Community correction scheme intelligent recommendation method and device, equipment and medium
By constructing a community corrections case library and using a pre-trained word embedding model for personalized program recommendations, the problem of a lack of personalization in community corrections programs was solved, achieving accurate program recommendations and improving correction efficiency and effectiveness.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- NO 15 INST OF CHINA ELECTRONICS TECH GRP
- Filing Date
- 2025-11-28
- Publication Date
- 2026-04-17
AI Technical Summary
Existing community correction programs lack personalization, resulting in low correction efficiency and failing to achieve precise correction with "one policy for each person".
A community corrections case database is constructed. By serializing key-value items of electronic file information, corrections cases, and recidivism information, word vector conversion is performed using a pre-trained word embedding model. Combined with semantic summarization and similarity matching, candidate cases are screened and weighted to generate personalized correction plans.
This improved the efficiency and accuracy of community corrections, ensuring that correction programs closely match the actual needs of offenders and enhancing the effectiveness of corrections.
Smart Images
Figure CN121880642A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of this disclosure relate to the field of personalized recommendation, and more specifically, to a method, apparatus, device, and medium for intelligent recommendation of a community corrections program. Background Technology
[0002] Community corrections refers to a non-custodial form of punishment in which eligible offenders are placed in a community environment and, under the guidance of specialized state agencies, in conjunction with relevant social groups, non-governmental organizations, and social volunteers, work to correct the offenders' criminal psychology and behavioral vices within the time limit determined by the judgment, ruling, or decision. The aim is to promote their smooth reintegration into society.
[0003] In related technologies, community correction programs are currently mainly formulated by judicial office staff based on experience. Due to the numerous and complex factors that need to be considered in correction strategies, such as the type of crime committed by the community correction subject, psychological state, education level, family environment, and whether they are minors, the grassroots level does not have enough energy and ability to accurately formulate correction programs. The programs adopted by the same judicial office are basically the same, which is still far from the precise correction goals of "one policy for one person" and "personalized correction".
[0004] However, using existing technology results in low correction efficiency. Summary of the Invention
[0005] The embodiments described herein provide a method, apparatus, device, and medium for intelligent recommendation of community correction programs, overcoming the aforementioned problems.
[0006] Firstly, based on the content of this disclosure, a method for intelligently recommending community correction programs is provided, including: A community corrections case database is constructed based on the electronic records, corrections cases, recidivism information, and resettlement and assistance cases of historical corrections subjects. The electronic file information and / or recidivism information corresponding to the target correctional object are serialized using key-value pairs to transform the electronic file information and / or recidivism information corresponding to the target correctional object into the corresponding target case text; The target case text corresponding to the target correction object is input into the pre-trained word embedding model to perform word vector conversion on the target case text corresponding to the target correction object, so as to obtain the case mapping word embedding vector corresponding to the target correction object; Semantic summarization of the text words in the community correction case database is performed to obtain case semantic representation vectors; Perform similarity matching between the semantic representation vector of the case and the embedding vector of the case mapping word corresponding to the target correction object; Based on the similarity matching results between the case mapping word embedding vector corresponding to the target correction object and the case semantic representation vector, multiple candidate correction cases are selected from the community correction case library; Obtain the target correction plan corresponding to each candidate correction case from the correction plan library; Based on the similarity between different target correction schemes, the target correction schemes corresponding to multiple candidate correction cases are weighted and integrated to obtain the recommended correction scheme corresponding to the target correction object.
[0007] Secondly, according to the content of this disclosure, a smart recommendation device for community correction programs is provided, comprising: The module is used to build a community corrections case library based on the electronic records of historical corrections subjects, corrections cases, recidivism information, and resettlement and assistance cases. The conversion module is used to serialize the electronic file information and / or recidivism information corresponding to the target correctional object into key-value items, so as to convert the electronic file information and / or recidivism information corresponding to the target correctional object into the corresponding target case text; The determination module is used to input the target case text corresponding to the target correction object into the pre-trained word embedding model to perform word vector conversion on the target case text corresponding to the target correction object, so as to obtain the case mapping word embedding vector corresponding to the target correction object; The aggregation module is used to perform semantic aggregation on the text words in the community correction case library to obtain the case semantic representation vector; The matching module is used to perform similarity matching between the semantic representation vector of the case and the embedding vector of the case mapping word corresponding to the target correction object; The filtering module is used to filter out multiple candidate correction cases from the community correction case library based on the similarity matching result between the case mapping word embedding vector corresponding to the target correction object and the case semantic representation vector; The acquisition module is used to acquire the target correction plan corresponding to each candidate correction case from the correction plan library; The integration module is used to perform weighted integration of the target correction schemes corresponding to multiple candidate correction cases based on the similarity between different target correction schemes, so as to obtain the recommended correction scheme corresponding to the target correction object.
