CN111475613A

CN111475613A - Case classification method and device, computer equipment and storage medium

Info

Publication number: CN111475613A
Application number: CN202010151309.6A
Authority: CN
Inventors: 林凡
Original assignee: OneConnect Financial Technology Co Ltd Shanghai
Current assignee: OneConnect Smart Technology Co Ltd; OneConnect Financial Technology Co Ltd Shanghai
Priority date: 2020-03-06
Filing date: 2020-03-06
Publication date: 2020-07-31

Abstract

The invention discloses a case classification method, a case classification device, computer equipment and a storage medium, wherein the method comprises the following steps: the method comprises the steps of carrying out image recognition on a case material image to obtain case text information, carrying out text semantic clustering on the case text information to determine a target case type, improving accuracy of case type determination, obtaining a legal knowledge graph corresponding to the target case type from a preset legal knowledge graph system, extracting evidence elements, related parties and appeal elements from the case text information according to the legal knowledge graph, dynamically generating weight information according to scoring results of the elements of historical cases corresponding to the target case type, enabling the weight information of the cases to dynamically change along with changes of the elements of the cases, determining a target scoring score of the cases corresponding to the case material image according to the weight information, and carrying out difficult classification on the cases according to the target scoring score, so that accuracy of case classification is improved.

Description

Case classification method and device, computer equipment and storage medium

Technical Field

The present invention relates to the field of data processing, and in particular, to a case classification method and apparatus, a computer device, and a storage medium.

Background

With the rapid development of social economy, the living standard of people is greatly improved, the legal consciousness is stronger, and when some equity disputes are involved, legal weapons are often used for maintaining own equity, which is an embodiment of social progress.

Generally, after receiving case information, a judge will issue a case and assign the case to a judge for processing, different judges are not in the same field of expertise, and different judges have different professional qualities due to factors such as working experience and capability, so before assigning the case, it is necessary to classify the case and assign the case to an appropriate judge according to the category and difficulty level of the case, so as to improve the case judging efficiency.

At present, key words in case materials are mainly identified in a mode of pattern matching, so that the types of cases are determined, and then the difficulty of the cases is determined according to preset rules, which easily causes the case classification to be inaccurate.

Disclosure of Invention

The embodiment of the invention provides a case classification method, a case classification device, computer equipment and a storage medium, and aims to improve the accuracy of current case classification.

In order to solve the above technical problem, an embodiment of the present application provides a case classification method, including:

performing image preprocessing and gray level clustering processing on a scheme material image to obtain a basic image, and performing optical identification on the basic image to obtain scheme text information corresponding to the basic image;

performing text semantic clustering on the set case text information, and determining a target case type corresponding to the set case text information;

acquiring a legal knowledge graph corresponding to the type of the target case from a preset legal knowledge graph system to serve as a target knowledge graph;

extracting an evidence element, an associated party and a appeal element from the filing text information according to the target knowledge map;

generating weight information respectively corresponding to the evidence element, the related party and the appeal element according to the historical filing material image corresponding to the target case type and the grading result corresponding to the historical filing material image;

determining a target score of a case corresponding to the protocol material image based on the evidence element, the associated party, the appeal element and corresponding weight information;

and determining the difficulty category of the case corresponding to the case of the case material image according to a preset classification score interval and the target score.

Optionally, the optically recognizing the basic image to obtain the case text information corresponding to the basic image includes:

carrying out character positioning on the basic image by adopting a scene text detection algorithm, and determining a target character area;

and identifying the target character area by adopting a long-time memory neural network model generated by pre-training to obtain the scheme text information corresponding to the target character area.

Optionally, the performing semantic clustering on the solution text information, and determining the target case type corresponding to the solution text information includes:

performing Chinese automatic word segmentation on the case text information in a preset word segmentation mode to obtain basic word segmentation;

training the basic participles in a word vector mode to obtain space word vectors corresponding to the basic participles;

performing clustering analysis on the space word vectors based on a K-Means aggregation algorithm to obtain a clustering analysis result;

and calculating the Euclidean distance between the clustering analysis result and a preset case type, and taking the preset case type with the minimum Euclidean distance value with the clustering analysis result as the target case type.

