WO2021073410A1 - 法律证据的排序和推荐方法、装置、设备及存储介质 - Google Patents

法律证据的排序和推荐方法、装置、设备及存储介质 Download PDF

Info

Publication number
WO2021073410A1
WO2021073410A1 PCT/CN2020/118287 CN2020118287W WO2021073410A1 WO 2021073410 A1 WO2021073410 A1 WO 2021073410A1 CN 2020118287 W CN2020118287 W CN 2020118287W WO 2021073410 A1 WO2021073410 A1 WO 2021073410A1
Authority
WO
WIPO (PCT)
Prior art keywords
preset
evidence
target
weight
target feature
Prior art date
Application number
PCT/CN2020/118287
Other languages
English (en)
French (fr)
Inventor
凌岚
刘嘉伟
于修铭
陈晨
李可
汪伟
Original Assignee
平安科技（深圳）有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 平安科技（深圳）有限公司 filed Critical 平安科技（深圳）有限公司
Publication of WO2021073410A1 publication Critical patent/WO2021073410A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/3331—Query processing
    • G06F16/3332—Query translation
    • G06F16/3334—Selection or weighting of terms from queries, including natural language queries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/31—Indexing; Data structures therefor; Storage structures
    • G06F16/313—Selection or weighting of terms for indexing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/31—Indexing; Data structures therefor; Storage structures
    • G06F16/316—Indexing structures
    • G06F16/322—Trees
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/335—Filtering based on additional data, e.g. user or group profiles
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/338—Presentation of query results
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
    • G06F16/367—Ontology