[0008] Thirdly, a computer device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the steps of the intelligent recommendation method for community correction schemes as described in any of the above embodiments.
[0009] Fourthly, a computer-readable storage medium is provided, on which a computer program is stored, and when executed by a processor, the computer program implements the steps of the intelligent recommendation method for community correction schemes as described in any of the above embodiments.
[0010] The intelligent recommendation method for community correction programs provided in this application constructs a community correction case library based on the electronic file information, correction cases, recidivism information, and resettlement and assistance cases corresponding to historical correction subjects. It performs key-value serialization on the electronic file information and / or recidivism information corresponding to the target correction subject to transform it into corresponding target case text. The target case text corresponding to the target correction subject is then input into a pre-trained word embedding model to perform word vector conversion, obtaining the case mapping word embedding vector corresponding to the target correction subject. The process involves: semantically summarizing textual terms in the community corrections case database to obtain case semantic representation vectors; performing similarity matching between the case semantic representation vectors and the case mapping word embedding vectors corresponding to the target corrections object; selecting multiple candidate corrections cases from the community corrections case database based on the similarity matching results; obtaining the target correction plan corresponding to each candidate corrections case from the correction plan database; and weighting and integrating the target correction plans corresponding to multiple candidate corrections cases based on the similarity between different target correction plans to obtain a recommended correction plan for the target corrections object. In this way, by constructing a community corrections case database to summarize historical community corrections cases and matching them with actual offender cases to obtain recommended correction plans, the efficiency of community offender correction can be effectively improved.
[0011] The above description is merely an overview of the technical solutions of the embodiments of this application. In order to better understand the technical means of the embodiments of this application and to implement them in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the embodiments of this application more obvious and understandable, specific implementation methods of this application are described below. Attached Figure Description
[0012] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments will be briefly described below. It should be understood that the drawings described below only relate to some embodiments of this disclosure and are not intended to limit this disclosure, wherein: Figure 1 This is a flowchart illustrating an intelligent recommendation method for community correction programs provided in this publication.
[0013] Figure 2 This is a schematic diagram of the structure of an intelligent recommendation device for community correction programs provided in this disclosure.
[0014] Figure 3 This is a schematic diagram of the structure of a computer device provided in this disclosure.
[0015] It should be noted that the elements in the attached diagram are schematic and not drawn to scale. Detailed Implementation
[0016] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this disclosure. All other embodiments obtained by those skilled in the art based on the described embodiments of this disclosure without creative effort are also within the scope of protection of this disclosure.
[0017] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this subject matter pertains. It will be further understood that terms such as those defined in commonly used dictionaries shall be interpreted as having the meaning consistent with their meaning in the context of the specification and in the relevant art, and shall not be interpreted in an idealized or overly formal form unless otherwise explicitly defined herein. As used herein, the statement of “connecting” or “coupling” two or more parts together shall mean that these parts are directly joined together or joined through one or more intermediate components.
[0018] The term "embodiment" as used herein means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of the phrase "embodiment" in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0019] In this document, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists, A and B exist simultaneously, or B exists. Additionally, the character " / " generally indicates that the preceding and following related objects have an "or" relationship. Terms such as "first" and "second" are only used to distinguish one component (or part of a component) from another component (or another part of a component).
[0020] In the description of this application, unless otherwise stated, "multiple" means two or more (including two), and similarly, "multiple groups" means two or more (including two groups).
[0021] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.
[0022] Figure 1 This is a flowchart illustrating an intelligent recommendation method for community correction programs provided in an embodiment of this disclosure, as shown below. Figure 1 As shown, the specific process of the intelligent recommendation method for community correction programs includes: S110. Construct a community correction case database based on the electronic file information, correction cases, recidivism information, and resettlement and assistance cases corresponding to historical correction subjects.
[0023] The electronic files for individuals under past correctional supervision may include, but are not limited to, basic personal information, criminal records, behavioral performance during correction, psychological assessment reports, and recidivism risk assessments. Correctional cases are specific cases involving individuals under past correction during their correction period, reflecting their behavioral patterns, psychological state, and key events in the correction process. Recidivism information indicates whether an individual under past correction has a history of recidivism during or after correction, reflecting the effectiveness of correctional measures and the individual's recidivism risk. Resettlement and assistance cases are specific examples of individuals under past correction after their correction period, detailing the various measures taken, their effectiveness, and the individual's social adaptation.