Optionally, before the obtaining of the legal knowledge base corresponding to the target case type from the preset legal knowledge base system as the target knowledge base, the case classification method further includes:

acquiring historical case information corresponding to the preset case type, wherein the historical case information comprises historical case setting materials and historical case judgment results;

acquiring a reference knowledge point from the historical case judgment result in a regular matching mode, wherein the reference knowledge point comprises an evidence element, an associated party, an appeal element and a case trial result;

acquiring basic knowledge points corresponding to the reference knowledge points from the historical case planning materials;

acquiring the content corresponding to the reference knowledge point and the content corresponding to the basic knowledge point, and storing the contents into a relational database;

constructing a relation link between the reference knowledge point and the basic knowledge point through a Resource Description Framework (RDF), and storing the reference knowledge point, the basic knowledge point and the relation link into a legal knowledge map database;

constructing a mapping relation between the relational database and the knowledge map database;

and generating a legal knowledge graph system corresponding to the basic field information based on the relational database, the knowledge graph database and the mapping relation.

Optionally, the weight information includes a weight value, and the determining a target score of a case corresponding to the case material image based on the evidence element, the associated party, the appeal element, and the corresponding weight information includes:

calculating the target score of the case corresponding to the protocol material image by the following formula:

wherein S is the target score value, N₁As to the number of the evidence elements,N₂for the number of associated parties, N₃Number of said appeal elements, Q₁A weight value, Q, corresponding to said evidence element₂For the weight value, Q, corresponding to the party concerned₃And the weight value is the weight value corresponding to the appeal element.

In order to solve the above technical problem, an embodiment of the present application further provides a case classifying device, including:

the image processing module is used for carrying out image preprocessing and gray level clustering processing on the scheme material image to obtain a basic image, and carrying out optical identification on the basic image to obtain scheme text information corresponding to the basic image;

the type determining module is used for performing text semantic clustering on the set case text information and determining a target case type corresponding to the set case text information;

the map acquisition module is used for acquiring a legal knowledge map corresponding to the type of the target case from a preset legal knowledge map system to serve as a target knowledge map;

the element extraction module is used for extracting evidence elements, related parties and appeal elements from the scheme text information according to the target knowledge graph;

the weight generation module is used for generating weight information corresponding to the evidence element, the related party and the appeal element respectively according to the historical filing material image corresponding to the target case type and the grading result corresponding to the historical filing material image;

the scoring module is used for determining a target scoring score of a case corresponding to the case material image based on the evidence element, the associated party, the appeal element and corresponding weight information;

and the classification module is used for determining the difficulty and easiness category of the case corresponding to the case material image according to a preset classification score interval and the target score.

Optionally, the image processing module comprises:

the area positioning unit is used for carrying out character positioning on the basic image by adopting a scene text detection algorithm and determining a target character area;

and the character recognition unit is used for recognizing the target character area by adopting a long-time memory neural network model generated by pre-training to obtain the scheme text information corresponding to the target character area.

Optionally, the type determining module includes:

the text word segmentation unit is used for carrying out Chinese automatic word segmentation on the set text information in a preset word segmentation mode to obtain basic word segmentation;

the word vector training unit is used for training the basic participles in a word vector mode to obtain space word vectors corresponding to the basic participles;

the clustering unit is used for carrying out clustering analysis on the space word vectors based on a K-Means aggregation algorithm to obtain a clustering analysis result;

and the type determining unit is used for calculating the Euclidean distance between the clustering analysis result and a preset case type, and taking the preset case type with the minimum Euclidean distance value with the clustering analysis result as the target case type.

Optionally, the case classification method further includes:

a historical information obtaining module, configured to obtain historical case information corresponding to the preset case type, where the historical case information includes historical case setting materials and historical case decision results;

a reference knowledge point acquisition module, configured to acquire a reference knowledge point from the historical case judgment result in a regular matching manner, where the reference knowledge point includes an evidence element, an associated party, an appeal element, and a case judgment result;

a basic knowledge point acquisition module, configured to acquire basic knowledge points corresponding to the reference knowledge points from the historical case planning material;

the data storage module is used for acquiring the content corresponding to the reference knowledge point and the content corresponding to the basic knowledge point and storing the content into a relational database;

the correlation module is used for constructing a relation link between the reference knowledge point and the basic knowledge point through a Resource Description Framework (RDF), and storing the reference knowledge point, the basic knowledge point and the relation link into a legal knowledge map database;

the mapping module is used for constructing a mapping relation between the relational database and the knowledge map database;

and the system generation module is used for generating a legal knowledge graph system corresponding to the basic field information based on the relation database, the knowledge graph database and the mapping relation.

Optionally, the scoring module comprises:

the score calculating unit is used for calculating the target score of the case corresponding to the case material image according to the following formula:

wherein S is the target score value, N₁Is the number of evidence elements, N₂For the number of associated parties, N₃Number of said appeal elements, Q₁A weight value, Q, corresponding to said evidence element₂For the weight value, Q, corresponding to the party concerned₃And the weight value is the weight value corresponding to the appeal element.

