Case auditing method based on multi-dimension reduced penalty
By constructing a multi-dimensional attribute association mapping table and calculating the matching overlap, a traceable scoring chain for sentence reduction and parole is generated, which solves the problem of high ambiguity rate in rule matching and achieves transparent and accurate scoring of sentence reduction and parole cases.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-26
- Publication Date
- 2026-03-27
AI Technical Summary
In handling cases of sentence reduction and parole, existing technologies suffer from high ambiguity in rule matching, unstable scoring, and difficulty in generating traceable recommendations for sentence reduction.
By acquiring data from the criminal record system and the case handling system, a multi-dimensional attribute association mapping table is constructed, the matching overlap between attributes is calculated, the target rule item set is selected, and a traceable commutation and parole scoring chain is generated.
Reduce the ambiguity rate of rule matching, enhance the interpretability of the scoring composition, enable traceable output of the reduction of sentence recommendations, and improve the transparency and accuracy of the review process.
Smart Images
Figure CN121743897A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing and information retrieval technology, and more specifically, to a method for reviewing commutation and parole cases based on multiple dimensions. Background Technology
[0002] In the context of handling commutation and parole cases, rule-based evaluation and suggestions are usually generated based on structured data from the offender file system and the case handling system. Under the constraints of limited processing time and multi-user concurrency, the review reference results are output. Existing technologies mostly adopt ideas such as rule condition filtering, multi-dimensional attribute matching, sub-item scoring summary, threshold determination and result tracking to achieve automated processing. Its applicable boundaries usually depend on the premise that the data is consistent, the rule items are mutually exclusive, and the indicators change slowly during the assessment period.
[0003] In actual operation, rule base item iteration and scenario coverage expansion introduce rule item overlap and boundary overlap. Cross-system data synchronization and manual supplementation introduce differences in caliber and missing fields. These factors can cause multiple rules to be hit simultaneously and trigger different scoring paths when multi-dimensional attribute matching is performed. As a result, it is difficult to form a stable sub-attribution chain for the scoring composition. Ultimately, it is difficult to generate traceable output results for the reduction of sentence. Therefore, the technical problem that needs to be solved is how to achieve traceable output of the reduction of sentence recommendation while reducing the ambiguity rate of rule matching and enhancing the interpretability of the scoring composition.
[0004] In view of this, the present invention proposes a multi-dimensional method for reviewing commutation and parole cases to solve the above problems. Summary of the Invention
[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method for reviewing commutation and parole cases based on multiple dimensions.
[0006] To achieve the above objectives, the present invention provides the following technical solution: Firstly, it provides a multi-dimensional method for reviewing commutation and parole cases, including: Data on prisoners' performance during their sentences is obtained from the prisoner file system, and historical data on commutation and parole cases is obtained from the case handling system. Based on the prisoner performance data and historical data, a set of original prisoner data is formed. Extract a multi-dimensional attribute set of commutation and parole rules from the original dataset of criminals, and construct an attribute association mapping table based on the multi-dimensional attribute set; The matching overlap between attributes is calculated based on the attribute association mapping table, and the set of overlapping target rule entries is determined based on the matching overlap. The applicable frequency of each rule entry is calculated based on the historical matching records of the target rule entry set, and a set of rule entries to be used is selected based on the applicable frequency. The original dataset of criminals is matched and analyzed based on the set of rules to be used to generate a commutation and parole scoring chain, and a traceable commutation range recommendation is generated based on the commutation and parole scoring chain.
[0007] In some embodiments, a multidimensional attribute set of commutation and parole rules is extracted from the original dataset of offenders, and an attribute association mapping table is constructed based on the multidimensional attribute set, including: Identify the specific attribute categories involved in the rules for commutation and parole based on the historical processing data in the original criminal data set, and assign the attribute categories under the same processing situation to the corresponding category set; Extract the values of the corresponding attributes for each attribute category set, and categorize and organize the attribute values to form an attribute category value set; The cross-comparison method is used to compare the value sets of each pair of attribute categories one by one in order to identify the cross attribute values and record them as attribute cross-mapping sets; The above attribute cross-mapping set is structured according to the cross-strength of values between attribute categories to obtain an attribute association mapping table.
[0008] In some embodiments, a cross-comparison method is used to compare each pair of attribute category value sets one by one to identify cross-attribute values and record them as an attribute cross-mapping set, including: Extract specific attribute values from the two attribute category value sets respectively, and compare the numerical values or categories of the attribute values one by one to see if they are the same; Record whether there is an overlap for each attribute pair based on the attribute value comparison results, and accumulate the number of overlapping attribute values. The total number of attribute values and the number of cross attribute values within the value set of each attribute category are calculated to obtain the attribute cross value ratio; The cross attribute values are determined based on the attribute categories whose cross value ratio exceeds a preset ratio, and recorded as an attribute cross mapping set.