Definitions

  • This application relates to the field of intelligent recommendation, and in particular to a method, device, device, and storage medium for sorting and recommending legal evidence.
  • Knowledge graph technology has increasingly become the foundation of artificial intelligence, and it is an important method for machines to understand natural language and construct knowledge networks.
  • knowledge graphs in the judicial field has quietly emerged. It helps practitioners to quickly retrieve relevant legal content online, thereby improving the quality and efficiency of court trials.
  • An important application of the legal knowledge map is to query or infer the evidence needed to resolve the dispute focus based on the legal knowledge map after determining the dispute focus of the parties to the litigation, and it is easy to ignore but plays an important role in past cases.
  • the evidence gives hints.
  • the main purpose of this application is to solve the technical problem of low accuracy in judging evidence or weight of evidence that can only rely on legal provisions or personal experience.
  • the first aspect of this application provides a method for sorting and recommending legal evidence, including: obtaining target characteristics of preset evidence items, where the preset evidence items are a set of evidence items in the focus of a preset dispute,
  • the target feature includes a preset evidence item frequency and a preset inverse dispute focus frequency; the target feature is processed according to a preset tree model to obtain the weight of the target feature; the comparison between the target feature and the target feature
  • the weight is calculated by weighted average to obtain the target weight of the preset evidence item; the target weight of the preset evidence item is stored in the preset legal knowledge graph; when the retrieval request sent by the target terminal is received, the target weight is obtained from the
  • the evidence guidance data is determined in the preset legal knowledge graph, and the evidence guidance data is pushed to the target terminal.
  • the evidence guidance data is used to indicate the preset evidence items whose target weight is greater than a preset threshold.
  • the second aspect of the present application provides a device for sorting and recommending legal evidence, including a memory, a processor, and computer-readable instructions stored on the memory and running on the processor, and the processor executes all
  • the computer-readable instruction implements the following steps: obtaining the target feature of the preset evidence item, the preset evidence item is a set of evidence items in the preset dispute focus, and the target feature includes the preset evidence item frequency and preset inverse The frequency of dispute focus; the target feature is processed according to the preset tree model to obtain the weight of the target feature; the weighted average calculation is performed on the target feature and the weight of the target feature to obtain the preset evidence item Target weight; store the target weight of the preset evidence item in a preset legal knowledge map; when a retrieval request sent by the target terminal is received, determine the evidence guide data from the preset legal knowledge map, and store all
  • the evidence guidance data is pushed to the target terminal, and the evidence guidance data is used to indicate a preset evidence item whose target weight is greater than a prese
  • the third aspect of the present application provides a computer-readable storage medium, the computer-readable storage medium stores computer instructions, and when the computer instructions run on the computer, the computer executes the following steps: Obtain preset evidence items
  • the preset evidence item is a set of evidence items in the preset dispute focus, and the target feature includes the preset evidence item frequency and the preset inverse dispute focus frequency; the target feature is performed according to the preset tree model Process to obtain the weight of the target feature; perform a weighted average calculation on the target feature and the weight of the target feature to obtain the target weight of the preset evidence item; store the target weight of the preset evidence item in In the preset legal knowledge map; when a retrieval request sent by the target terminal is received, evidence guidance data is determined from the preset legal knowledge map, and the evidence guidance data is pushed to the target terminal, the evidence guidance The data is used to indicate the preset evidence items whose target weight is greater than the preset threshold.
  • the fourth aspect of the application provides a device for sorting and recommending legal evidence, including: an acquiring unit for acquiring target characteristics of preset evidence items, where the preset evidence items are a set of evidence items in the focus of a preset dispute,
  • the target feature includes a preset evidence item frequency and a preset inverse dispute focus frequency;
  • a processing unit configured to process the target feature according to a preset tree model, to obtain the weight of the target feature;
  • a calculation unit configured to The weighted average calculation of the target feature and the weight of the target feature is performed to obtain the target weight of the preset evidence item;
  • a storage unit for storing the target weight of the preset evidence item in a preset legal knowledge graph ;
  • the push unit when receiving a search request sent by the target terminal, is used to determine evidence guidance data from the preset legal knowledge graph, and push the evidence guidance data to the target terminal, the evidence guidance data It is used to indicate the preset evidence item whose target weight is greater than the preset threshold.
  • the target feature of the preset evidence item is obtained, the preset evidence item is a set of evidence items in the preset dispute focus, and the target feature includes the preset evidence item frequency and the preset inverse dispute focus Frequency; process the target feature according to the preset tree model to obtain the weight of the target feature; perform a weighted average calculation on the target feature and the weight of the target feature to obtain the target weight of the preset evidence item ; Store the target weight of the preset evidence item in a preset legal knowledge map; when receiving a search request sent by the target terminal, determine the evidence guide data from the preset legal knowledge map, and combine the evidence The guidance data is pushed to the target terminal, and the evidence guidance data is used to indicate a preset evidence item whose target weight is greater than a preset threshold.
  • FIG. 1 is a schematic diagram of an embodiment of a method for sorting and recommending legal evidence in an embodiment of this application;
  • FIG. 2 is a schematic diagram of another embodiment of a method for sorting and recommending legal evidence in an embodiment of this application;
  • FIG. 3 is a schematic diagram of an embodiment of a device for sorting and recommending legal evidence in an embodiment of this application;
  • FIG. 4 is a schematic diagram of another embodiment of a device for sorting and recommending legal evidence in an embodiment of this application;
  • Fig. 5 is a schematic diagram of an embodiment of a device for sorting and recommending legal evidence in an embodiment of the application.
  • the embodiments of the present application provide a method, device, device, and storage medium for sorting and recommending legal evidence, which are used to extract the target feature and the weight of the target feature of the preset evidence item, and according to the target feature and the weight of the preset evidence item
  • the weight of the target feature is calculated to the weight of the preset evidence item, and the evidence item is sorted and pushed according to the weight of the preset evidence item, which improves the accuracy and efficiency of evidence guidance.
  • An embodiment of the method for sorting and recommending legal evidence in the embodiment of the present application includes:
  • the preset evidence items are the set of evidence items in the preset dispute focus, and the target characteristics include the preset evidence item frequency and the preset inverse dispute focus frequency;
  • the server obtains the target feature of the preset evidence item, the preset evidence item is a set of evidence items in the preset dispute focus, and the target feature includes the preset evidence item frequency and the preset inverse dispute focus frequency.
  • the frequency of preset evidence items is the ratio coefficient between the number of occurrences of preset evidence items in the cases corresponding to the same preset dispute focus and the total number of cases corresponding to the same preset dispute focus
  • the preset frequency of inverse dispute focus is the ratio of the preset evidence items in all preset evidence items.
  • the target features corresponding to the preset evidence items include the frequency of preset evidence items and the frequency of preset inverse dispute focus.
  • the purpose of constructing these two features is that the importance of the preset evidence items is as important as the preset evidence items in the case. The number of occurrences in all disputes increases proportionally, but at the same time, the occurrence frequency of preset evidence items in all disputes decreases inversely.
  • the target characteristics also include the number of times the evidence item supports the small element in the case corresponding to the same dispute focus/total number of cases corresponding to the same dispute focus, whether it is original evidence, whether it is transmitted evidence, whether it is direct evidence, whether it is Indirect evidence, whether it is documentary evidence, whether it is physical evidence, whether it is audio-visual material, whether it is witness testimony, whether it is a statement by the parties, whether it is an appraisal conclusion, and whether it is an inquest transcript, the specifics are not limited here.
  • the server processes the target feature according to the preset tree model to obtain the weight of the target feature.
  • the server uses a preset tree model to process the target feature to obtain the weight of the target feature, and the value range is greater than or equal to 0 and less than or equal to 1.
  • the preset tree model can be an extreme gradient boosting (XGBoost) algorithm.
  • the XGBoost algorithm gradually accumulates the prediction results by constructing an additive model to achieve the best prediction value given the target feature.
  • the best predicted value is the weight of the target feature.
  • weight refers to the degree of importance of a certain factor or indicator relative to a certain thing. It is different from the general proportion. It reflects not only the percentage of a certain factor or indicator, but also emphasizes the factor or indicator. The relative importance of, tends to contribute or importance.
  • the server performs a weighted average calculation on the target feature and the weight of the target feature to obtain the target weight of the preset evidence item. Specifically, the server reads the target feature and the weight of the target feature from the preset data dictionary; the server performs a weighted average calculation on the target feature and the weight of the target feature according to the preset formula to obtain the target weight of the preset evidence item; The feature is set as the key, the target weight is set as the value, and the target weight is stored in the preset data dictionary.
  • the preset data dictionary refers to the definition and description of data items, data structure, data flow, data storage, processing logic, etc. in advance, and its purpose is to make a detailed description of each element in the data flow chart.
  • the preset data dictionary includes a collection of preset evidence items and target characteristics.
  • the server stores the target weight of the preset evidence item in the preset legal knowledge graph. Specifically, the server sets the target weight of the preset evidence item as the attribute of the preset evidence item; the server imports the preset evidence item and the target weight of the preset evidence item into the preset legal knowledge graph, and the preset legal knowledge graph is based on Storage is performed in the form of triples, where triples are composed of entities, attributes associated with the entities, and attribute values corresponding to the attributes.
  • the knowledge graph refers to the use of graphs as the data structure to represent knowledge.
  • the knowledge graph includes the edges between nodes and nodes. Among them, nodes represent entities or attribute values of entities, and edges between nodes are used to represent entities. Attributes, preset legal knowledge map to express legal knowledge in a structured way of knowledge map.
  • the evidence guidance data is determined from the preset legal knowledge graph, and the evidence guidance data is pushed to the target terminal.
  • the evidence guidance data is used to indicate the preset with the target weight greater than the preset threshold. Evidence item.
  • the server determines the evidence guidance data from the preset legal knowledge graph and pushes the evidence guidance data to the target terminal.
  • the evidence guidance data is used to indicate the preset evidence whose target weight is greater than the preset threshold. item.
  • the server receives the search request sent by the target terminal; the server parses the search request sent by the target terminal to obtain the unique identification and query text of the target terminal, where the unique identification is used to indicate the target terminal; the server searches for preset legal knowledge according to the query text Atlas, get the preset evidence items and the weights of the preset evidence items; sort based on the weight of the preset evidence items, and set the preset evidence items with the target weight greater than the preset threshold as the evidence guidance data.
  • the evidence guidance data is in the instruction The location accurately finds the push information corresponding to the query text according to the preset legal knowledge graph; the server pushes the evidence guidance data to the target terminal according to the unique identifier.
  • the weight of the preset evidence item is calculated by extracting the target feature and the weight of the target feature of the preset evidence item, and calculating the weight of the preset evidence item according to the preset evidence item's target feature and the weight of the target feature.
  • the weight of the evidence is sorted and pushed to improve the accuracy and efficiency of evidence guidance.
  • another embodiment of the method for ranking and recommending legal evidence in the embodiment of the present application includes:
  • the preset evidence items are the set of evidence items in the preset dispute focus, and the target characteristics include the preset evidence item frequency and the preset inverse dispute focus frequency;
  • the target feature ⁇ in corresponding to the preset evidence item includes the preset evidence item frequency and the preset inverse dispute focus frequency.
  • the purpose of constructing these two features is that the importance of the preset evidence item increases with the preset evidence item frequency. The number of occurrences in the case increases proportionally, but at the same time, the occurrence frequency of the preset evidence items in all disputes decreases inversely.
  • the target characteristics also include the number of times the evidence item supports the small element in the case corresponding to the same dispute focus/total number of cases corresponding to the same dispute focus, whether it is original evidence, whether it is transmitted evidence, whether it is direct evidence, whether it is Indirect evidence, whether it is documentary evidence, whether it is physical evidence, whether it is audio-visual material, whether it is witness testimony, whether it is a statement by the parties, whether it is an appraisal conclusion, and whether it is an inquest transcript, the specifics are not limited here.
  • the server sets ( ⁇ i1 , ⁇ i2 ,..., ⁇ in ) according to the above target feature sequence as (0.3,0.4,0.8,1,0,1,0,1,0,0,0,0, 0,0).