[0024] Based on the electronic records of historical correctional subjects, correctional cases, recidivism information, and resettlement and assistance cases, a community corrections case database is constructed. This may include: classifying and organizing the electronic records of historical correctional subjects, extracting key data fields, and forming a structured archive database; analyzing behavioral patterns and psychological characteristics in correctional cases, and summarizing representative case types; evaluating the actual effectiveness of different correctional measures in reducing the risk of recidivism through recidivism information, and quantifying the relevant data; and evaluating the effectiveness of assistance measures in resettlement and assistance cases, selecting assistance programs with high success rates, thereby constructing the community corrections case database.
[0025] S120. Perform key-value serialization on the electronic file information and / or recidivism information corresponding to the target correctional object to transform the electronic file information and / or recidivism information corresponding to the target correctional object into the corresponding target case text.
[0026] Key-value item serialization is used to structure relevant information of the target correction object according to preset rules, ensuring that each piece of data can be stored and retrieved in a unified format.
[0027] In some embodiments, the electronic file information and / or recidivism information corresponding to the target correctional object are serialized using key-value pairs to transform the electronic file information and / or recidivism information corresponding to the target correctional object into the corresponding target case text. This includes: structuring the electronic file information and / or recidivism information of the target correctional object according to preset rules to form a target data set in the form of key-value pairs; each key-value pair contains a specific attribute and its corresponding attribute value; performing data anonymization processing on each key-value pair; and serializing each key-value pair using a preset serialization processing strategy to obtain the target case text corresponding to the target correctional object.
[0028] For example, information such as the age, gender, and type of crime of the target correctional subject can be used as keys, and their specific values or descriptions can be used as values to form a clear data structure. By anonymizing these key-value pairs, personal privacy can be effectively protected while ensuring data security and compliance.
[0029] The pre-defined serialization strategy uses JSON format to serialize key-value pairs, converting each pair into a string and ensuring that the logical relationships between each pair are clearly discernible. Furthermore, by setting specific encoding rules, the generated target example text can be guaranteed to be compatible and readable across different platforms.
[0030] Therefore, by pre-setting a serialization processing strategy, the storage and transmission of data are further standardized, making the generated target case text not only easy to analyze later, but also able to flow efficiently between different systems, effectively improving the efficiency of data processing.
[0031] S130. Input the target case text corresponding to the target correction object into the pre-trained word embedding model to perform word vector conversion on the target case text corresponding to the target correction object, and obtain the case mapping word embedding vector corresponding to the target correction object.
[0032] The pre-trained word embedding model, trained on a large-scale corpus, captures semantic information and contextual relationships within the text. By inputting the target case text into this model, it transforms the string into a numerical representation in a high-dimensional vector space. This numerical representation not only preserves the core features of the original text but also effectively reflects the similarities and differences between different texts. Specifically, the pre-trained word embedding model processes each word in the target case text and generates corresponding word vectors based on their distribution in the corpus. Then, by performing weighted averaging or specific aggregation operations on these word vectors, it forms a case-mapped word embedding vector representing the entire target case text.
[0033] S140. Semantically summarize the text words in the community correction case database to obtain the case semantic representation vector.
[0034] By semantically summarizing the text words in the community corrections case database, vector representations reflecting the core features of the cases can be extracted. Specifically, natural language processing techniques can be combined during semantic summarization to identify and extract key information from the text, while contextual relationships are used to enhance the understanding of word meanings.
[0035] In some embodiments, semantic summarization of text words in the community correction case library is performed to obtain case semantic representation vectors, including: extracting corresponding text words from the community correction case library using a text word extraction model; performing word segmentation and part-of-speech tagging on the text words in the community correction case library to obtain correction key terms corresponding to the community correction case library; performing term mapping on the correction key terms corresponding to the community correction case library using a pre-trained word vector model to obtain a term mapping representation vector corresponding to each correction key term; performing weighted processing on the term mapping representation vector corresponding to each correction key term using context information to obtain a weighted mapping representation vector corresponding to each correction key term; and aggregating the weighted mapping representation vectors corresponding to multiple correction key terms to obtain a case semantic representation vector.
[0036] Among them, the text words are representative words selected from the community correction case database to reflect the core content and characteristics of the cases.