[0009] In some embodiments, the matching overlap degree between attributes is calculated based on the attribute association mapping table, and a set of overlapping target rule entries is determined based on the matching overlap degree, including: The attribute association mapping table is used to parse the attribute combination of rule entries in the preset commutation and parole rule base, and the attribute combination feature set corresponding to each rule entry is recorded. Cross-matching is performed on the attribute combination feature sets of any two rule entries, and the number of overlapping features of corresponding attributes between each pair of rule entries is counted one by one. The overlapping attribute ratio is calculated based on the number of overlapping attribute features in each pair of rule entries, and the pair of rule entries is sorted according to the overlapping attribute ratio to obtain a set of overlapping rule rankings. Select rule entries from the rule overlap sorting set whose overlap attribute ratio meets the preset overlap conditions, and extract their corresponding rule entries to form the target rule entry set.
[0010] In some embodiments, the overlapping attribute ratio is calculated based on the number of overlapping attribute features in each rule entry pair, and the rule entry pairs are sorted according to the overlapping attribute ratio to obtain a rule overlap sorting set, including: Calculate the number of all attribute combinations for each rule entry; Calculate the number of overlapping attribute features shared between each pair of rule entries; The specific percentage of overlapping attributes is obtained by dividing the number of overlapping attribute features by the total number of attribute combination features for each rule entry pair. Based on the specific overlapping attribute ratio values, the rule entries are sorted in descending order to obtain the rule overlap sorted set.
[0011] In some embodiments, the applicable frequency of each rule entry is calculated based on the historical matching records of the target rule entry set, and a set of rule entries to be used is selected based on the applicable frequency, including: Based on the target set of rule entries, extract the set of matching records for each rule entry from the historical processing data, and calculate the number of matching records for each rule entry; Count the number of matched records for all target rule entries, and calculate the matching percentage for each rule entry accordingly; Based on the matching ratio, the rule entries in the target rule entry set are sorted in descending order to obtain the rule application frequency set; Select rule entries with a matching rate higher than a preset rate threshold from the set of rule application frequencies, and filter out the set of rule entries to be used.
[0012] In some embodiments, the rule entries in the target rule entry set are sorted in descending order according to the matching ratio to obtain a rule application frequency set, including: Calculate the percentage of each rule entry relative to the total number of matches in the target rule entry set based on the number of matched records; The rule entries are arranged in order of their matching percentage to form a rule percentage sorting list; Mark the matching percentage value corresponding to each rule entry in the sorted list according to the rule percentage sorting list; Based on the rule proportion sorting list and the matching proportion values of its tags, a set of rule application frequencies is generated.
[0013] In some embodiments, a matching analysis is performed on the original dataset of offenders based on a set of pending rule entries to generate a commutation and parole scoring chain, and a traceable commutation range recommendation is generated based on the commutation and parole scoring chain, including: Based on the set of rule entries to be used, extract the attribute combination features and scoring methods corresponding to each rule entry to generate a set of scoring feature maps; Extract specific attribute data for each criminal case from the original criminal data set, and then compare each specific attribute data item by item with the scoring feature mapping set; Based on the precisely compared attribute data and the set of scoring feature maps, the specific scoring source and scoring value are determined, and a clear scoring chain for sentence reduction and parole is generated. Suggestions for the extent of sentence reduction are generated based on the specific score values and corresponding attribute combinations in the sentence reduction and parole scoring chain.
[0014] In some embodiments, the specific scoring source and scoring value are determined based on precisely compared attribute data and a set of scoring feature maps, generating a clear scoring chain for sentence reduction and parole, including: Based on the specific attribute data of the criminal case and the attribute combination characteristics corresponding to the rule entries to be used, the specific score corresponding to each attribute is determined item by item. Construct a set of rating records for each attribute based on the specific rating; Extract all attribute score values from the score record set, and calculate the weighted score values of the attributes according to the weight requirements of the score feature mapping set; The weighted scoring values are organized and recorded in the logical order of the rule entries to be used, forming a clear scoring chain for sentence reduction and parole.
[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention forms a raw dataset of criminals from the criminal file system and the case handling system, unifying the data on incarceration performance and historical case handling data. Based on this, a multi-dimensional attribute set of commutation and parole rules is extracted, and an attribute association mapping table is constructed. Under the same structure, the relationships between attributes are explicitly presented. Based on the attribute association mapping table, the matching overlap is calculated. First, a set of target rule entries with overlap is located to reduce the ambiguity of multiple rules hitting simultaneously. Then, the applicability frequency is calculated based on the historical matching records of the target rule entry set, and a set of pending rule entries is obtained. This ensures that the rule entries entering subsequent matching have measurable stability. Subsequently, the set of pending rule entries is used to perform matching analysis on the raw dataset of criminals, generating a commutation and parole scoring chain. The sources and paths of the sub-scorings are sequentially unfolded on the chain. Finally, a traceable commutation range suggestion is generated based on the commutation and parole scoring chain. This reduces the rule matching ambiguity rate while enhancing the interpretability of the scoring composition, achieving traceable output of the commutation range suggestion. Attached Figure Description
[0016] Figure 1 This is a flowchart illustrating a multi-dimensional method for reviewing commutation and parole cases according to the present invention. Figure 2 This is a schematic diagram of a multi-dimensional commutation and parole case review system according to the present invention. Detailed Implementation
[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0018] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains. The terms “first,” “second,” and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as “comprising” or “including” mean that the element or object preceding the term covers the element or object listed after the term and its equivalents, without excluding other elements or objects. Terms such as “connection” or “linked” are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect.