  • the server processes the target feature according to the preset tree model to obtain the weight of the target feature. Specifically, the server generates the structural risk minimization model of the preset tree model according to the loss function L of the preset minimization target feature, and the structural risk minimization model is Among them, J(f) is the complexity of the structural risk minimization model; the server determines the complexity of each classification regression tree ⁇ (f), ⁇ (f) is Among them, T is the number of child nodes of the classification and regression tree,
  • is the modulus of the child node vector of the classification and regression tree, ⁇ is the difficulty of node segmentation, and ⁇ is the L2 regularization coefficient; the server sets a preset minimization target
  • the feature loss function L and the complexity of each classification regression tree ⁇ (f) are superimposed to obtain the initial function, which is as follows: The server calculates according to the initial function to obtain the objective function obj * ; the server determines all the segmentation points of the target feature according to the objective function o
  • the value range of ⁇ n is greater than or equal to 0 and less than or equal to 1.
  • weight refers to the degree of importance of a certain factor or indicator relative to a certain thing. It is different from the general proportion. It reflects not only the percentage of a certain factor or indicator, but also emphasizes the factor or indicator. The relative importance of, tends to contribute or importance.
  • the server performs a weighted average calculation on the target feature and the weight of the target feature to obtain the target weight of the preset evidence item. Specifically, the server reads the target feature ⁇ in and the target feature weight ⁇ n from the preset data dictionary; the server reads the target feature ⁇ in and the target feature weight ⁇ n according to the preset data dictionary.
  • the preset formula performs a weighted average calculation on the target feature ⁇ in and the target feature weight ⁇ n to obtain the target weight ⁇ i of the preset evidence item.
  • the target weight of the preset evidence item is proportional to the number of times the preset evidence item appears in the case, and the target weight of the preset evidence item is proportional to the frequency of the preset evidence item in all preset disputes. Inversely.
  • the server stores the target weight of the preset evidence item in the preset legal knowledge graph.
  • the server sets the target weight of the preset evidence item as the attribute of the preset evidence item; the server imports the preset evidence item and the target weight of the preset evidence item into the preset legal knowledge graph, and the preset legal knowledge graph is a triplet
  • the form is stored, where the triplet is composed of the entity, the attributes associated with the entity, and the attribute values corresponding to the attributes.
  • the knowledge graph refers to the use of graphs as the data structure to represent knowledge.
  • the knowledge graph includes the edges between nodes and nodes. Among them, nodes represent entities or attribute values of entities, and edges between nodes are used to represent entities. Attributes, preset legal knowledge map to express legal knowledge in a structured way of knowledge map.
  • the target terminal When receiving the retrieval request sent by the target terminal, read the target entity data from the preset legal knowledge graph, and the target entity data includes the preset evidence items and the target weights of the preset evidence items;
  • the server When receiving the retrieval request sent by the target terminal, the server reads the target entity data from the preset legal knowledge graph.
  • the target entity data includes the preset evidence items and the target weights of the preset evidence items.
  • the server parses the search request sent by the target terminal to obtain the query text; the server extracts keywords from the query text according to the preset keyword extraction algorithm; the server queries the preset legal knowledge according to the preset similarity algorithm and keywords Atlas, the target entity data is obtained, and the target entity data includes preset evidence items and target weights corresponding to the preset evidence items.
  • the server parses the retrieval request sent by the target terminal, it also obtains the unique identifier of the target terminal, and the unique identifier is used to indicate the target terminal to which the server pushes the target entity data.
  • the server sorts the preset evidence items in descending order based on the target weights of the preset evidence items. It can be understood that the server sorts the target entity data in descending order according to the target weights of the preset evidence items in descending order Sort. For example, the server obtains the target entity data from the preset legal knowledge graph.
  • the target entity data includes 7 preset evidence items A, B, C, D, E, F, and G, and the corresponding target weights of the preset evidence items As 1, 1.1, 2.4, 0.5, 0.9, 0.3 and 3, the server sorts the preset evidence items in descending order based on the target weight of the preset evidence items, and the target entity data is (G, 3) , (C, 2.4), (B, 1.1), (A, 1), (E, 0.9), (D, 0.5) and (F, 0.3).
  • the server sets the preset evidence items whose target weight is greater than the preset threshold value as evidence guidance data. Specifically, the server reads the preset threshold; the server determines whether the target weight is greater than the preset threshold; if the target weight is greater than the preset threshold, the server sets the preset evidence items with the target weight greater than the preset threshold as evidence guidance data; if the target If the weight is less than or equal to the preset threshold, the server discards the preset evidence items whose target weight is less than or equal to the preset threshold.
  • the preset threshold is 2, and the preset evidence items are (G, 3), (C, 2.4), (B, 1.1), (A, 1), (E, 0.9), (D, 0.5) and ( F, 0.3), the server sets the preset evidence items (G, 3) and (C, 2.4) greater than 2 as evidence guidance data.
  • the server pushes the evidence guidance data to the target terminal, and the target terminal is used to display the evidence guidance data as prompt information. Specifically, the server confirms the target terminal according to the unique identifier of the target terminal; the server constructs the evidence guidance data in a preset format, and calls the preset push interface to push the constructed evidence guidance data to the target terminal, and the target terminal is used to send the evidence The guidance data is displayed with prompt information. For example, if the server confirms that the unique identifier 001 corresponds to the target terminal A, the server pushes the evidence guidance data to the target terminal A.
  • the unique identifier of the target terminal is set by the target terminal. It can be a device identifier, which is a string of 32-bit numbers and lowercase letters, or a password token identifier. Make a limit.
  • the weight of the preset evidence item is calculated by extracting the target feature and the weight of the target feature of the preset evidence item, and calculating the weight of the preset evidence item according to the preset evidence item's target feature and the weight of the target feature.
  • the weight of the evidence is sorted and pushed to improve the accuracy and efficiency of evidence guidance.
  • the ordering and recommending method of legal evidence in the embodiment of this application is described above, and the ordering and recommending device of legal evidence in the embodiment of this application is described below. Please refer to FIG. 3, the ordering and recommending device of legal evidence in the embodiment of this application One embodiment includes:
  • the obtaining unit 301 is configured to obtain the target feature of the preset evidence item, the preset evidence item is a set of evidence items in the preset dispute focus, and the target feature includes the preset evidence item frequency and the preset counter dispute focus frequency;
  • the processing unit 302 is configured to process the target feature according to the preset tree model to obtain the weight of the target feature
  • the calculation unit 303 is configured to perform a weighted average calculation on the target feature and the weight of the target feature to obtain the target weight of the preset evidence item;
  • the storage unit 304 is configured to store the target weight of the preset evidence item in the preset legal knowledge graph
  • the pushing unit 305 when receiving a retrieval request sent by the target terminal, is used to determine evidence guidance data from the preset legal knowledge graph and push the evidence guidance data to the target terminal.
  • the evidence guidance data is used to indicate that the target weight is greater than the preset The preset evidence item of the threshold.
  • the weight of the preset evidence item is calculated by extracting the target feature and the weight of the target feature of the preset evidence item, and calculating the weight of the preset evidence item according to the preset evidence item's target feature and the weight of the target feature.
  • the weight of the evidence is sorted and pushed to improve the accuracy and efficiency of evidence guidance.
  • another embodiment of the device for sorting and recommending legal evidence in the embodiment of the present application includes:
  • the obtaining unit 301 is configured to obtain the target feature of the preset evidence item, the preset evidence item is a set of evidence items in the preset dispute focus, and the target feature includes the preset evidence item frequency and the preset counter dispute focus frequency;
  • the processing unit 302 is configured to process the target feature according to the preset tree model to obtain the weight of the target feature
  • the calculation unit 303 is configured to perform a weighted average calculation on the target feature and the weight of the target feature to obtain the target weight of the preset evidence item;
  • the storage unit 304 is configured to store the target weight of the preset evidence item in the preset legal knowledge graph
  • the pushing unit 305 when receiving a retrieval request sent by the target terminal, is used to determine evidence guidance data from the preset legal knowledge graph and push the evidence guidance data to the target terminal.
  • the evidence guidance data is used to indicate that the target weight is greater than the preset The preset evidence item of the threshold.
  • processing unit 302 may also be specifically configured to:
  • the structure risk minimization model of the preset tree model is generated, and the structure risk minimization model is Among them, J(f) is the complexity of the structural risk minimization model;
  • T is the number of child nodes of the classification regression tree
  • is the modulus of the child node vector of the classification regression tree
  • ⁇ is the difficulty of node segmentation
  • ⁇ represents the L2 regularization coefficient
  • the results of the target classification regression tree are accumulated to obtain the weight ⁇ n of the target feature, and the value range of ⁇ n is greater than or equal to 0 and less than or equal to 1;
  • the weight ⁇ n of the target feature is written into the preset data dictionary.
  • the pushing unit 305 may further include:
  • the reading sub-unit 3051 is used to read the target entity data from the preset legal knowledge graph when the retrieval request sent by the target terminal is received, the target entity data includes the preset evidence items and the target weights of the preset evidence items;
  • the sorting subunit 3052 is used to sort the preset evidence items in descending order based on the target weight of the preset evidence items;
  • a setting subunit 3053 is used to set preset evidence items with a target weight greater than a preset threshold as evidence guidance data
  • the push subunit 3054 is used to push the evidence guidance data to the target terminal, and the target terminal is used to display the evidence guidance data as prompt information.
  • the reading subunit 3051 may also be specifically used for:
  • the preset legal knowledge graph is queried to obtain the target entity data.
  • the target entity data includes the preset evidence items and the target weights corresponding to the preset evidence items.
  • the obtaining unit 301 may also be specifically configured to:
  • Target characteristics include preset frequency of evidence items and preset frequency of inverse dispute focus;
  • the preset evidence item ⁇ k and the target feature ⁇ in are stored in the preset data dictionary according to the corresponding relationship.
  • calculation unit 303 may also be specifically configured to:
  • the storage unit 304 may also be specifically used for:
  • the weight of the preset evidence item is calculated by extracting the target feature and the weight of the target feature of the preset evidence item, and calculating the weight of the preset evidence item according to the preset evidence item's target feature and the weight of the target feature.
  • the weight of the evidence is sorted and pushed to improve the accuracy and efficiency of evidence guidance.
  • FIG. 5 is a schematic structural diagram of a device for sorting and recommending legal evidence provided by an embodiment of the present application.
  • the device for sorting and recommending legal evidence 500 may have relatively large differences due to different configurations or performances, and may include one or more A processor (central processing units, CPU) 510 (for example, one or more processors) and a memory 520, one or more storage media 530 (for example, one or more storage devices with a large amount of data) storing application programs 533 or data 532.
  • the memory 520 and the storage medium 530 may be short-term storage or persistent storage.
  • the program stored in the storage medium 530 may include one or more modules (not shown in the figure), and each module may include a sequence of legal evidence and a series of instruction operations in the recommendation device 500. Further, the processor 510 may be configured to communicate with the storage medium 530, and execute a series of instruction operations in the storage medium 530 on the device 500 for sorting and recommending legal evidence.
  • the device for sorting and recommending legal evidence 500 may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input and output interfaces 560, and/or one or more operating systems 531, For example, Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc.
  • Windows Serve Windows Serve
  • Mac OS X Unix
  • Linux FreeBSD
  • FIG. 5 do not constitute a limitation on the sorting of legal evidence and the recommendation device, and may include more or less components than those shown in the figure, or a combination of certain components. Some components, or different component arrangements.
  • the present application also provides a device for sorting and recommending legal evidence, including: a memory and at least one processor, the memory stores instructions, the memory and the at least one processor are interconnected by wires; the at least one processor The processor invokes the instructions in the memory, so that the device for sorting and recommending legal evidence executes the steps in the above-mentioned method for sorting and recommending legal evidence.
  • the present application also provides a computer-readable storage medium.
  • the computer-readable storage medium may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium.
  • the computer-readable storage medium stores computer instructions, and when the computer instructions are executed on the computer, the computer executes the following steps:
  • the preset evidence item is a set of evidence items in a preset dispute focus
  • the target feature includes a preset evidence item frequency and a preset inverse dispute focus frequency
  • evidence guidance data is determined from the preset legal knowledge graph, and the evidence guidance data is pushed to the target terminal, and the evidence guidance data is used to indicate the target A preset evidence item with a weight greater than a preset threshold.
  • the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium.
  • the technical solution of the present application essentially or the part that contributes to the existing technology or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , Including several instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in the various embodiments of the present application.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (read-only memory, ROM), random access memory (random access memory, RAM), magnetic disks or optical disks and other media that can store program codes. .