[0037] The process involves segmenting and tagging the text words in the community corrections case database to obtain the corresponding corrections key terms. This can include: segmenting the text words using a word segmentation tool to identify the independent semantic unit of each word; and combining this with a part-of-speech tagging algorithm to determine the grammatical role and attributes of each independent semantic unit in the sentence, so as to select corrections key terms closely related to the field of community corrections.
[0038] By weighting the term mapping representation vector corresponding to each correction key term using contextual information, a weighted mapping representation vector corresponding to each correction key term is obtained. This may include: calculating the corresponding weight coefficient by analyzing the position, frequency and correlation of each correction key term in the context with other terms; and multiplying the weight coefficient with the corresponding term mapping representation vector to generate the weighted mapping representation vector.
[0039] Therefore, vector aggregation processing can not only improve the efficiency of data processing, but also enhance the accuracy of semantic expression, making the recommendation of community correction programs more in line with actual needs and making it easier to capture the core semantic information of community correction cases more accurately.
[0040] S150. Perform similarity matching between the case semantic representation vector and the case mapping word embedding vector corresponding to the target correction object; based on the similarity matching results between the case mapping word embedding vector corresponding to the target correction object and the case semantic representation vector, select multiple candidate correction cases from the community correction case library.
[0041] Specifically, the semantic association between the case semantic representation vector and the case mapping word embedding vector corresponding to the target correction object can be quantified by calculating the cosine similarity between the two. During the screening process, candidate correction cases with higher similarity are prioritized to ensure the relevance and targeting of the recommendation results.
[0042] In some embodiments, similarity matching is performed between the case semantic representation vector and the case mapping word embedding vector corresponding to the target correction object, including: calculating the spatial similarity between the case semantic representation vector and the case mapping word embedding vector corresponding to the target correction object; using the target correction object as a constraint, and filtering out multiple cases that have a matching relationship with the target correction object based on the spatial similarity between the case semantic representation vector and the case mapping word embedding vector; the cases include correction cases and / or placement and assistance cases.
[0043] Spatial similarity measures the closeness between the semantic representation vector of a case and the word embedding vector of the case mapping in the semantic space. By introducing the target correction object as a constraint, the selection process can be further optimized to ensure that the selected cases are highly consistent with the actual needs of the target correction object.
[0044] S160. Obtain the target correction plan corresponding to each candidate correction case from the correction plan library; based on the similarity between different target correction plans, perform weighted integration of the target correction plans corresponding to multiple candidate correction cases to obtain the recommended correction plan corresponding to the target correction object.
[0045] When weighted integrating multiple corrective measures, the characteristics of multiple dimensions such as corrective effect, implementation difficulty and resource consumption can be comprehensively considered; and by assigning corresponding weights to each dimension, the actual applicability of different corrective measures can be reflected more scientifically.
[0046] In some embodiments, obtaining the target correction plan corresponding to each candidate correction case from the correction plan library includes: extracting the historical correction plan corresponding to each candidate correction case from the correction plan library; analyzing the correction measure type, correction implementation cycle and effect evaluation index of each historical correction plan; and modifying each historical correction plan according to the analysis results of each historical correction plan and the correction goal of the target correction object to obtain the modified correction plan corresponding to each candidate correction case; and introducing correction expert knowledge to verify the modified correction plan corresponding to each candidate correction case to obtain the target correction plan corresponding to each candidate correction case.
[0047] Correctional measures can take various forms, including psychological counseling, behavioral restraint, education and training, and social services. Each type has its specific applicable scenarios and target groups, and can provide personalized support to meet the needs of different individuals undergoing correction.
[0048] The implementation cycle for corrective measures can be dynamically adjusted based on the specific circumstances of the target group and historical data. The length of the implementation cycle is affected by the complexity of the corrective goals, the type of corrective measures, and resource allocation. For short-term corrective goals, a shorter implementation cycle can be set to quickly verify the effects and adjust the plan in a timely manner; for long-term goals, the cycle needs to be designed in stages, and the effects should be evaluated and the plan optimized after each stage.
[0049] The setting of effectiveness evaluation indicators can comprehensively consider multiple dimensions such as the achievement of correctional goals, feedback from the corrected individuals, and resource utilization efficiency. Effectiveness evaluation indicators can include quantitative data and qualitative analysis. Quantitative data includes the completion rate of corrective measures and specific numerical values of behavioral improvement in corrected individuals; qualitative analysis includes subjective evaluations such as changes in the corrected individuals' psychological state and improvements in social adaptability.