[0019] In the technical solution of this invention, the user information (including but not limited to user personal information, user image information, user device information, such as location information) and data (including but not limited to data used for analysis, stored data, and displayed data) involved are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data are all carried out in accordance with relevant laws, regulations, and standards, and necessary confidentiality measures have been taken. They do not violate public order and good morals, and corresponding operation entry points are provided for users to choose to authorize or refuse. Example
[0020] Figure 1 This disclosure illustrates at least one embodiment of a method for reviewing sentence reduction and parole cases based on multiple dimensions, including: S10: Obtain data on the prisoner's performance during imprisonment from the prisoner file system and historical data on the handling of commutation and parole cases from the case handling system, and form a set of original prisoner data based on the prisoner's performance data and historical data. In this embodiment, data on the prisoner's performance during imprisonment is obtained from the prisoner file system, and historical data on the handling of commutation and parole cases is obtained from the case handling system. This aims to ensure the accuracy and completeness of the data sources during the commutation and parole review process, thereby avoiding ambiguity in subsequent data processing due to differences in data definitions or missing fields. The prisoner's performance data specifically includes the prisoner's daily performance, award records, punishment status, and point level data during the assessment period. The historical data on the handling of commutation and parole cases specifically includes the judge's historical record of rule application, trial results, number of rule hits, and review opinions.
[0021] For example, for case number 001, the data vector of incarceration performance obtained from the incarceration file system can be represented as (performance level: excellent, number of rewards: 3, number of punishments: 0, point level: level 1), and the set of historical case handling data obtained from the case handling system can be represented as {(case number: 2022001, applicable rule number: R12, sentence reduction range: 6 months), (case number: 2022002, applicable rule number: R15, sentence reduction range: 3 months)}. Then, the above two sets of data are integrated to form the original data set of the incarceration, providing a unified multi-dimensional data input for subsequent steps.
[0022] S20: Extract a multi-dimensional attribute set of commutation and parole rules from the original dataset of criminals, and construct an attribute association mapping table based on the multi-dimensional attribute set; A multidimensional attribute set of commutation and parole rules is extracted from the original dataset of criminals, and an attribute association mapping table is constructed based on the multidimensional attribute set, including: Identify the specific attribute categories involved in the rules for commutation and parole based on the historical processing data in the original criminal data set, and assign the attribute categories under the same processing situation to the corresponding category set; Extract the values of the corresponding attributes for each attribute category set, and categorize and organize the attribute values to form an attribute category value set; The cross-comparison method is used to compare the value sets of each pair of attribute categories one by one in order to identify the cross attribute values and record them as attribute cross-mapping sets; The above attribute cross-mapping set is structured according to the cross-strength of values between attribute categories to obtain an attribute association mapping table.
[0023] Understandably, extracting a multi-dimensional attribute set of commutation and parole rules from the original dataset of offenders and constructing an attribute association mapping table aims to transform the review criteria, which were originally scattered across different rule entries and data fields, into a structured attribute system that can be processed uniformly. This provides a stable data foundation for subsequent rule overlap identification and scoring chain construction. The multi-dimensional attribute set refers to the attribute categories directly related to commutation and parole rules extracted from offender performance data and historical processing data, such as offender's type of sentence, crime type, term of imprisonment, reward records, punishment records, and historical commutation status. These attributes may appear in different combinations in different rules. Based on the actual application of the rules in historical processing data, the specific attribute categories involved in each commutation and parole rule are identified, and attribute categories that appear together in the same processing situation are merged into the same category set.
[0024] For example, in the aforementioned case of criminal number 001, historical cases show that both rule R12 and rule R15 involve the type of punishment, the term of imprisonment, and reward records. These attributes are then assigned to the same set of attribute categories. Subsequently, the corresponding attribute values are extracted from each set of attribute categories and classified and organized to form a set of attribute category values. For example, the set of punishment type values can be represented as {fixed-term imprisonment, life imprisonment}, and the set of reward record values can be represented as {0 times, 1 time, 3 times}.
[0025] Furthermore, by cross-comparing the value sets of different attribute categories one by one, the cross-attribute values that appear simultaneously in different rules are identified, and these cross-values are recorded as attribute cross-mapping sets. Finally, the attribute cross-mapping sets are structured according to the cross-value strength between attribute categories to form an attribute association mapping table. This attribute association mapping table clearly depicts the cross-relationship of different attribute categories in the application of rules. Compared with the existing method of matching only according to rule text or a single attribute, this embodiment effectively reveals the potential overlap between rule entries by constructing an attribute association mapping table, providing clear data support for reducing the ambiguity rate of rule matching and enhancing the interpretability of subsequent scoring.
[0026] A cross-comparison method is used to compare each pair of attribute category value sets one by one to identify the cross-attribute values and record them as an attribute cross-mapping set, including: Extract specific attribute values from the two attribute category value sets respectively, and compare the numerical values or categories of the attribute values one by one to see if they are the same; Record whether there is an overlap for each attribute pair based on the attribute value comparison results, and accumulate the number of overlapping attribute values. The total number of attribute values and the number of cross attribute values within the value set of each attribute category are calculated to obtain the attribute cross value ratio; The cross attribute values are determined based on the attribute categories whose cross value ratio exceeds a preset ratio, and recorded as an attribute cross mapping set.