In order to solve the above technical problem, an embodiment of the present application further provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the case classification method when executing the computer program.

In order to solve the above technical problem, an embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored, and the computer program, when executed by a processor, implements the steps of the case classification method described above.

On one hand, the case classification method, the device, the computer equipment and the storage medium provided by the embodiment of the invention can be used for obtaining the case text information corresponding to the basic image by carrying out image recognition on the case material image, further carrying out text semantic clustering on the case text information, determining the target case type corresponding to the case text information, improving the accuracy of case type determination and being beneficial to improving the accuracy of subsequent case difficult and easy classification, on the other hand, obtaining the legal knowledge map corresponding to the target case type from a preset legal knowledge map system to be used as the target knowledge map, extracting evidence elements, related parties and appeal elements from the case text information according to the target knowledge map, and generating grading results corresponding to the historical case material image and the historical case material image corresponding to the target case type respectively and the evidence elements, And meanwhile, determining a target score of the case corresponding to the case material image based on the evidence element, the related party, the appeal element and the corresponding weight information, and then determining the difficulty category of the case corresponding to the case material image according to a preset classification score interval and the target score. The accuracy of case classification is improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments of the present invention will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without inventive labor.

FIG. 1 is an exemplary system architecture diagram in which the present application may be applied;

FIG. 2 is a flow chart of one embodiment of a case classification method of the present application;

FIG. 3 is a schematic structural view of one embodiment of a case sorting apparatus according to the present application;

FIG. 4 is a schematic block diagram of one embodiment of a computer device according to the present application.

Detailed Description

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs; the terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "including" and "having," and any variations thereof, in the description and claims of this application and the description of the above figures are intended to cover non-exclusive inclusions. The terms "first," "second," and the like in the description and claims of this application or in the above-described drawings are used for distinguishing between different objects and not for describing a particular order.

Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Referring to fig. 1, as shown in fig. 1, a system architecture 100 may include

terminal devices

101, 102, 103, a network 104 and a server 105. The network 104 serves as a medium for providing communication links between the

terminal devices

101, 102, 103 and the server 105. Network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.

The user may use the

terminal devices

101, 102, 103 to interact with the server 105 via the network 104 to receive or send messages or the like.

The

terminal devices

101, 102, 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, E-book readers, MP3 players (Moving Picture E interface displays the properties Group Audio L layer III, mpeg compression standard Audio layer 3), MP4(Moving Picture E interface displays the properties Group Audio L layer IV, mpeg compression standard Audio layer 4) players, laptop portable computers, desktop computers, and the like.

The server 105 may be a server providing various services, such as a background server providing support for pages displayed on the

terminal devices

101, 102, 103.

It should be noted that the case classification method provided in the embodiments of the present application is executed by a server, and accordingly, the case classification apparatus is disposed in the server.

It should be understood that the number of terminal devices, networks, and servers in fig. 1 is merely illustrative. Any number of terminal devices, networks and servers may be provided according to implementation needs, and the

terminal devices

101, 102 and 103 in this embodiment may specifically correspond to an application system in actual production.

Referring to fig. 2, fig. 2 shows a case classification method according to an embodiment of the present invention, which is described by taking the case classification method applied to the server in fig. 1 as an example, and is detailed as follows:

s201: and performing image preprocessing and gray level clustering processing on the scheme material image to obtain a basic image, and performing optical identification on the basic image to obtain scheme text information corresponding to the basic image.

Specifically, the image preprocessing includes, but is not limited to, a sharpening process, a graying process, a perspective transformation process, a binarization process, and the like. Specifically, since the solution material image may include a plurality of colors, the colors themselves are very susceptible to the influence of factors such as illumination, and the colors of objects of the same type vary greatly, it is difficult for the colors themselves to provide key information, and therefore, it is necessary to perform graying processing on the solution material image to obtain a grayscale map so as to eliminate interference and reduce the complexity of the image and the information processing amount.

The gray processing is a process of converting a color picture into a gray picture, and aims to improve the image quality and make the display effect of the picture clearer. Grayscale processing includes, but is not limited to: component, maximum, average, weighted average, and the like.

The sharpening process is to compensate the outline of the picture, enhance the edge and the gray level jump part of the picture, make the picture clear, and is divided into a spatial domain process and a frequency domain process, and the sharpening process is to highlight the edge and the outline of the ground object on the picture or the characteristics of some linear target elements.

The binarization processing is to set the gray value of a pixel point on the picture to be 0 or 255, namely, to make the whole picture have an obvious black-and-white effect, and the binarization of the picture greatly reduces the data volume in the picture, so that the outline of the target can be highlighted.