[0050] Therefore, revising the treatment plan can effectively improve its accuracy and adaptability. Furthermore, incorporating expert knowledge to validate the plan not only enhances its authority but also effectively mitigates potential implementation risks.
[0051] In some embodiments, the target correction schemes corresponding to multiple candidate correction cases are weighted and integrated based on the similarity between different target correction schemes to obtain a recommended correction scheme for the target correction object. This includes: determining the corresponding weighting coefficient based on the matching degree of key elements of each target correction scheme; adjusting the weighting coefficient of each target correction scheme based on the expert verification instructions for the weighting coefficient of each target correction scheme and the feedback of historical data; and integrating the core correction measures, correction implementation cycle, and correction evaluation indicators in each target correction scheme based on the weighting coefficient of each target correction scheme to obtain a recommended correction scheme for the target correction object.
[0052] For example, in behavior improvement programs for juvenile community correction subjects, if certain cases show a high degree of fit in core correction measures, they can be assigned a higher weighting coefficient. Simultaneously, the weighting of each program can be dynamically adjusted by combining the practical experience of correction experts with feedback from successful cases in historical data. This not only enhances the scientific rigor of recommended programs but also ensures that correction measures are more targeted and feasible. Furthermore, by comparing the differences in implementation cycles and evaluation indicators among different programs, resource allocation can be further optimized, improving overall correction efficiency.
[0053] In this embodiment, a community corrections case library is constructed based on the electronic file information, corrections cases, recidivism information, and resettlement and assistance cases corresponding to historical corrections subjects. Key-value serialization is performed on the electronic file information and / or recidivism information corresponding to the target corrections subject to convert them into corresponding target case text. The target case text corresponding to the target corrections subject is input into a pre-trained word embedding model to perform word vector conversion, obtaining the case mapping word embedding vector corresponding to the target corrections subject. The community corrections case... The text words in the database are semantically summarized to obtain case semantic representation vectors. Similarity matching is performed between these case semantic representation vectors and the case mapping word embedding vectors corresponding to the target correctional object. Based on the similarity matching results between the case mapping word embedding vectors and the case semantic representation vectors, multiple candidate correctional cases are selected from the community correctional case database. The target correctional plan corresponding to each candidate correctional case is obtained from the correctional plan database. The target correctional plans corresponding to multiple candidate correctional cases are weighted and integrated based on the similarity between different target correctional plans to obtain a recommended correctional plan for the target correctional object. In this way, by constructing a community correctional case database and summarizing historical community correctional cases, recommended correctional plans are obtained by matching them with actual offender cases in correctional settings, thereby effectively improving the efficiency of community offender correction.
[0054] In some embodiments, the method further includes: conducting community correction for the target correction subject using the recommended correction plan corresponding to the target correction subject, and collecting behavioral feedback data of the target correction subject during the correction period; optimizing the recommended correction plan corresponding to the target correction subject based on the behavioral feedback data of the target correction subject during the correction period, and continuing community correction for the target correction subject using the optimized recommended correction plan.
[0055] This can be achieved through short-term reinforcement learning of the recommended correction plan, allowing it to quickly adapt to changes in the target individual's behavior and feedback characteristics. During reinforcement learning, real-time behavioral data from the target individual can be used to dynamically adjust key parameters and execution strategies within the recommended correction plan, thereby improving the accuracy and adaptability of the correction effect. Simultaneously, multi-dimensional evaluation indicators can be introduced to continuously monitor and analyze the optimized correction plan, ensuring that the correction measures effectively address the actual needs and potential risks of the target individual.
[0056] In summary, the method in this embodiment can process community correction technical specifications, recidivism influencing factors, etc., through knowledge management and accumulate them into a domain knowledge base. For new cases, the recommended solutions are corrected by manual intervention to form new revised solutions, which are then added to the database. Subsequently, the intelligent recommendation algorithm learns and corrects the solutions, and combines feedback information such as solution evaluation to actively optimize the recommendation strategy and gradually improve the accuracy and speed of recommendations.
[0057] Figure 2 This is a schematic diagram of the structure of an intelligent recommendation device for community correction programs provided in this embodiment. The intelligent recommendation device for community correction programs may include: Module 210 is used to build a community correction case library based on the electronic file information, correction cases, recidivism information and resettlement and assistance cases of historical correction subjects.