[0027] Understandably, the purpose of using the cross-comparison method to compare each set of attribute category values one by one and identify the cross-attribute values is to clarify the specific cross-correspondence between the values of different attribute categories in the rules for sentence reduction and parole cases, thereby providing clear data support for the subsequent determination of rule overlap. Specifically, specific attribute values are extracted from the two sets of attribute category values respectively.
[0028] For example, in the case of criminal number 001, attribute category value set 1 (type of punishment) is {fixed-term imprisonment, life imprisonment}, and attribute category value set 2 (duration of imprisonment) is {1 year, 2 years, 3 years, 5 years, life imprisonment}. Then, value set 1 and set 2 are cross-compared one by one, that is, the attribute values such as "life imprisonment" are compared one by one to determine the specific cross-attribute values that are the same between the two sets. For example, in this case, "life imprisonment" and "life imprisonment" clearly overlap. Next, based on the above comparison results, the cross-attribute values of each attribute pair are recorded one by one, and the total number of cross-attribute values is recorded. For example, the above attribute pair (type of punishment: life imprisonment, duration of imprisonment: life imprisonment)... The method involves identifying and counting crossovers, then calculating the total number of attributes in each attribute category's value set against the number of crossover attribute values. This clarifies the specific proportion of crossover attributes to all attribute values. Finally, attribute categories with crossover attribute proportions exceeding a preset ratio are further identified as valid crossover attribute values and structurally recorded as attribute crossover mapping sets. This embodiment clarifies the basis for crossover matching between different attribute categories through this crossover comparison method. Compared to existing technologies that compare single rules one by one or rely solely on human experience, this embodiment effectively reduces the ambiguity rate of rule matching, providing reliable data support for the construction of the scoring chain in the commutation and parole review process and the clarity of commutation recommendations.
[0029] S30: Calculate the matching overlap between attributes based on the attribute association mapping table, and determine the set of overlapping target rule entries based on the matching overlap. The matching overlap degree between attributes is calculated based on the attribute association mapping table, and the set of overlapping target rule entries is determined based on the matching overlap degree, including: The attribute association mapping table is used to parse the attribute combination of rule entries in the preset commutation and parole rule base, and the attribute combination feature set corresponding to each rule entry is recorded. Cross-matching is performed on the attribute combination feature sets of any two rule entries, and the number of overlapping features of corresponding attributes between each pair of rule entries is counted one by one. The overlapping attribute ratio is calculated based on the number of overlapping attribute features in each pair of rule entries, and the pair of rule entries is sorted according to the overlapping attribute ratio to obtain a set of overlapping rule rankings. Select rule entries from the rule overlap sorting set whose overlap attribute ratio meets the preset overlap conditions, and extract their corresponding rule entries to form the target rule entry set.
[0030] In this embodiment, the purpose of calculating the matching overlap between attributes based on the attribute association mapping table and determining the set of overlapping target rule entries based on the matching overlap is to accurately identify the possible attribute overlap between the rules for handling commutation and parole cases, thereby effectively reducing the processing ambiguity caused by multiple rules hitting at the same time. Specifically, the attribute association mapping table previously constructed is first used to perform attribute combination analysis on each rule entry in the preset commutation and parole rule base to clarify the specific attribute combination feature set corresponding to each rule entry.
[0031] For example, in the handling of a specific case, the attribute combination feature set of rule item R12 is {Sentence type: fixed-term imprisonment, Number of rewards: 3, Number of punishments: 0}, and the attribute combination feature set of rule item R15 is {Sentence type: fixed-term imprisonment, Number of rewards: 3, Number of punishments: 1}. Then, the attribute combination feature sets of any two rule items are cross-matched. For each pair of rule items, features with the same attribute value are counted one by one. Specifically, in the cross-matching of rule items R12 and R15, the number of overlapping features is 2 (the sentence type and number of rewards are exactly the same). Then, based on the overlapping attributes of each pair of rule items... The number of features is used to calculate the proportion of overlapping attributes. Specifically, the proportion of overlapping attributes is obtained by dividing the number of overlapping attribute features by the total number of attribute features in a single rule entry. For example, in the rule entry pair R12 and R15 above, rule entry R12 has 3 attributes and 2 overlapping attributes. Therefore, the proportion of overlapping attributes in this rule entry pair is 2 / 3. Then, all rule entry pairs are sorted in descending order according to the calculated proportion of overlapping attributes to form a clear set of overlapping rules. Finally, rule entry pairs that meet the preset overlap conditions are selected from the set of overlapping rules, and the corresponding rule entries in these rule entry pairs are extracted to form the target set of rule entries.
[0032] This embodiment, through the above-mentioned matching overlap calculation and rule filtering method, significantly improves the matching accuracy of commutation and parole rules compared with the existing technology that only relies on single attribute comparison. It effectively reduces attribute matching ambiguity and ensures the clarity and interpretability of case handling results.
[0033] The overlapping attribute ratio is calculated based on the number of overlapping attribute features in each rule pair, and the rule pairs are sorted according to the overlapping attribute ratio to obtain a rule overlap sorting set, including: Calculate the number of all attribute combinations for each rule entry; Calculate the number of overlapping attribute features shared between each pair of rule entries; The specific percentage of overlapping attributes is obtained by dividing the number of overlapping attribute features by the total number of attribute combination features for each rule entry pair. Based on the specific overlapping attribute ratio values, the rule entries are sorted in descending order to obtain the rule overlap sorted set.
