on the multiple different case element variables and the multiple different case level variables, and to obtain the weight and value corresponding to the multiple different case level variables when they are respectively assigned true values, wherein the multiple different hard constraints are all analyzed as true values; The analysis conclusion acquisition unit is used to obtain the target case analysis conclusion corresponding to the case to be analyzed based on the case level variable corresponding to the case with the highest weighted sum value.
9. A computer device, characterized in that, include: A memory on which computer programs are stored; A processor for executing the computer program in the memory to implement the steps of the case analysis method according to any one of claims 1-7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the case analysis method according to any one of claims 1-7.