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

一种法律证据的排序和推荐方法、装置、设备及存储介质。法律证据的排序和推荐方法包括：获取预置证据项的目标特征，预置证据项为预置争议焦点中的证据项集合，目标特征包括预置证据项频率和预置逆争议焦点频率（101）；根据预置树模型对目标特征进行处理，得到目标特征的权重（102）；对目标特征和目标特征的权重进行加权平均计算，得到预置证据项的目标权重（103）；将预置证据项的目标权重存储到预置法律知识图谱中（104）；当接收到目标终端发送的检索请求时，从预置法律知识图谱中确定证据指引数据，并将证据指引数据推送到目标终端，证据指引数据用于指示目标权重大于预置阈值的预置证据项（105）。上述方法通过确定证据项的目标特征和证据项的权重，根据证据项的权重进行证据项排序和推送，提高证据指引的正确率和效率。

Description

法律证据的排序和推荐方法、装置、设备及存储介质
本申请要求于2019年10月18日提交中国专利局、申请号为201910991648.2、发明名称为“法律证据的排序和推荐方法、装置、设备及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在申请中。
技术领域
本申请涉及智能推荐领域，尤其涉及一种法律证据的排序和推荐方法、装置、设备及存储介质。
背景技术
知识图谱技术日益成为人工智能的基础，它是机器理解自然语言和构建知识网络的重要方法。近年来，知识图谱在司法领域的运用悄然兴起，它帮助从业人员快速地在线检索相关的法务内容，从而提高法院审判工作质量和效率。
发明人意识到，由于一个案件往往有多个证据，而且各个证据相互间存在相互印证，存在相互矛盾，所以有必要通过庭审质证以及法官核实判断证据，结合全案证据，对所有证据之间存在的客观联系，以及各个证据证明力的大小进行判断，并就案件事实作出符合客观实际的结论。法律知识图谱的一个重要应用便是在确定诉讼双方的争议焦点后根据法律知识图谱查询或者推断出解决这一争议焦点所需要的证据，并且对容易忽视但是在以往案例中起到重要的作用的证据给出提示。
目前在法律领域法律从业人员只能依靠法条或者个人从业经验来判断需要哪些证据以及这些证据对案件的重要性程度，这种证据指引方式准确率低。
发明内容
本申请的主要目的在于解决了只能依靠法条或者个人从业经验来判断证据或者证据权重准确率低的技术问题。
为实现上述目的，本申请第一方面提供了一种法律证据的排序和推荐方法，包括：获取预置证据项的目标特征，所述预置证据项为预置争议焦点中的证据项集合，所述目标特征包括预置证据项频率和预置逆争议焦点频率；根据预置树模型对所述目标特征进行处理，得到所述目标特征的权重；对所述目标特征和所述目标特征的权重进行加权平均计算，得到所述预置证据项的目标权重；将所述预置证据项的目标权重存储到预置法律知识图谱中；当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中确定证据指引数据，并将所述证据指引数据推送到所述目标终端，所述证据指引数据用于指示所述目标权重大于预置阈值的预置证据项。
本申请第二方面提供了一种法律证据的排序和推荐设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机可读指令，所述处理器执行所述计算机可读指令时实现如下步骤：获取预置证据项的目标特征，所述预置证据项为预置争议焦点中的证据项集合，所述目标特征包括预置证据项频率和预置逆争议焦点频率；根据预置树模型对所述目标特征进行处理，得到所述目标特征的权重；对所述目标特征和所述目标特征的权重进行加权平均计算，得到所述预置证据项的目标权重；将所述预置证据项的目标权重存储到预置法律知识图谱中；当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中确定证据指引数据，并将所述证据指引数据推送到所述目标终端，所述证据指引数据用于指示所述目标权重大于预置阈值的预置证据项。
本申请第三方面提供了一种计算机可读存储介质，所述计算机可读存储介质中存储有计算机指令，当所述计算机指令在计算机上运行时，使得计算机执行如下步骤：获取预置证据项的目标特征，所述预置证据项为预置争议焦点中的证据项集合，所述目标特征包括预置证据项频率和预置逆争议焦点频率；根据预置树模型对所述目标特征进行处理，得到 所述目标特征的权重；对所述目标特征和所述目标特征的权重进行加权平均计算，得到所述预置证据项的目标权重；将所述预置证据项的目标权重存储到预置法律知识图谱中；当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中确定证据指引数据，并将所述证据指引数据推送到所述目标终端，所述证据指引数据用于指示所述目标权重大于预置阈值的预置证据项。
本申请第四方面提供了一种法律证据的排序和推荐装置，包括：获取单元，用于获取预置证据项的目标特征，所述预置证据项为预置争议焦点中的证据项集合，所述目标特征包括预置证据项频率和预置逆争议焦点频率；处理单元，用于根据预置树模型对所述目标特征进行处理，得到所述目标特征的权重；计算单元，用于对所述目标特征和所述目标特征的权重进行加权平均计算，得到所述预置证据项的目标权重；存储单元，用于将所述预置证据项的目标权重存储到预置法律知识图谱中；推送单元，当接收到目标终端发送的检索请求时，用于从所述预置法律知识图谱中确定证据指引数据，并将所述证据指引数据推送到所述目标终端，所述证据指引数据用于指示所述目标权重大于预置阈值的预置证据项。
本申请提供的技术方案中，获取预置证据项的目标特征，所述预置证据项为预置争议焦点中的证据项集合，所述目标特征包括预置证据项频率和预置逆争议焦点频率；根据预置树模型对所述目标特征进行处理，得到所述目标特征的权重；对所述目标特征和所述目标特征的权重进行加权平均计算，得到所述预置证据项的目标权重；将所述预置证据项的目标权重存储到预置法律知识图谱中；当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中确定证据指引数据，并将所述证据指引数据推送到所述目标终端，所述证据指引数据用于指示所述目标权重大于预置阈值的预置证据项。本申请中，通过提取预置证据项的目标特征和目标特征的权重，并根据预置证据项的目标特征和目标特征的权重计算的到预置证据项的权重，根据预置证据项的权重进行证据项排序和推送，提高证据指引的正确率和效率。
附图说明
图1为本申请实施例中法律证据的排序和推荐方法的一个实施例示意图；
图2为本申请实施例中法律证据的排序和推荐方法的另一个实施例示意图；
图3为本申请实施例中法律证据的排序和推荐装置的一个实施例示意图；
图4为本申请实施例中法律证据的排序和推荐装置的另一个实施例示意图；
图5为本申请实施例中法律证据的排序和推荐设备的一个实施例示意图。
具体实施方式
本申请实施例提供了一种法律证据的排序和推荐方法、装置、设备及存储介质，用于通过提取预置证据项的目标特征和目标特征的权重，并根据预置证据项的目标特征和目标特征的权重计算的到预置证据项的权重，根据预置证据项的权重进行证据项排序和推送，提高证据指引的正确率和效率。
本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”、“第三”、“第四”等(如果存在)是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的实施例能够以除了在这里图示或描述的内容以外的顺序实施。此外，术语“包括”或“具有”及其任何变形，意图在于覆盖不排他的包含，例如，包含了一系列步骤或单元的过程、方法、系统、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。
为便于理解，下面对本申请实施例的具体流程进行描述，请参阅图1，本申请实施例中法律证据的排序和推荐方法的一个实施例包括：
101、获取预置证据项的目标特征，预置证据项为预置争议焦点中的证据项集合，目标特征包括预置证据项频率和预置逆争议焦点频率；
服务器获取预置证据项的目标特征，预置证据项为预置争议焦点中的证据项集合，目标特征包括预置证据项频率和预置逆争议焦点频率。其中，预置证据项频率为同一预置争议焦点对应的案例中预置证据项出现次数与同一预置争议焦点对应案例总数的比例系数，预置逆争议焦点频率为预置证据项在全部预置争议焦点中出现的次数与预置争议焦点总数的商，例如，对7个预置争议焦点，预置证据项存在3个预置争议焦点对应的案例中，则预置逆争议焦点频率就是3/7。
需要说明的是，预置证据项对应的目标特征包括预置证据项频率和预置逆争议焦点频率，构建这两个特征的目的是预置证据项的重要性随着预置证据项在案例中出现的次数成正比增加，但同时随着预置证据项在所有争议焦点中出现频率成反比下降。进一步地，目标特征还包括同一争议焦点对应的案例中该证据项支持小要素成立的次数/同一争议焦点对应的案例总数、是否是原始证据、是否是传来证据、是否是直接证据、是否是间接证据、是否是书证、是否是物证、是否是视听资料、是否是证人证言、是否是当事人陈述、是否是鉴定结论和是否是勘验笔录，具体此处不做限定。
102、根据预置树模型对目标特征进行处理，得到目标特征的权重；
服务器根据预置树模型对目标特征进行处理，得到目标特征的权重。具体的，服务器采用预置树模型对目标特征进行处理，得到目标特征的权重，的取值范围为大于等于0并且小于等于1。其中，预置树模型可以为极端梯度提升(extreme gradient boosting，XGBoost)算法，XGBoost算法通过构造加法模型逐步累加预测结果，以达到在给定目标特征的情况下，得出最佳的预测值，该最佳的预测值为目标特征的权重。
需要说明的是，权重是指某一因素或指标相对于某一事物的重要程度，其不同于一般的比重，体现的不仅仅是某一因素或指标所占的百分比，强调的是因素或指标的相对重要程度，倾向于贡献度或重要性。