[0034] In this embodiment, the proportion of overlapping attributes is calculated based on the number of overlapping attribute features of each rule entry pair, and the rule entry pairs are sorted to obtain a rule overlap sorting set. The purpose of this is to accurately assess the degree of attribute overlap between rule entries, so as to clarify the overlap of rule entries and thus effectively reduce ambiguity in the matching process of commutation and parole rules. Specifically, the number of all attribute combination features contained in each rule entry pair is first calculated as a clear benchmark for subsequent calculations.
[0035] For example, in actual cases of sentence reduction and parole, rule item R12 includes attribute combination features (type of sentence: fixed-term imprisonment, number of rewards: 3, number of punishments: 0), with a total of 3 features; rule item R15 includes attribute combination features (type of sentence: fixed-term imprisonment, number of rewards: 3, number of punishments: 1), also with a total of 3 features. Then, the number of overlapping attribute features shared between each pair of rule items is calculated. Taking rule items R12 and R15 as examples, the overlap is identified by comparing the attribute combination features one by one. The attribute combination features are (crime type: fixed-term imprisonment, number of rewards: 3 times), with 2 overlapping features. Then, by dividing the number of overlapping attribute features by the total number of attribute combination features for each rule entry, the specific overlapping attribute ratio is obtained. For example, the overlapping attribute ratio for rule entry pair (R12, R15) is 2 / 3, approximately 0.67. Finally, based on the aforementioned specific overlapping attribute ratio, all rule entry pairs are sorted in descending order to form a clear set of rule overlap rankings. For example, the rule overlap ranking set can be represented as: {(R12,R15):0.67,(R15,R20):0.5,(R12,R20):0.33}.
[0036] In this embodiment, the rule overlap sorting set obtained through the above steps significantly improves the accuracy and objectivity of the rule overlap assessment in commutation and parole cases compared with the existing technology that relies solely on human experience or simple rule text comparison. It effectively reduces the ambiguity problem of multiple rules being hit simultaneously and provides a solid data foundation for the subsequent formation of reliable commutation range recommendations.
[0037] S40: Calculate the applicable frequency of each rule entry based on the historical matching records of the target rule entry set, and filter out the set of rule entries to be used based on the applicable frequency; The applicability frequency of each rule entry is calculated based on the historical matching records of the target rule entry set, and a set of rule entries to be used is selected based on the applicability frequency, including: Based on the target set of rule entries, extract the set of matching records for each rule entry from the historical processing data, and calculate the number of matching records for each rule entry; Count the number of matched records for all target rule entries, and calculate the matching percentage for each rule entry accordingly; Based on the matching ratio, the rule entries in the target rule entry set are sorted in descending order to obtain the rule application frequency set; Select rule entries with a matching rate higher than a preset rate threshold from the set of rule application frequencies, and filter out the set of rule entries to be used.
[0038] In this embodiment, the application frequency of each rule entry is calculated based on the historical matching records of the target rule entry set, and a set of rule entries to be used is selected. This aims to clarify the frequency with which different rule entries are actually used in the historical case handling process, so as to reasonably select high-frequency rules, thereby reducing ambiguity in rule matching and avoiding interference from low-frequency rules. Extracting the matching record set of each rule entry from the historical case handling data refers to extracting the case number of each rule entry that is actually applied from the historical data of commutation and parole cases in the case handling system, and counting the corresponding number of matching records. For example, for the three rule entries R12, R15, and R20 in the target rule entry set, the following historical matching information can be obtained, as shown in Table 1: Table 1 rule entries Match case number Number of matching records Matching percentage R15 2022002,2022004,2022006,2022007,2022008,2022009 6 0.6 R12 2022001,2022003,2022005 3 0.3 R20 2022010 1 0.1 Furthermore, based on the aforementioned statistical data, the matching percentage of each rule entry is calculated. Specifically, the number of matching records for a single rule entry is divided by the total number of matching records for all rule entries. In the example above, the total number of matches is 10, so the matching percentage of R15 is 0.6, the matching percentage of R12 is 0.3, and the matching percentage of R20 is 0.1. Subsequently, the target rule entry set is sorted in descending order based on the calculated matching percentage to form a rule application frequency set. Rule entries with a matching percentage higher than a preset threshold (such as 0.2) are selected from the rule application frequency set, thus filtering out the set of rule entries to be used.
[0039] The rule entries in the target rule entry set are sorted in descending order based on the matching ratio to obtain the rule application frequency set, which includes: Calculate the percentage of each rule entry relative to the total number of matches in the target rule entry set based on the number of matched records; The rule entries are arranged in order of their matching percentage to form a rule percentage sorting list; Mark the matching percentage value corresponding to each rule entry in the sorted list according to the rule percentage sorting list; Based on the rule proportion sorting list and the matching proportion values of its tags, a set of rule application frequencies is generated.