In the present case, the processing method of perspective transformation includes, but is not limited to, using a perspective transform () function in OpenCv to perform perspective transformation processing, OpenCv is a cross-platform computer vision library including a large number of open source APIs (interfaces), and provides interfaces in languages of Python, Ruby, MAT L AB, etc., so that many general algorithms in image processing and computer vision are implemented.

Specifically, the gray level clustering processing refers to clustering and segmentation of maximum and minimum values of the preprocessed images.

The gray level clustering of the images uses the idea of 'kernel probability density estimation', and the occurrence times of each color level of the preprocessed images are counted and clustered by solving a color density extreme value.

Wherein, several obvious maximum and minimum points can be obviously seen after the maximum and minimum segmentation is carried out on the color gradation. How many maximum points there are, how many classes there are, and the minimum point is taken as the boundary between the classes.

Further, in order to avoid interference caused by individual isolated noise points to subsequent character recognition, denoising processing needs to be performed on the image after gray level clustering, in this embodiment, denoising processing is performed by isolated region exclusion, and the starting point of isolated region exclusion is as follows: the space between characters and strokes should be relatively compact, and if one region is significantly isolated from other regions, then it is likely that this region is not a character region.

S202: and performing text semantic clustering on the set case text information, and determining the target case type corresponding to the set case text information.

Specifically, the server is preset with a plurality of case types, clustering is performed according to historical cases corresponding to the case types, and a clustering center is used as a word vector of the case type.

S203: and acquiring the legal knowledge graph corresponding to the type of the target case from a preset legal knowledge graph system to serve as the target knowledge graph.

Specifically, the legal knowledge map corresponding to each preset case type is stored in a legal knowledge map library preset by the server, and the legal knowledge map corresponding to the target case type is acquired as the target knowledge map, so that case elements can be extracted and case analysis can be carried out according to the target knowledge map.

The Knowledge map (Knowledge Graph) is a series of different graphs displaying the relationship between the Knowledge development process and the structure, describes Knowledge resources and carriers thereof by using a visualization technology, and excavates, analyzes, constructs, draws and displays Knowledge and the mutual relation between the Knowledge resources and the Knowledge resources. The legal knowledge map in the embodiment is a knowledge map which is used for constructing the association of the evidence elements, the associated parties, the appeal elements and the case trial results and providing the association for the subsequent case combing and trial as reference.

It should be noted that the legal knowledge base in this embodiment is constructed according to the historical case corresponding to the case type, and the specific construction process may refer to the description of the subsequent embodiment, and is not described here again to avoid repetition.

S204: and extracting the evidence elements, the associated parties and the appeal elements from the filing text information according to the target knowledge graph.

Specifically, according to the target knowledge graph, an evidence element, an associated party and a appeal element are extracted from the text information in a natural language semantic recognition mode.

The appeal elements refer to trial results of appeal of case clients, and the more the appeal elements in a reasonable range, the more the cases tend to be complex.

Wherein, the related parties refer to the parties involved in the case, one or more individuals with the same property are generally taken as one party, and the more related parties are generally involved, the more complicated the case becomes.

Wherein, the evidence element refers to key evidence information for providing support for the case.

S205: and generating weight information respectively corresponding to the evidence element, the related party and the appeal element according to the historical filing material image corresponding to the target case type and the grading result corresponding to the historical filing material image.

Specifically, the association between the historical filing material image corresponding to the target case type and the scoring result corresponding to the historical filing material image is analyzed, and weight information corresponding to the evidence element, the association party and the appeal element is generated.

Easily understood, in order to more accurately generate the weight information according to the influence of the elements on the scoring result in the historical case on different types of evidence elements, related parties and appeal elements, the obtained weight information is closer to the actual case, and the accuracy degree is higher.

It should be noted that, in this embodiment, the evidence elements may be of a plurality of different types, and the evidence elements of different types may correspond to different weights, and may be specifically generated according to an actual situation, which is not limited herein.

S206: and determining a target score of the case corresponding to the case material image based on the evidence element, the related party, the appeal element and the corresponding weight information.

Specifically, a preset score calculation mode is adopted to determine a target score of a case corresponding to the case material image according to the evidence element, the associated party, the appeal element and the corresponding weight information.

The preset score calculation mode may be set according to actual requirements, and is not limited herein.

S207: and determining the difficulty category of the case corresponding to the case of the set material image according to the preset classification score interval and the target score.

Specifically, in this embodiment, fixed score intervals are preset, each score interval is classified according to a difficulty level, and the difficulty category of the case corresponding to the record material image can be determined by determining the score interval in which the target score falls.

The preset classification score interval can be set according to actual needs.