103、对目标特征和目标特征的权重进行加权平均计算，得到预置证据项的目标权重；
服务器对目标特征和目标特征的权重进行加权平均计算，得到预置证据项的目标权重。具体的，服务器从预置数据字典中读取目标特征和目标特征的权重；服务器根据预置公式对目标特征和目标特征的权重进行加权平均计算，得到预置证据项的目标权重；服务器将目标特征设置为键，将目标权重设置为值，并将目标权重存储到预置数据字典中。
需要说明的是，预置数据字典是指预先对数据的数据项、数据结构、数据流、数据存储、处理逻辑等进行定义和描述，其目的是对数据流程图中的各个元素做出详细的说明，预置数据字典包括预置证据项和目标特征的信息集合。
104、将预置证据项的目标权重存储到预置法律知识图谱中；
服务器将预置证据项的目标权重存储到预置法律知识图谱中。具体的，服务器将预置证据项的目标权重设置为预置证据项的属性；服务器将预置证据项和预置证据项的目标权重导入到预置法律知识图谱中，预置法律知识图谱以三元组形式进行存储，其中，三元组是由实体、与实体关联的属性以及属性对应的属性值组成。
需要说明的是，知识图谱是指用图作为表示知识的数据结构，知识图谱包括节点和节点之间的边，其中，节点表示实体或者实体的属性值，节点之间的边用于表示实体的属性，预置法律知识图谱将法律知识用知识图谱结构化的方式表示。
105、当接收到目标终端发送的检索请求时，从预置法律知识图谱中确定证据指引数据，并将证据指引数据推送到目标终端，证据指引数据用于指示目标权重大于预置阈值的预置证据项。
当接收到目标终端发送的检索请求时，服务器从预置法律知识图谱中确定证据指引数据，并将证据指引数据推送到目标终端，证据指引数据用于指示目标权重大于预置阈值的预置证据项。具体的，服务器接收目标终端发送的检索请求；服务器解析目标终端发送的检索请求，得到目标终端的唯一标识和查询文本，其中，唯一标识用于指示目标终端；服务器根据查询文本查找预置法律知识图谱，得到预置证据项和预置证据项的权重；基于预置证据项的权重进行排序，并将目标权重大于预置阈值的预置证据项设置为证据指引数据，该证据指引数据于指示位根据预置法律知识图谱准确查找到对应查询文本的推送信息；服务器根据唯一标识将证据指引数据推送到目标终端。
本申请实施例中，通过提取预置证据项的目标特征和目标特征的权重，并根据预置证据项的目标特征和目标特征的权重计算的到预置证据项的权重，根据预置证据项的权重进行证据项排序和推送，提高证据指引的正确率和效率。
请参阅图2，本申请实施例中法律证据的排序和推荐方法的另一个实施例包括：
201、获取预置证据项的目标特征，预置证据项为预置争议焦点中的证据项集合，目标特征包括预置证据项频率和预置逆争议焦点频率；
服务器获取预置证据项的目标特征，预置证据项为预置争议焦点中的证据项集合，目标特征包括预置证据项频率和预置逆争议焦点频率。具体的，服务器根据预置法条确定预置争议焦点中的证据项集合，得到预置证据项χ k，其中，k为正整数，k用于指示预置争议焦点中预置证据项的数量，例如，目标案件争议焦点为借款是否为夫妻的共同财产，服务器确定目标案件争议焦点包括3个预置证据项，预置证据项为借款合同、离婚协议书和借款使用说明，则k＝3，将预置证据项标记为(χ 1,χ 2,χ 3)；服务器根据预置证据项确定目标特征，并将目标特征标记为α in，i为小于或者等于k的正整数，α in的取值范围为大于等于0并且小于等于1，n为目标特征的个数，目标特征包括预置证据项频率和预置逆争议焦点频率，其中，i用于指示预置证据项的序号，k在预置争议焦点中是不同的，但预置争议焦点中每个预置证据项对应的n是相同的，例如，3个预置证据项(χ 1,χ 2,χ 3)，若每个预置证据项包括2个目标特征，则n＝2，服务器设置预置证据项χ 1的目标特征为(α 11,α 12)，服务器设置预置证据项χ 2的目标特征为(α 21,α 22)，服务器设置预置证据项χ 3的目标特征为(α 31,α 32)，其中，0≤α in≤1；服务器将预置证据项χ k和目标特征α in按照对应关系存储到预置数据字典中。
需要说明的是，预置证据项对应的目标特征α in包括预置证据项频率和预置逆争议焦点频率，构建这两个特征的目的是预置证据项的重要性随着预置证据项在案例中出现的次数成正比增加，但同时随着预置证据项在所有争议焦点中出现频率成反比下降。进一步地，目标特征还包括同一争议焦点对应的案例中该证据项支持小要素成立的次数/同一争议焦点对应的案例总数、是否是原始证据、是否是传来证据、是否是直接证据、是否是间接证据、是否是书证、是否是物证、是否是视听资料、是否是证人证言、是否是当事人陈述、是否是鉴定结论和是否是勘验笔录，具体此处不做限定。例如，服务器按照上述目标特征顺序设置的(α i1,α i2,...,α in)为(0.3,0.4,0.8,1,0,1,0,1,0,0,0,0,0,0)。
202、根据预置树模型对目标特征进行处理，得到目标特征的权重；
服务器根据预置树模型对目标特征进行处理，得到目标特征的权重。具体的，服务器根据预置最小化目标特征的损失函数L生成预置树模型的结构风险最小化模型，结构风险最小化模型为
Figure PCTCN2020118287-appb-000001
其中，J(f)为结构风险最小化模型的复 杂度；服务器确定每棵分类回归树的复杂度Ω(f)，Ω(f)为
Figure PCTCN2020118287-appb-000002
其中，T为分类回归树的子节点数，||ω||为分类回归树的子节点向量的模，γ为节点切分的难度，λ表示L2正则化系数；服务器对预置最小化目标特征的损失函数L和每棵分类回归树的复杂度Ω(f)进行叠加，得到初始函数，初始函数如下所示：
Figure PCTCN2020118287-appb-000003
服务器根据初始函数进行计算得到目标函数obj *；服务器根据目标函数obj *确定目标特征的全部切分点，并对目标特征的全部切分点进行切分，得到目标特征的多个目标增益值；服务器从目标特征的多个目标增益值中确定最大增益值；服务器设置最大增益值对应的切分点为最佳分裂点，并在最佳分裂点处切分节点，得到目标分类回归树；服务器累加目标分类回归树的结果，得到目标特征的权重β n，β n的取值范围为大于等于0并且小于等于1；服务器将目标特征的权重β n写入到预置数据字典中。其中，β n相加后总和为1，例如，n＝4，则目标特征为(0.1，0.4，0.3，0.2)。
需要说明的是，权重是指某一因素或指标相对于某一事物的重要程度，其不同于一般的比重，体现的不仅仅是某一因素或指标所占的百分比，强调的是因素或指标的相对重要程度，倾向于贡献度或重要性。
203、对目标特征和目标特征的权重进行加权平均计算，得到预置证据项的目标权重；
服务器对目标特征和目标特征的权重进行加权平均计算，得到预置证据项的目标权重，具体的，服务器从预置数据字典中读取目标特征α in和目标特征的权重β n；服务器根据预置公式对目标特征α in和目标特征的权重β n进行加权平均计算，得到预置证据项的目标权重ρ i，预置公式为ρ i＝β 1×α i1+β 2×α i2+...+β n×α in，其中，ρ i＞0；服务器将目标特征设置为键，将目标权重ρ i设置为值，并将目标权重ρ i存储到预置数据字典中。
需要说明的是，预置证据项的目标权重与该预置证据项在案例中出现的次数成正比，预置证据项的目标权重与预置证据项在所有预置争议焦点中出现的频率成反比。
204、将预置证据项的目标权重存储到预置法律知识图谱中；
服务器将预置证据项的目标权重存储到预置法律知识图谱中。服务器将预置证据项的目标权重设置为预置证据项的属性；服务器将预置证据项和预置证据项的目标权重导入到预置法律知识图谱中，预置法律知识图谱以三元组形式进行存储，其中，三元组是由实体、与实体关联的属性以及属性对应的属性值组成。
需要说明的是，知识图谱是指用图作为表示知识的数据结构，知识图谱包括节点和节点之间的边，其中，节点表示实体或者实体的属性值，节点之间的边用于表示实体的属性，预置法律知识图谱将法律知识用知识图谱结构化的方式表示。
205、当接收到目标终端发送的检索请求时，从预置法律知识图谱中读取目标实体数据，目标实体数据包括预置证据项和预置证据项的目标权重；
当接收到目标终端发送的检索请求时，服务器从预置法律知识图谱中读取目标实体数据，目标实体数据包括预置证据项和预置证据项的目标权重，具体的，当接收到目标终端发送的检索请求时，服务器解析目标终端发送的检索请求，得到查询文本；服务器根据预置关键词提取算法从查询文本中提取关键词；服务器根据预置相似度算法和关键词查询预置法律知识图谱，得到目标实体数据，目标实体数据包括预置证据项和预置证据项对应的目标权重。
需要说明的是，服务器解析目标终端发送的检索请求后，还得到目标终端的唯一标识， 唯一标识用于指示服务器将目标实体数据推送的目标终端。
206、基于预置证据项的目标权重将预置证据项按照从大到小的顺序进行排序；
服务器基于预置证据项的目标权重将预置证据项按照从大到小的顺序进行排序，可以理解的是，服务器按照预置证据项的目标权重从大到小的顺序对目标实体数据进行降序排序。例如，服务器从预置法律知识图谱中查询得到目标实体数据，目标实体数据包括A、B、C、D、E、F和G共7项预置证据项，预置证据项各自对应的目标权重为1、1.1、2.4、0.5、0.9、0.3和3，服务器基于预置证据项的目标权重将预置证据项按照从大到小的顺序进行排序后，得到目标实体数据为(G，3)、(C，2.4)、(B，1.1)、(A，1)、(E，0.9)、(D，0.5)和(F，0.3)。
207、将目标权重大于预置阈值的预置证据项设置为证据指引数据；
服务器将目标权重大于预置阈值的预置证据项设置为证据指引数据。具体的，服务器读取预置阈值；服务器判断目标权重是否大于预置阈值；若目标权重大于预置阈值，则服务器将目标权重大于预置阈值的预置证据项设置为证据指引数据；若目标权重小于或者等于预置阈值，则服务器将目标权重小于或者等于预置阈值的预置证据项丢弃。例如，预置阈值为2，预置证据项为(G，3)、(C，2.4)、(B，1.1)、(A，1)、(E，0.9)、(D，0.5)和(F，0.3)，则服务器将大于2的预置证据项(G，3)、(C，2.4)设置为证据指引数据。