[0058] The conversion module 220 is used to serialize the electronic file information and / or recidivism information corresponding to the target correctional object into key-value items, so as to convert the electronic file information and / or recidivism information corresponding to the target correctional object into the corresponding target case text.
[0059] The determination module 230 is used to input the target case text corresponding to the target correction object into the pre-trained word embedding model, so as to perform word vector conversion on the target case text corresponding to the target correction object and obtain the case mapping word embedding vector corresponding to the target correction object.
[0060] The summary module 240 is used to semantically summarize the text words in the community correction case library to obtain the case semantic representation vector.
[0061] The matching module 250 is used to perform similarity matching between the semantic representation vector of the case and the embedding vector of the case mapping word corresponding to the target correction object.
[0062] The filtering module 260 is used to filter out multiple candidate correction cases from the community correction case library based on the similarity matching results between the case mapping word embedding vector and the case semantic representation vector corresponding to the target correction object.
[0063] The acquisition module 270 is used to acquire the target correction plan corresponding to each candidate correction case from the correction plan library.
[0064] The integration module 280 is used to perform weighted integration of the target correction schemes corresponding to multiple candidate correction cases based on the similarity between different target correction schemes, so as to obtain the recommended correction scheme corresponding to the target correction object.
[0065] In this embodiment, optionally, the conversion module 220 is specifically used for: The electronic file information and / or recidivism information of the target correctional object are structured according to preset rules to form a target data set in the form of key-value pairs; each key-value pair contains a specific attribute and its corresponding attribute value; data anonymization processing is performed on each key-value pair; and each key-value pair is serialized through a preset serialization processing strategy to obtain the target case text corresponding to the target correctional object.
[0066] In this embodiment, optionally, the summarization module 240 is specifically used for: Text words are extracted from the community correction case database using a text word extraction model. These words are then segmented and tagged with parts of speech to obtain key correction terms. A pre-trained word vector model is used to map these key terms, resulting in a term mapping vector for each key term. The term mapping vectors for each key term are then weighted using contextual information to obtain a weighted mapping vector. Finally, the weighted mapping vectors for multiple key terms are aggregated to obtain a case semantic representation vector.
[0067] In this embodiment, optionally, the matching module 250 is specifically used for: Calculate the spatial similarity between the case semantic representation vector and the case mapping word embedding vector corresponding to the target correction object. The spatial similarity is used to measure the closeness between the case semantic representation vector and the case mapping word embedding vector in the semantic space. Using the target correction object as a constraint, select multiple cases that have a matching relationship with the target correction object based on the spatial similarity between the case semantic representation vector and the case mapping word embedding vector. The cases include correction cases and / or placement and assistance cases.
[0068] In this embodiment, optionally, the acquisition module 270 is specifically used for: The historical correction schemes corresponding to each candidate correction case are extracted from the correction scheme library; the correction measure type, correction implementation cycle and effect evaluation indicators of each historical correction scheme are analyzed; and based on the analysis results of each historical correction scheme and the correction goals of the target correction object, each historical correction scheme is revised to obtain the revised correction scheme corresponding to each candidate correction case; correction expert knowledge is introduced to verify the revised correction scheme corresponding to each candidate correction case to obtain the target correction scheme corresponding to each candidate correction case.
[0069] In this embodiment, optionally, the integration module 280 is specifically used for: The corresponding weighting coefficients are determined based on the degree of matching of the key elements of each target correction plan; in response to the expert verification instructions for the weighting coefficients of each target correction plan, the weighting coefficients of each target correction plan are adjusted based on the expert verification results and historical data feedback; based on the weighting coefficients of each target correction plan, the core correction measures, correction implementation cycle and correction evaluation indicators in each target correction plan are integrated to obtain the recommended correction plan for the target correction object.
[0070] In this embodiment, optionally, a processing module may also be included.
[0071] The processing module is used to conduct community correction for the target correction subject according to the recommended correction plan, and to collect behavioral feedback data of the target correction subject during the correction period; based on the behavioral feedback data of the target correction subject during the correction period, to optimize the recommended correction plan for the target correction subject, and to continue to conduct community correction for the target correction subject according to the optimized recommended correction plan.
[0072] The intelligent recommendation device for community correction programs provided in this disclosure can execute the above-described method embodiments. Its specific implementation principle and technical effects can be found in the above-described method embodiments, and will not be repeated here.
[0073] This application also provides a computer device. Please refer to the following for details. Figure 3 , Figure 3 This is a basic structural block diagram of the computer device in this embodiment.