[0040] In this embodiment, the rule entries in the target rule entry set are sorted in descending order according to the matching ratio to obtain the rule application frequency set. The purpose is to clearly identify the frequency differences in the actual application of rule entries, so as to reasonably select rule entries with higher application frequency in the review of historical commutation and parole cases, further reduce the ambiguity of rule application and improve processing efficiency. Specifically, firstly, based on the number of historical matching records of each rule entry, the specific percentage of each rule entry in the total number of matching records of the target rule entry set is calculated to obtain a clear matching ratio.
[0041] For example, in the above example of the target rule entry set (R15, R12, R20), the total number of matches is 10, of which R15 has 6 matches, R12 has 3 matches, and R20 has 1 match. The corresponding match percentages are R15:60%, R12:30%, and R20:10%, respectively. Then, based on the clearly calculated match percentages, all rule entries are sorted in descending order according to the numerical value of the match percentage, forming a clear rule percentage sorting list. For example, the rule percentage sorting list formed in the above example is: {R15:60%, R12:30%, R20:10%}. And according to this sorting list, each rule entry is clearly marked with its corresponding specific match percentage value to ensure the accurate correspondence between the match percentage and the rule entry.
[0042] Finally, based on the above rule proportion ranking list and the matching proportion values of the tags, a rule application frequency set is generated. The rule application frequency set clearly expresses the actual historical frequency of rule entries. This embodiment overcomes the defects of existing subjective experience judgment or fuzzy matching by using a clear and structured rule entry application frequency calculation method. It effectively improves the accuracy and objectivity of rule selection in the review process of commutation and parole cases, significantly reduces the degree of ambiguity in the rule selection process, and provides clear and accurate data basis for the subsequent construction of the commutation and parole scoring chain.
[0043] S50: Match and analyze the original data set of criminals based on the set of rules to be used, generate a commutation and parole scoring chain, and generate a traceable commutation range suggestion based on the commutation and parole scoring chain.
[0044] The original dataset of offenders is matched and analyzed based on the set of rules to be used to generate a commutation and parole scoring chain. Based on this scoring chain, a traceable commutation range recommendation is generated, including: Based on the set of rule entries to be used, extract the attribute combination features and scoring methods corresponding to each rule entry to generate a set of scoring feature maps; Extract specific attribute data for each criminal case from the original criminal data set, and then compare each specific attribute data item by item with the scoring feature mapping set; Based on the precisely compared attribute data and the set of scoring feature maps, the specific scoring source and scoring value are determined, and a clear scoring chain for sentence reduction and parole is generated. Suggestions for the extent of sentence reduction are generated based on the specific score values and corresponding attribute combinations in the sentence reduction and parole scoring chain.
[0045] Understandably, the purpose of matching and analyzing the original data set of criminals with the set of pending rule entries to generate a commutation and parole scoring chain and generating traceable commutation range suggestions based on this scoring chain is to ensure that the source of commutation and parole scoring results is clear and the data is explicit, thereby effectively solving the technical problems of unclear scoring sources and difficulty in retrospection of results in traditional methods. Specifically, firstly, based on the selected set of pending rule entries, specific attribute combinations and corresponding scoring methods are clearly extracted from each rule entry to form a clear scoring feature mapping set.
[0046] For example, the attribute combination features of the pending rule entry R15 include the type of punishment as fixed-term imprisonment, the number of rewards as 3, and the number of punishments as 0. The corresponding scoring method is: "the type of punishment matches and gets a base score of 60 points, and the number of rewards matches and gets a score of 30 points." Based on this, a scoring feature mapping set for R15 is constructed. Subsequently, the specific attribute data corresponding to each criminal case is extracted from the original criminal data set, and the specific attribute data is precisely compared with the above scoring feature mapping set one by one.
[0047] For example, the specific attribute data of a criminal with the ID 001 is: type of punishment: fixed-term imprisonment, number of rewards: 3, number of punishments: 0. This data perfectly matches the R15 scoring feature mapping set. Therefore, it is determined that the attribute data perfectly matches rule R15. Furthermore, based on the above precise comparison results, the scoring source and specific scoring value of each attribute are clearly determined, and the matching relationship of each attribute score corresponding to the rule entry is recorded in a structured manner, forming a clear scoring chain for sentence reduction and parole. A specific scoring chain example is: (Criminal 001 → Rule R15 → type of punishment: 60 points, number of rewards: 30 points, number of punishments: 10 points, total score: 100 points). Finally, based on the specific scoring values of the clear scoring chain and the corresponding attribute combination features, a sentence reduction recommendation that can be clearly traced back to the rule source is generated. For example, based on the total score of 100 points in the clear scoring chain, the recommended sentence reduction is 6 months.
[0048] This embodiment significantly improves the transparency and traceability of the scoring process for sentence reduction and parole recommendations by clearly defining the source of the scoring and the correspondence between the attributes, effectively solving the problems of ambiguity and difficulty in retrospection in the traditional scoring method.
[0049] Based on precisely compared attribute data and a set of scoring feature maps, the specific scoring source and score value are determined, generating a clear scoring chain for sentence reduction and parole, including: Based on the specific attribute data of the criminal case and the attribute combination characteristics corresponding to the rule entries to be used, the specific score corresponding to each attribute is determined item by item. Construct a set of rating records for each attribute based on the specific rating; Extract all attribute score values from the score record set, and calculate the weighted score values of the attributes according to the weight requirements of the score feature mapping set; The weighted scoring values are organized and recorded in the logical order of the rule entries to be used, forming a clear scoring chain for sentence reduction and parole.