208、将证据指引数据推送到目标终端，目标终端用于将证据指引数据以提示信息进行展示。
服务器将证据指引数据推送到目标终端，目标终端用于将证据指引数据以提示信息进行展示。具体的，服务器根据目标终端的唯一标识确认目标终端；服务器将证据指引数据按照预置格式进行构建，并调用预置推送接口将构建后的证据指引数据推送到目标终端，目标终端用于将证据指引数据以提示信息进行展示。例如，服务器确认唯一标识001对应目标终端A，则服务器将证据指引数据推送到目标终端A。
需要说明的是，目标终端的唯一标识由目标终端进行设置，可以为设备标识，也就是长度为32位的数字和小写字母的组合的字符串，也可以为密码令牌标识，具体此处不做限定。
本申请实施例中，通过提取预置证据项的目标特征和目标特征的权重，并根据预置证据项的目标特征和目标特征的权重计算的到预置证据项的权重，根据预置证据项的权重进行证据项排序和推送，提高证据指引的正确率和效率。
上面对本申请实施例中法律证据的排序和推荐方法进行了描述，下面对本申请实施例中法律证据的排序和推荐装置进行描述，请参阅图3，本申请实施例中法律证据的排序和推荐装置一个实施例包括：
获取单元301，用于获取预置证据项的目标特征，预置证据项为预置争议焦点中的证据项集合，目标特征包括预置证据项频率和预置逆争议焦点频率；
处理单元302，用于根据预置树模型对目标特征进行处理，得到目标特征的权重；
计算单元303，用于对目标特征和目标特征的权重进行加权平均计算，得到预置证据项的目标权重；
存储单元304，用于将预置证据项的目标权重存储到预置法律知识图谱中；
推送单元305，当接收到目标终端发送的检索请求时，用于从预置法律知识图谱中确定证据指引数据，并将证据指引数据推送到目标终端，证据指引数据用于指示目标权重大于预置阈值的预置证据项。
本申请实施例中，通过提取预置证据项的目标特征和目标特征的权重，并根据预置证据项的目标特征和目标特征的权重计算的到预置证据项的权重，根据预置证据项的权重进 行证据项排序和推送，提高证据指引的正确率和效率。
请参阅图4，本申请实施例中法律证据的排序和推荐装置的另一个实施例包括：
获取单元301，用于获取预置证据项的目标特征，预置证据项为预置争议焦点中的证据项集合，目标特征包括预置证据项频率和预置逆争议焦点频率；
处理单元302，用于根据预置树模型对目标特征进行处理，得到目标特征的权重；
计算单元303，用于对目标特征和目标特征的权重进行加权平均计算，得到预置证据项的目标权重；
存储单元304，用于将预置证据项的目标权重存储到预置法律知识图谱中；
推送单元305，当接收到目标终端发送的检索请求时，用于从预置法律知识图谱中确定证据指引数据，并将证据指引数据推送到目标终端，证据指引数据用于指示目标权重大于预置阈值的预置证据项。
可选的，处理单元302还可以具体用于：
根据预置最小化目标特征的损失函数L生成预置树模型的结构风险最小化模型，结构风险最小化模型为
Figure PCTCN2020118287-appb-000004
其中，J(f)为结构风险最小化模型的复杂度；
确定每棵分类回归树的复杂度Ω(f)，
Figure PCTCN2020118287-appb-000005
其中，T为分类回归树的子节点数，||ω||为分类回归树的子节点向量的模，γ为节点切分的难度，λ表示L2正则化系数；
对预置最小化目标特征的损失函数L和每棵分类回归树的复杂度Ω(f)进行叠加，得到初始函数，初始函数如下所示：
Figure PCTCN2020118287-appb-000006
根据初始函数进行计算得到目标函数obj *；
根据目标函数obj *确定目标特征的全部切分点，并对目标特征的全部切分点进行切分，得到目标特征的多个目标增益值；
从目标特征的多个目标增益值中确定最大增益值；
设置最大增益值对应的切分点为最佳分裂点，并在最佳分裂点处切分节点，得到目标分类回归树；
累加目标分类回归树的结果，得到目标特征的权重β n，β n的取值范围为大于等于0并且小于等于1；
将目标特征的权重β n写入到预置数据字典中。
可选的，推送单元305还可进一步包括：
读取子单元3051，当接收到目标终端发送的检索请求时，用于从预置法律知识图谱中读取目标实体数据，目标实体数据包括预置证据项和预置证据项的目标权重；
排序子单元3052，用于基于预置证据项的目标权重将预置证据项按照从大到小的顺序进行排序；
设置子单元3053，用于将目标权重大于预置阈值的预置证据项设置为证据指引数据；
推送子单元3054，用于将证据指引数据推送到目标终端，目标终端用于将证据指引数据以提示信息进行展示。
可选的，读取子单元3051还可以具体用于：
当接收到目标终端发送的检索请求时，解析目标终端发送的检索请求，得到查询文本；
根据预置关键词提取算法从查询文本中提取关键词；
根据预置相似度算法和关键词查询预置法律知识图谱，得到目标实体数据，目标实体数据包括预置证据项和预置证据项对应的目标权重。
可选的，获取单元301还可以具体用于：
根据预置法条确定预置争议焦点中的证据项集合，得到预置证据项χ k，其中，k为正整数，k用于指示预置争议焦点中预置证据项的数量；
根据预置证据项确定目标特征，并将目标特征标记为α in，i为小于或者等于k的正整数，α in的取值范围为大于等于0并且小于等于1，n为目标特征的个数，目标特征包括预置证据项频率和预置逆争议焦点频率；
将预置证据项χ k和目标特征α in按照对应关系存储到预置数据字典中。
可选的，计算单元303还可以具体用于：
从预置数据字典中读取目标特征α in和目标特征的权重β n；
根据预置公式对目标特征α in和目标特征的权重β n进行加权平均计算，得到预置证据项的目标权重ρ i，预置公式为ρ i＝β 1×α i1+β 2×α i2+...+β n×α in，其中，ρ i＞0；
将目标特征设置为键，将目标权重ρ i设置为值，并将目标权重ρ i存储到预置数据字典中。
可选的，存储单元304还可以具体用于：
将预置证据项的目标权重设置为预置证据项的属性；
将预置证据项和预置证据项的目标权重导入到预置法律知识图谱中，预置法律知识图谱以三元组形式进行存储。
本申请实施例中，通过提取预置证据项的目标特征和目标特征的权重，并根据预置证据项的目标特征和目标特征的权重计算的到预置证据项的权重，根据预置证据项的权重进行证据项排序和推送，提高证据指引的正确率和效率。
上面图3和图4从模块化功能实体的角度对本申请实施例中的法律证据的排序和推荐装置进行详细描述，下面从硬件处理的角度对本申请实施例中法律证据的排序和推荐设备进行详细描述。
图5是本申请实施例提供的一种法律证据的排序和推荐设备的结构示意图，该法律证据的排序和推荐设备500可因配置或性能不同而产生比较大的差异，可以包括一个或一个以上处理器(central processing units，CPU)510(例如，一个或一个以上处理器)和存储器520，一个或一个以上存储应用程序533或数据532的存储介质530(例如一个或一个以上海量存储设备)。其中，存储器520和存储介质530可以是短暂存储或持久存储。存储在存储介质530的程序可以包括一个或一个以上模块(图示没标出)，每个模块可以包括对法律证据的排序和推荐设备500中的一系列指令操作。更进一步地，处理器510可以设置为与存储介质530通信，在法律证据的排序和推荐设备500上执行存储介质530中的一系列指令操作。
法律证据的排序和推荐设备500还可以包括一个或一个以上电源540，一个或一个以上有线或无线网络接口550，一个或一个以上输入输出接口560，和/或，一个或一个以上 操作系统531，例如Windows Serve，Mac OS X，Unix，Linux，FreeBSD等等。本领域技术人员可以理解，图5示出的法律证据的排序和推荐设备结构并不构成对法律证据的排序和推荐设备的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。
本申请还提供一种法律证据的排序和推荐设备，包括：存储器和至少一个处理器，所述存储器中存储有指令，所述存储器和所述至少一个处理器通过线路互连；所述至少一个处理器调用所述存储器中的所述指令，以使得所述法律证据的排序和推荐设备执行上述法律证据的排序和推荐方法中的步骤。
本申请还提供一种计算机可读存储介质，该计算机可读存储介质可以为非易失性计算机可读存储介质，也可以为易失性计算机可读存储介质。计算机可读存储介质存储有计算机指令，当所述计算机指令在计算机上运行时，使得计算机执行如下步骤：
获取预置证据项的目标特征，所述预置证据项为预置争议焦点中的证据项集合，所述目标特征包括预置证据项频率和预置逆争议焦点频率；
根据预置树模型对所述目标特征进行处理，得到所述目标特征的权重；
对所述目标特征和所述目标特征的权重进行加权平均计算，得到所述预置证据项的目标权重；
将所述预置证据项的目标权重存储到预置法律知识图谱中；
当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中确定证据指引数据，并将所述证据指引数据推送到所述目标终端，所述证据指引数据用于指示所述目标权重大于预置阈值的预置证据项。
所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(read-only memory，ROM)、随机存取存储器(random access memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述，以上实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围。