[0074] The computer device includes a memory 310 and a processor 320 that are interconnected via a system bus. It should be noted that only a computer device with memory 310 and processor 320 is shown in the figure; however, it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented alternatively. Those skilled in the art will understand that the computer device described herein is a device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.
[0075] Computer devices can include desktop computers, laptops, handheld computers, and cloud servers. These devices allow for human-computer interaction with users through keyboards, mice, remote controls, touchpads, or voice-activated devices.
[0076] The memory 310 includes at least one type of readable storage medium, including non-volatile memory or volatile memory, such as flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. RAM may include static RAM or dynamic RAM. In some embodiments, the memory 310 may be an internal storage unit of a computer device, such as the hard disk or memory of the computer device. In other embodiments, the memory 310 may also be an external storage device of the computer device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, or flash card equipped on the computer device. Of course, the memory 310 may include both internal storage units and external storage devices of the computer device. In this embodiment, the memory 310 is typically used to store the operating system and various application software installed on the computer device, such as the program code of the method described above. In addition, the memory 310 can also be used to temporarily store various types of data that have been output or will be output.
[0077] Processor 320 is typically used to perform overall operations of a computer device. In this embodiment, memory 310 is used to store program code or instructions, including computer operation instructions, and processor 320 is used to execute the program code or instructions stored in memory 310 or process data, such as program code that runs the methods described above.
[0078] In this article, the bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. This bus system can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.
[0079] Another embodiment of this application also provides a computer-readable medium, which may be a computer-readable signal medium or a computer-readable medium. A processor in a computer reads computer-readable program code stored in the computer-readable medium, enabling the processor to execute the functional actions specified in each step or combination of steps in the above method; and to generate means for implementing the functional actions specified in each block or combination of blocks in the block diagram.
[0080] Computer-readable media include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared memory or semiconductor systems, devices or apparatuses, or any suitable combination thereof, wherein the memory is used to store program code or instructions, the program code including computer operation instructions, and the processor is used to execute the program code or instructions of the above-described methods stored in the memory.
[0081] The definitions of memory and processor can be found in the description of the foregoing computer device embodiments, and will not be repeated here.
[0082] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.
[0083] In the various embodiments of this application, the functional units or modules can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0084] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0085] In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" as described in this application does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. This application can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims listing several means, several units of these means may be embodied by the same item of hardware. The use of "first," "second," and "third," etc., does not indicate any order and these words should be interpreted as names. Unless otherwise specified, the steps in the above embodiments should not be construed as limiting the order of execution.
[0086] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.
Claims
1. A method for intelligently recommending community correction programs, characterized in that, include: A community corrections case database is constructed based on the electronic records, corrections cases, recidivism information, and resettlement and assistance cases of historical corrections subjects. The electronic file information and / or recidivism information corresponding to the target correctional object are serialized using key-value pairs to transform the electronic file information and / or recidivism information corresponding to the target correctional object into the corresponding target case text; The target case text corresponding to the target correction object is input into the pre-trained word embedding model to perform word vector conversion on the target case text corresponding to the target correction object, so as to obtain the case mapping word embedding vector corresponding to the target correction object; Semantic summarization of the text words in the community correction case database is performed to obtain case semantic representation vectors; Perform similarity matching between the semantic representation vector of the case and the embedding vector of the case mapping word corresponding to the target correction object; Based on the similarity matching results between the case mapping word embedding vector and the case semantic representation vector corresponding to the target correction object, multiple candidate correction cases are selected from the community correction case library; Obtain the target correction plan corresponding to each candidate correction case from the correction plan library; Based on the similarity between different target correction schemes, the target correction schemes corresponding to multiple candidate correction cases are weighted and integrated to obtain the recommended correction scheme corresponding to the target correction object.
2. The method according to claim 1, characterized in that, The electronic record information and / or recidivism information corresponding to the target correctional subject are serialized using key-value pairs to transform the electronic record information and / or recidivism information corresponding to the target correctional subject into the corresponding target case text, including: The electronic file information and / or recidivism information of the target correctional subjects are structured according to preset rules to form a target data set in the form of key-value pairs; each key-value pair contains a specific attribute and its corresponding attribute value; Data anonymization processing is performed on each key-value pair; and each key-value pair is serialized using a preset serialization processing strategy to obtain the target case text corresponding to the target correction object.