[0050] In this embodiment, the specific scoring source and scoring value are determined based on the accurately compared attribute data and the scoring feature mapping set, and a clear scoring chain for commutation and parole is generated. The purpose is to ensure that the scoring composition in the commutation and parole case review process is clear, specific and traceable, so as to solve the problems of unclear scoring source and difficult data traceability in the prior art. Specifically, firstly, based on the specific attribute data of the criminal case and the attribute combination characteristics of the rule entries to be used, the specific score corresponding to each attribute is determined item by item to ensure that there is a clear and specific one-to-one correspondence between the attribute data and the scoring standards of the rule entries. For example, for criminal case number 001, the specific attribute data includes (crime type: fixed-term imprisonment, number of rewards: 3, number of punishments: 0). The attribute combination feature mapping of the pending rule entry R15 clarifies that the crime type score is 60 points, the number of rewards score is 30 points, and the number of punishments score is 10 points. Subsequently, based on the specific scores determined above, a score record set from each attribute to the specific score is constructed item by item. The score record set can be specifically represented as {(crime type, 60 points), (number of rewards, 30 points), (number of punishments, 10 points)}.
[0051] Next, all attribute score values are extracted from the score record set, and the weighted score value of each attribute is calculated according to the explicit weight requirements of each attribute in the score feature mapping set. For example, the weight of the type of punishment is 0.5, the weight of the number of rewards is 0.3, and the weight of the number of punishments is 0.2. The specific weighted score value after calculation is: 60×0.5+30×0.3+10×0.2=39 points. Finally, the weighted score values obtained above are organized and recorded according to the explicit logical order of the rules to be used, forming a clear and complete score chain for sentence reduction and parole.
[0052] This embodiment significantly improves the objectivity and transparency of the scoring process through the clear construction and organization of the scoring chain described above, making the review process and final recommendation output of commutation and parole cases clearly traceable, and overcoming the problems of unclear scoring sources and unclear composition in the prior art. Example
[0053] Please see Figure 2 As shown, based on the same inventive concept, this embodiment discloses a multi-dimensional review system for commutation and parole cases. For details not covered in this embodiment, please refer to the relevant sections of Embodiment 1. The system includes: Data aggregation module: used to obtain data on prisoners' performance during their sentences from the prisoner file system and historical data on the handling of commutation and parole cases from the case handling system, and to form a set of original prisoner data based on the prisoner performance data and historical data. Attribute mapping module: used to extract a multi-dimensional attribute set of commutation and parole rules from the original criminal data set, and to build an attribute association mapping table based on the multi-dimensional attribute set; Overlap identification module: used to calculate the matching overlap degree between attributes based on the attribute association mapping table, and determine the set of target rule entries with overlap based on the matching overlap degree; Frequency filtering module: used to calculate the applicable frequency of each rule entry based on the historical matching records of the target rule entry set, and to filter out the set of rule entries to be used based on the applicable frequency; Scoring Link Module: This module is used to match and analyze the original data set of criminals based on the set of rules to be used, generate a scoring chain for sentence reduction and parole, and generate traceable suggestions for the extent of sentence reduction based on the scoring chain for sentence reduction and parole.
[0054] The accompanying drawings of the embodiments of this invention only involve the structures involved in the embodiments of this invention. Other structures can refer to the general design. In the absence of conflict, the features of the same embodiment and different embodiments of this invention can be combined with each other. The above are only specific implementations of this invention, but the protection scope of this invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this invention should be included within the protection scope of this invention. Therefore, the protection scope of this invention should be determined by the protection scope of the claims.
Claims
1. A method for reviewing commutation and parole cases based on multiple dimensions, characterized in that, include: Data on prisoners' performance during their sentences is obtained from the prisoner file system, and historical data on commutation and parole cases is obtained from the case handling system. Based on the prisoner performance data and historical data, a set of original prisoner data is formed. Extract a multi-dimensional attribute set of commutation and parole rules from the original dataset of criminals, and construct an attribute association mapping table based on the multi-dimensional attribute set; The matching overlap between attributes is calculated based on the attribute association mapping table, and the set of overlapping target rule entries is determined based on the matching overlap. The applicable frequency of each rule entry is calculated based on the historical matching records of the target rule entry set, and a set of rule entries to be used is selected based on the applicable frequency. The original dataset of criminals is matched and analyzed based on the set of rules to be used to generate a commutation and parole scoring chain, and a traceable commutation range recommendation is generated based on the commutation and parole scoring chain.
2. The method for reviewing commutation and parole cases based on multiple dimensions as described in claim 1, characterized in that, A multidimensional attribute set of commutation and parole rules is extracted from the original dataset of criminals, and an attribute association mapping table is constructed based on the multidimensional attribute set, including: Identify the specific attribute categories involved in the rules for commutation and parole based on the historical processing data in the original criminal data set, and assign the attribute categories under the same processing situation to the corresponding category set; Extract the values of the corresponding attributes for each attribute category set, and categorize and organize the attribute values to form an attribute category value set; A cross-comparison method is used to compare each pair of attribute category value sets one by one in order to identify the cross attribute values and record them as attribute cross-mapping sets; The above attribute cross-mapping set is structured according to the cross-strength of values between attribute categories to obtain an attribute association mapping table.