Claims (20)

  1. 一种法律证据的排序和推荐方法，其中，包括：
    获取预置证据项的目标特征，所述预置证据项为预置争议焦点中的证据项集合，所述目标特征包括预置证据项频率和预置逆争议焦点频率；
    根据预置树模型对所述目标特征进行处理，得到所述目标特征的权重；
    对所述目标特征和所述目标特征的权重进行加权平均计算，得到所述预置证据项的目标权重；
    将所述预置证据项的目标权重存储到预置法律知识图谱中；
    当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中确定证据指引数据，并将所述证据指引数据推送到所述目标终端，所述证据指引数据用于指示所述目标权重大于预置阈值的预置证据项。
  2. 根据权利要求1所述的法律证据的排序和推荐方法，其中，所述根据预置树模型对所述目标特征进行处理，得到所述目标特征的权重，包括：
    根据预置最小化目标特征的损失函数L生成预置树模型的结构风险最小化模型，所述结构风险最小化模型为
    Figure PCTCN2020118287-appb-100001
    其中，J(f)为结构风险最小化模型的复杂度；
    确定每棵分类回归树的复杂度Ω(f)，所述Ω(f)为
    Figure PCTCN2020118287-appb-100002
    其中，T为所述分类回归树的子节点数，||ω||为所述分类回归树的子节点向量的模，γ为节点切分的难度，λ表示L2正则化系数；
    对所述预置最小化目标特征的损失函数L和所述每棵分类回归树的复杂度Ω(f)进行叠加，得到初始函数，所述初始函数如下所示：
    Figure PCTCN2020118287-appb-100003
    根据所述初始函数进行计算得到目标函数obj *；
    根据所述目标函数obj *确定所述目标特征的全部切分点，并对所述目标特征的全部切分点进行切分，得到所述目标特征的多个目标增益值；
    从所述目标特征的多个目标增益值中确定最大增益值；
    设置所述最大增益值对应的切分点为最佳分裂点，并在最佳分裂点处切分节点，得到目标分类回归树；
    累加目标分类回归树的结果，得到目标特征的权重β n，所述β n的取值范围为大于等于0并且小于等于1；
    将所述目标特征的权重β n写入到预置数据字典中。
  3. 根据权利要求1所述的法律证据的排序和推荐方法，其中，所述当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中确定证据指引数据，并将所述证据指引数据推送到所述目标终端，所述证据指引数据用于指示所述目标权重大于预置阈值的预置证据项，包括：
    当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中读取目标实体数据，所述目标实体数据包括所述预置证据项和所述预置证据项的目标权重；
    基于所述预置证据项的目标权重将所述预置证据项按照从大到小的顺序进行排序；
    将所述目标权重大于预置阈值的预置证据项设置为证据指引数据；
    将所述证据指引数据推送到目标终端，所述目标终端用于将所述证据指引数据以提示信息进行展示。
  4. 根据权利要求3所述的法律证据的排序和推荐方法，其中，所述当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中读取目标实体数据，所述目标实体数据包括所述预置证据项和所述预置证据项的目标权重，包括：
    当接收到目标终端发送的检索请求时，解析所述目标终端发送的检索请求，得到查询文本；
    根据预置关键词提取算法从所述查询文本中提取关键词；
    根据预置相似度算法和所述关键词查询所述预置法律知识图谱，得到目标实体数据，所述目标实体数据包括所述预置证据项和所述预置证据项对应的目标权重。
  5. 根据权利要求1所述的法律证据的排序和推荐方法，其中，所述获取预置证据项的目标特征，所述预置证据项为预置争议焦点中的证据项集合，所述目标特征包括预置证据项频率和预置逆争议焦点频率，包括：
    根据预置法条确定预置争议焦点中的证据项集合，得到预置证据项χ k，其中，所述k为正整数，所述k用于指示预置争议焦点中预置证据项的数量；
    根据所述预置证据项χ k确定目标特征，并将目标特征标记为α in，所述i为小于或者等于k的正整数，所述α in的取值范围为大于等于0并且小于等于1，所述n为目标特征的个数，所述目标特征包括预置证据项频率和预置逆争议焦点频率；
    将所述预置证据项χ k和所述目标特征α in按照对应关系存储到预置数据字典中。
  6. 根据权利要求5所述的法律证据的排序和推荐方法，其中，所述对所述目标特征和所述目标特征的权重进行加权平均计算，得到所述预置证据项的目标权重，包括：
    从所述预置数据字典中读取所述目标特征α in和所述目标特征的权重β n；
    根据预置公式对所述目标特征α in和所述目标特征的权重β n进行加权平均计算，得到所述预置证据项的目标权重ρ i，所述预置公式为ρ i＝β 1×α i1+β 2×α i2+...+β n×α in，其中，ρ i＞0；
    将所述目标特征设置为键，将所述目标权重ρ i设置为值，并将所述目标权重ρ i存储到所述预置数据字典中。
  7. 根据权利要求1-6中任意一项所述的法律证据的排序和推荐方法，其中，所述将所述预置证据项的目标权重存储到预置法律知识图谱中，包括：
    将所述预置证据项的目标权重设置为所述预置证据项的属性；
    将所述预置证据项和所述预置证据项的目标权重导入到预置法律知识图谱中，所述预置法律知识图谱以三元组形式进行存储。
  8. 一种法律证据的排序和推荐设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机可读指令，所述处理器执行所述计算机可读指令时实现如下步骤：
    获取预置证据项的目标特征，所述预置证据项为预置争议焦点中的证据项集合，所述目标特征包括预置证据项频率和预置逆争议焦点频率；
    根据预置树模型对所述目标特征进行处理，得到所述目标特征的权重；
    对所述目标特征和所述目标特征的权重进行加权平均计算，得到所述预置证据项的目标权重；
    将所述预置证据项的目标权重存储到预置法律知识图谱中；
    当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中确定证据指引数据，并将所述证据指引数据推送到所述目标终端，所述证据指引数据用于指示所述目标权重大于预置阈值的预置证据项。
  9. 根据权利要求8所述的法律证据的排序和推荐设备，所述处理器执行所述计算机程序时还实现以下步骤：
    根据预置最小化目标特征的损失函数L生成预置树模型的结构风险最小化模型，所述结构风险最小化模型为
    Figure PCTCN2020118287-appb-100004
    其中，J(f)为结构风险最小化模型的复杂度；
    确定每棵分类回归树的复杂度Ω(f)，所述Ω(f)为
    Figure PCTCN2020118287-appb-100005
    其中，T为所述分类回归树的子节点数，||ω||为所述分类回归树的子节点向量的模，γ为节点切分的难度，λ表示L2正则化系数；
    对所述预置最小化目标特征的损失函数L和所述每棵分类回归树的复杂度Ω(f)进行叠加，得到初始函数，所述初始函数如下所示：
    Figure PCTCN2020118287-appb-100006
    根据所述初始函数进行计算得到目标函数obj *；
    根据所述目标函数obj *确定所述目标特征的全部切分点，并对所述目标特征的全部切分点进行切分，得到所述目标特征的多个目标增益值；
    从所述目标特征的多个目标增益值中确定最大增益值；
    设置所述最大增益值对应的切分点为最佳分裂点，并在最佳分裂点处切分节点，得到目标分类回归树；
    累加目标分类回归树的结果，得到目标特征的权重β n，所述β n的取值范围为大于等于0并且小于等于1；
    将所述目标特征的权重β n写入到预置数据字典中。
  10. 根据权利要求8所述的法律证据的排序和推荐设备，所述处理器执行所述计算机程序时还实现以下步骤：
    当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中读取目标实体数据，所述目标实体数据包括所述预置证据项和所述预置证据项的目标权重；
    基于所述预置证据项的目标权重将所述预置证据项按照从大到小的顺序进行排序；
    将所述目标权重大于预置阈值的预置证据项设置为证据指引数据；
    将所述证据指引数据推送到目标终端，所述目标终端用于将所述证据指引数据以提示信息进行展示。
  11. 根据权利要求10所述的法律证据的排序和推荐设备，所述处理器执行所述计算机程序时还实现以下步骤：
    当接收到目标终端发送的检索请求时，解析所述目标终端发送的检索请求，得到查询 文本；
    根据预置关键词提取算法从所述查询文本中提取关键词；
    根据预置相似度算法和所述关键词查询所述预置法律知识图谱，得到目标实体数据，所述目标实体数据包括所述预置证据项和所述预置证据项对应的目标权重。
  12. 根据权利要求8所述的法律证据的排序和推荐设备，所述处理器执行所述计算机程序时还实现以下步骤：
    根据预置法条确定预置争议焦点中的证据项集合，得到预置证据项χ k，其中，所述k为正整数，所述k用于指示预置争议焦点中预置证据项的数量；
    根据所述预置证据项χ k确定目标特征，并将目标特征标记为α in，所述i为小于或者等于k的正整数，所述α in的取值范围为大于等于0并且小于等于1，所述n为目标特征的个数，所述目标特征包括预置证据项频率和预置逆争议焦点频率；
    将所述预置证据项χ k和所述目标特征α in按照对应关系存储到预置数据字典中。
  13. 根据权利要求12所述的法律证据的排序和推荐设备，所述处理器执行所述计算机程序时还实现以下步骤：
    从所述预置数据字典中读取所述目标特征α in和所述目标特征的权重β n；
    根据预置公式对所述目标特征α in和所述目标特征的权重β n进行加权平均计算，得到所述预置证据项的目标权重ρ i，所述预置公式为ρ i＝β 1×α i1+β 2×α i2+...+β n×α in，其中，ρ i＞0；
    将所述目标特征设置为键，将所述目标权重ρ i设置为值，并将所述目标权重ρ i存储到所述预置数据字典中。
  14. 根据权利要求8-13中任意一项所述的法律证据的排序和推荐设备，所述处理器执行所述计算机程序时还实现以下步骤：
    将所述预置证据项的目标权重设置为所述预置证据项的属性；
    将所述预置证据项和所述预置证据项的目标权重导入到预置法律知识图谱中，所述预置法律知识图谱以三元组形式进行存储。
  15. 一种计算机可读存储介质，所述计算机可读存储介质中存储计算机指令，当所述计算机指令在计算机上运行时，使得计算机执行如下步骤：
    获取预置证据项的目标特征，所述预置证据项为预置争议焦点中的证据项集合，所述目标特征包括预置证据项频率和预置逆争议焦点频率；
    根据预置树模型对所述目标特征进行处理，得到所述目标特征的权重；
    对所述目标特征和所述目标特征的权重进行加权平均计算，得到所述预置证据项的目标权重；
    将所述预置证据项的目标权重存储到预置法律知识图谱中；
    当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中确定证据指引数据，并将所述证据指引数据推送到所述目标终端，所述证据指引数据用于指示所述目标权重大于预置阈值的预置证据项。
  16. 根据权利要求15所述的计算机可读存储介质，当所述计算机指令在计算机上运行时，使得计算机还执行以下步骤：
    根据预置最小化目标特征的损失函数L生成预置树模型的结构风险最小化模型，所述结构风险最小化模型为
    Figure PCTCN2020118287-appb-100007
    其中，J(f)为结构风险最小化模型的复杂度；
    确定每棵分类回归树的复杂度Ω(f)，所述Ω(f)为
    Figure PCTCN2020118287-appb-100008
    其中，T为所述分类回归树的子节点数，||ω||为所述分类回归树的子节点向量的模，γ为节点切分的难度，λ表示L2正则化系数；
    对所述预置最小化目标特征的损失函数L和所述每棵分类回归树的复杂度Ω(f)进行叠加，得到初始函数，所述初始函数如下所示：
    Figure PCTCN2020118287-appb-100009
    根据所述初始函数进行计算得到目标函数obj *；
    根据所述目标函数obj *确定所述目标特征的全部切分点，并对所述目标特征的全部切分点进行切分，得到所述目标特征的多个目标增益值；
    从所述目标特征的多个目标增益值中确定最大增益值；
    设置所述最大增益值对应的切分点为最佳分裂点，并在最佳分裂点处切分节点，得到目标分类回归树；
    累加目标分类回归树的结果，得到目标特征的权重β n，所述β n的取值范围为大于等于0并且小于等于1；
    将所述目标特征的权重β n写入到预置数据字典中。
  17. 根据权利要求15所述的计算机可读存储介质，当所述计算机指令在计算机上运行时，使得计算机还执行以下步骤：
    当接收到目标终端发送的检索请求时，从所述预置法律知识图谱中读取目标实体数据，所述目标实体数据包括所述预置证据项和所述预置证据项的目标权重；
    基于所述预置证据项的目标权重将所述预置证据项按照从大到小的顺序进行排序；
    将所述目标权重大于预置阈值的预置证据项设置为证据指引数据；
    将所述证据指引数据推送到目标终端，所述目标终端用于将所述证据指引数据以提示信息进行展示。
  18. 根据权利要求17所述的计算机可读存储介质，当所述计算机指令在计算机上运行时，使得计算机还执行以下步骤：
    当接收到目标终端发送的检索请求时，解析所述目标终端发送的检索请求，得到查询文本；
    根据预置关键词提取算法从所述查询文本中提取关键词；
    根据预置相似度算法和所述关键词查询所述预置法律知识图谱，得到目标实体数据，所述目标实体数据包括所述预置证据项和所述预置证据项对应的目标权重。
  19. 根据权利要求15所述的计算机可读存储介质，当所述计算机指令在计算机上运行执行以下步骤时，使得计算机还执行以下步骤：
    根据预置法条确定预置争议焦点中的证据项集合，得到预置证据项χ k，其中，所述k为正整数，所述k用于指示预置争议焦点中预置证据项的数量；
    根据所述预置证据项χ k确定目标特征，并将目标特征标记为α in，所述i为小于或者等 于k的正整数，所述α in的取值范围为大于等于0并且小于等于1，所述n为目标特征的个数，所述目标特征包括预置证据项频率和预置逆争议焦点频率；
    将所述预置证据项χ k和所述目标特征α in按照对应关系存储到预置数据字典中。
  20. 一种法律证据的排序和推荐装置，其中，所述法律证据的排序和推荐装置包括：
    获取单元，用于获取预置证据项的目标特征，所述预置证据项为预置争议焦点中的证据项集合，所述目标特征包括预置证据项频率和预置逆争议焦点频率；
    处理单元，用于根据预置树模型对所述目标特征进行处理，得到所述目标特征的权重；
    计算单元，用于对所述目标特征和所述目标特征的权重进行加权平均计算，得到所述预置证据项的目标权重；
    存储单元，用于将所述预置证据项的目标权重存储到预置法律知识图谱中；
    推送单元，当接收到目标终端发送的检索请求时，用于从所述预置法律知识图谱中确定证据指引数据，并将所述证据指引数据推送到所述目标终端，所述证据指引数据用于指示所述目标权重大于预置阈值的预置证据项。
PCT/CN2020/118287 2019-10-18 2020-09-28 法律证据的排序和推荐方法、装置、设备及存储介质 WO2021073410A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910991648.2 2019-10-18
CN201910991648.2A CN110928986B (zh) 2019-10-18 2019-10-18 法律证据的排序和推荐方法、装置、设备及存储介质