3. The method according to claim 1, characterized in that, Semantic summarization of the text words in the community corrections case database yields a case semantic representation vector, including: The corresponding text words are extracted from the community correction case database using a text word extraction model; and the text words in the community correction case database are segmented and tagged with parts of speech to obtain the correction key terms corresponding to the community correction case database. A pre-trained word vector model is used to map the correction key terms corresponding to the community correction case library to obtain the term mapping representation vector corresponding to each correction key term. The term mapping representation vector corresponding to each of the corrective key terms is weighted by the context information to obtain the weighted mapping representation vector corresponding to each of the corrective key terms; The weighted mapping representation vectors corresponding to multiple correction key terms are aggregated to obtain the case semantic representation vector.
4. The method according to claim 1, characterized in that, Perform similarity matching between the semantic representation vector of the case and the embedding vector of the case mapping word corresponding to the target correction object, including: Calculate the spatial similarity between the case semantic representation vector and the case mapping word embedding vector corresponding to the target correction object. The spatial similarity is used to measure the closeness between the case semantic representation vector and the case mapping word embedding vector in the semantic space. Using the target correction object as a constraint, multiple cases with matching relationships with the target correction object are selected based on the spatial similarity between the case semantic representation vector and the case mapping word embedding vector; the cases include correction cases and / or placement and assistance cases.
5. The method according to claim 1, characterized in that, Obtain the target correction plan corresponding to each candidate correction case from the correction plan library, including: Extract the historical correction plan corresponding to each candidate correction case from the correction plan library; Analyze the type of corrective measures, implementation period, and effectiveness evaluation indicators of each historical corrective scheme; and based on the analysis results of each historical corrective scheme and the corrective goals of the target corrective object, revise each historical corrective scheme to obtain the revised corrective scheme corresponding to each candidate corrective case. By introducing the knowledge of orthodontic experts, the modified orthodontic scheme corresponding to each candidate orthodontic case is verified, and the target orthodontic scheme corresponding to each candidate orthodontic case is obtained.
6. The method according to claim 1, characterized in that, Based on the similarity between different target correction schemes, the target correction schemes corresponding to multiple candidate correction cases are weighted and integrated to obtain the recommended correction scheme for the target correction object, including: The corresponding weighting coefficients are determined based on the degree of matching of the key elements of each target correction scheme. In response to the expert verification instruction for the weighting coefficients corresponding to each of the target correction schemes, the weighting coefficients corresponding to each of the target correction schemes are adjusted according to the expert verification results and historical data feedback for each of the target correction schemes. Based on the weighting coefficients corresponding to each target correction scheme, the core correction measures, correction implementation cycle, and correction evaluation indicators in each target correction scheme are integrated to obtain the recommended correction scheme for the target correction object.
7. The method according to claim 1, characterized in that, Also includes: The target correction subject is subject to community correction through the recommended correction plan, and the behavioral feedback data of the target correction subject during the correction period is collected. Based on the behavioral feedback data of the target correction subject during the correction period, the recommended correction plan corresponding to the target correction subject is optimized, and the target correction subject continues to undergo community correction through the optimized recommended correction plan.
8. A smart recommendation device for community correction programs, characterized in that, include: The module is used to build a community corrections case library based on the electronic records of historical corrections subjects, corrections cases, recidivism information, and resettlement and assistance cases. The conversion module is used to serialize the electronic file information and / or recidivism information corresponding to the target correctional object into key-value items, so as to convert the electronic file information and / or recidivism information corresponding to the target correctional object into the corresponding target case text; The determination module is used to input the target case text corresponding to the target correction object into the pre-trained word embedding model to perform word vector conversion on the target case text corresponding to the target correction object, so as to obtain the case mapping word embedding vector corresponding to the target correction object; The aggregation module is used to perform semantic aggregation on the text words in the community correction case library to obtain the case semantic representation vector; The matching module is used to perform similarity matching between the semantic representation vector of the case and the embedding vector of the case mapping word corresponding to the target correction object; The filtering module is used to filter out multiple candidate correction cases from the community correction case library based on the similarity matching result between the case mapping word embedding vector corresponding to the target correction object and the case semantic representation vector; The acquisition module is used to acquire the target correction plan corresponding to each candidate correction case from the correction plan library; The integration module is used to perform weighted integration of the target correction schemes corresponding to multiple candidate correction cases based on the similarity between different target correction schemes, so as to obtain the recommended correction scheme corresponding to the target correction object.
9. A computer device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the intelligent recommendation method for community correction programs as described in any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the intelligent recommendation method for community correction programs as described in any one of claims 1 to 7.