3. The method for reviewing commutation and parole cases based on multiple dimensions according to claim 2, characterized in that, A cross-comparison method is used to compare each pair of attribute category value sets one by one to identify the cross-attribute values and record them as an attribute cross-mapping set, including: Extract specific attribute values from the two attribute category value sets respectively, and compare the numerical values or categories of the attribute values one by one to see if they are the same. Based on the attribute value comparison results, record whether each attribute pair has an overlap, and accumulate the number of overlapping attribute values. The total number of attribute values and the number of cross attribute values within the value set of each attribute category are calculated to obtain the attribute cross value ratio; The cross attribute values are determined based on the attribute categories whose cross value ratio exceeds a preset ratio, and recorded as an attribute cross mapping set.
4. The method for reviewing commutation and parole cases based on multi-dimensional dimensions according to claim 1, characterized in that, The matching overlap degree between attributes is calculated based on the attribute association mapping table, and the set of overlapping target rule entries is determined based on the matching overlap degree, including: The attribute association mapping table is used to parse the attribute combination of rule entries in the preset commutation and parole rule base, and the attribute combination feature set corresponding to each rule entry is recorded. Cross-matching is performed on the attribute combination feature sets of any two rule entries, and the number of overlapping features of corresponding attributes between each pair of rule entries is counted one by one. The overlapping attribute ratio is calculated based on the number of overlapping attribute features in each rule entry pair, and the rule entry pairs are sorted according to the overlapping attribute ratio to obtain a rule overlap sorted set. Select rule entries from the rule overlap sorting set whose overlap attribute ratio meets the preset overlap conditions, and extract their corresponding rule entries to form the target rule entry set.
5. The method for reviewing commutation and parole cases based on multiple dimensions according to claim 4, characterized in that, The overlapping attribute ratio is calculated based on the number of overlapping attribute features in each rule pair, and the rule pairs are sorted according to the overlapping attribute ratio to obtain a rule overlap sorting set, including: Calculate the number of all attribute combinations for each rule entry; Calculate the number of overlapping attribute features shared between each pair of rule entries; The specific percentage of overlapping attributes is obtained by dividing the number of overlapping attribute features by the total number of attribute combination features for each rule entry pair. Based on the specific overlapping attribute ratio values, the rule entries are sorted in descending order to obtain the rule overlap sorted set.
6. The method for reviewing commutation and parole cases based on multiple dimensions according to claim 1, characterized in that, The applicability frequency of each rule entry is calculated based on the historical matching records of the target rule entry set, and a set of rule entries to be used is selected based on the applicability frequency, including: Based on the target set of rule entries, extract the set of matching records for each rule entry from the historical processing data, and calculate the number of matching records for each rule entry; Count the number of matched records for all target rule entries, and calculate the matching percentage for each rule entry accordingly; Based on the matching ratio, the rule entries in the target rule entry set are sorted in descending order to obtain the rule application frequency set; Select rule entries with a matching rate higher than a preset rate threshold from the set of rule application frequencies, and filter out the set of rule entries to be used.
7. The method for reviewing commutation and parole cases based on multiple dimensions as described in claim 6, characterized in that, The rule entries in the target rule entry set are sorted in descending order based on the matching ratio to obtain the rule application frequency set, which includes: Calculate the percentage of each rule entry relative to the total number of matches in the target rule entry set based on the number of matched records; The rule entries are arranged in order of their matching percentage to form a rule percentage sorting list; Mark the matching percentage value corresponding to each rule entry in the sorted list according to the rule percentage sorting list; Based on the rule proportion sorting list and the matching proportion values of its tags, a set of rule application frequencies is generated.
8. The method for reviewing commutation and parole cases based on multiple dimensions according to claim 1, characterized in that, The original dataset of offenders is matched and analyzed based on the set of rules to be used to generate a commutation and parole scoring chain. Based on this scoring chain, a traceable commutation range recommendation is generated, including: Based on the set of rule entries to be used, extract the attribute combination features and scoring methods corresponding to each rule entry to generate a set of scoring feature maps; Extract specific attribute data for each criminal case from the original criminal data set, and compare each specific attribute data item by item with the scoring feature mapping set; Based on the precisely compared attribute data and the set of scoring feature maps, the specific scoring source and scoring value are determined, and a clear scoring chain for sentence reduction and parole is generated. Suggestions for the extent of sentence reduction are generated based on the specific score values and corresponding attribute combinations in the sentence reduction and parole scoring chain.
9. A method for reviewing commutation and parole cases based on multiple dimensions as described in claim 8, characterized in that, Based on precisely compared attribute data and a set of scoring feature maps, the specific scoring source and score value are determined, generating a clear scoring chain for sentence reduction and parole, including: Based on the specific attribute data of the criminal case and the attribute combination characteristics corresponding to the rule entries to be used, the specific score corresponding to each attribute is determined item by item. Construct a set of rating records for each attribute based on the specific rating; Extract all attribute score values from the score record set, and calculate the weighted score values of the attributes according to the weight requirements of the score feature mapping set; The weighted scoring values are organized and recorded in the logical order of the rule entries to be used, forming a clear scoring chain for sentence reduction and parole.