Publications (1)

Publication Number Publication Date
WO2021073410A1 true WO2021073410A1 (zh) 2021-04-22

Family

ID=69849246

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/118287 WO2021073410A1 (zh) 2019-10-18 2020-09-28 法律证据的排序和推荐方法、装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN110928986B (zh)
WO (1) WO2021073410A1 (zh)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110928986B (zh) * 2019-10-18 2023-07-21 平安科技（深圳）有限公司 法律证据的排序和推荐方法、装置、设备及存储介质
CN112052304A (zh) * 2020-08-18 2020-12-08 中国建设银行股份有限公司 课程标签确定方法、装置及电子设备
CN114969384B (zh) * 2022-08-02 2022-10-21 联通(四川)产业互联网有限公司 高价值司法证据链获取存储方法、装置及可读储存介质
CN117743590A (zh) * 2023-11-30 2024-03-22 北京汉勃科技有限公司 一种基于大语言模型的法律辅助方法及系统

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107133276A (zh) * 2017-04-12 2017-09-05 安徽富驰信息技术有限公司 一种司法类案检索系统及其检索方法
US20170300470A1 (en) * 2016-04-15 2017-10-19 Marca Research & Development International, Llc Systems and methods for identifying evidentiary information
CN109359175A (zh) * 2018-09-07 2019-02-19 平安科技（深圳）有限公司 电子装置、诉讼数据处理的方法及存储介质
CN110175333A (zh) * 2019-06-04 2019-08-27 科大讯飞股份有限公司 一种证据指引方法、装置、设备及存储介质
CN110928986A (zh) * 2019-10-18 2020-03-27 平安科技（深圳）有限公司 法律证据的排序和推荐方法、装置、设备及存储介质

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108038508A (zh) * 2017-12-12 2018-05-15 携程旅游网络技术（上海）有限公司 中转航班的推送方法、系统、存储介质和电子设备
CN110110233B (zh) * 2019-05-09 2022-04-22 网易传媒科技（北京）有限公司 信息处理方法、装置、介质和计算设备
CN110263854B (zh) * 2019-06-20 2023-06-27 广州酷狗计算机科技有限公司 直播标签确定方法、装置及存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170300470A1 (en) * 2016-04-15 2017-10-19 Marca Research & Development International, Llc Systems and methods for identifying evidentiary information
CN107133276A (zh) * 2017-04-12 2017-09-05 安徽富驰信息技术有限公司 一种司法类案检索系统及其检索方法
CN109359175A (zh) * 2018-09-07 2019-02-19 平安科技（深圳）有限公司 电子装置、诉讼数据处理的方法及存储介质
CN110175333A (zh) * 2019-06-04 2019-08-27 科大讯飞股份有限公司 一种证据指引方法、装置、设备及存储介质
CN110928986A (zh) * 2019-10-18 2020-03-27 平安科技（深圳）有限公司 法律证据的排序和推荐方法、装置、设备及存储介质

Also Published As

Publication number Publication date
CN110928986B (zh) 2023-07-21
CN110928986A (zh) 2020-03-27

Similar Documents

Publication Publication Date Title
WO2021073410A1 (zh) 法律证据的排序和推荐方法、装置、设备及存储介质
US11017178B2 (en) Methods, devices, and systems for constructing intelligent knowledge base
US20180336193A1 (en) Artificial Intelligence Based Method and Apparatus for Generating Article
WO2017181612A1 (zh) 个性化视频推荐方法及装置
CN107180093B (zh) 信息搜索方法及装置和时效性查询词识别方法及装置
US8666984B2 (en) Unsupervised message clustering
CN110569496B (zh) 实体链接方法、装置及存储介质
US20180336202A1 (en) System and method to represent documents for search in a graph
WO2016180270A1 (zh) 网页分类方法和装置、计算设备以及机器可读存储介质
CN107918657B (zh) 一种数据源的匹配方法和装置
CN110019794B (zh) 文本资源的分类方法、装置、存储介质及电子装置
CN112100396B (zh) 一种数据处理方法和装置
WO2021082123A1 (zh) 信息推荐方法及装置、电子设备
CN105302807B (zh) 一种获取信息类别的方法和装置
WO2018176913A1 (zh) 搜索方法、装置及非临时性计算机可读存储介质
CN111708909B (zh) 视频标签的添加方法及装置、电子设备、计算机可读存储介质
CN106933878B (zh) 一种信息处理方法及装置
US20230086735A1 (en) Systems and methods for retrieving videos using natural language description
CN110019948B (zh) 用于输出信息的方法和装置
CN111325033B (zh) 实体识别方法、装置、电子设备及计算机可读存储介质
WO2021103594A1 (zh) 一种默契度检测方法、设备、服务器及可读存储介质
CN109753646B (zh) 一种文章属性识别方法以及电子设备
CN111737523B (zh) 一种视频标签、搜索内容的生成方法及服务器
CN107665222B (zh) 关键词的拓展方法和装置
US20220318318A1 (en) Systems and methods for automated information retrieval

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20877864

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20877864

Country of ref document: EP

Kind code of ref document: A1