In this embodiment, a basic image is obtained by performing image preprocessing and gray level clustering on a case material image, the basic image is optically identified to obtain case text information corresponding to the basic image, and then the case text information is subjected to text semantic clustering to determine a target case type corresponding to the case text information, so that accuracy of case type determination is improved, and accuracy of difficult and easy classification of subsequent cases is improved The method comprises the steps of associating the weight information corresponding to the party and the appeal element, enabling the weight information of the case to dynamically change along with the change of each element of the case, meanwhile, determining a target score of the case corresponding to a case of a case material image based on the evidence element, the associated party, the appeal element and the corresponding weight information, then determining the difficulty category of the case corresponding to the case material image according to a preset classification score interval and the target score, and improving the accuracy of case classification.

In some optional implementation manners of this embodiment, in step S201, optically recognizing the basic image, and obtaining the plan text information corresponding to the basic image includes:

Specifically, since the positions of characters in the base image are not fixed, that is, different base images have different character positions, it is necessary to determine the character area in the base image for character recognition after the content image is subjected to image preprocessing to obtain the base image.

The text area determination method includes, but is not limited to: hough Transform (Hough Transform) algorithm, Hidden Markov Model (HMM) based character recognition algorithm, region feature extraction (MSER) algorithm, and scene text detection (Connectionist textforward Network) algorithm.

Preferably, the embodiment of the present invention determines the text region in the base image by using a scene text detection algorithm, and the implementation manner thereof is as follows: training a basic image by using a Convolutional Neural Network (CNN) model to obtain a depth feature of the image; predicting character edges according to the depth features and a text line construction algorithm (SiDeRefinement), and putting characters with the character edges in the same line into the same rectangular frame according to the rectangular frame with a preset size; the rectangular frames are stringed into a sequence and input into a Recurrent Neural Networks (RNN) model for training, and finally, the training result is regressed by using a full connection layer to obtain a correct character edge, and the correct character edge is connected into a line, so that a character area in the basic image is obtained.

Understandably, when the scene text detection algorithm performs character positioning, the positioning is performed based on line level, that is, the position information of the rectangular frame corresponding to each line is returned.

It should be noted that, in the present embodiment, there may be one or more target text regions, which may be determined according to actual detection results, and the present invention is not limited thereto.

Further, the long-time memory neural network model generated by pre-training is adopted to identify the obtained image of the target character area, and text information contained in the image of the target character area is obtained.

The long-short term memory (L STM) Network is a time recursive Neural Network, is suitable for processing and predicting important events with time sequences and relatively long time sequence intervals and delays, and is a local connection Network, and compared with a full connection Network, the maximum characteristics of the Network are local connectivity and weight sharing.

Before the long-time memory neural network model generated by pre-training is adopted and the target character region is identified, the method also comprises the step of obtaining the pre-training to obtain the long-time memory neural network model, and the process is as follows: sequentially labeling training sample images in a training set, inputting the labeled training sample images into a convolutional neural network-long-and-short-term memory neural network for training, and updating network parameters of the convolutional neural network-long-and-short-term memory neural network by adopting a time sequence classification algorithm and an Adam optimizer so as to obtain the long-and-short-term memory neural network model.

The time sequence classification algorithm (CTC) is used for solving the problem of a time sequence with uncertain alignment relation between input features and output labels, and is an algorithm capable of optimizing model parameters and aligning a segmentation boundary end to end.

It should be noted that when n target character areas are obtained through the scene text detection algorithm, each target character area needs to be sequentially identified to obtain n pieces of text information, where n is a positive integer greater than 1.

In this embodiment, a scene text detection algorithm is used to perform text positioning on a basic image, determine a target text region, and then a long-term and short-term memory neural network model generated by pre-training is used to identify the target text region, so as to obtain the case text information corresponding to the target text region, thereby improving the accuracy of text identification in the basic image.

In some optional implementation manners of this embodiment, in step S202, performing semantic clustering on the case text information, and determining the target case type corresponding to the case text information includes:

and calculating the Euclidean distance between the clustering analysis result and the preset case type, and taking the preset case type with the minimum Euclidean distance value with the clustering analysis result as the target case type.

The method for automatically segmenting Chinese words mainly includes, but is not limited to: a rule-based word segmentation method, a statistic-based word segmentation method, an understanding-based word segmentation method, and a neural network word segmentation method.

The rule-based word segmentation method mainly comprises the following steps: minimum Matching Method (Minimum Matching), forward Maximum Matching Method (Maximum Matching), Reverse Maximum Matching Method (Reverse Directional Maximum Matching), bidirectional Maximum Matching Method (Bi-Directional Maximum Matching, BMM), mark segmentation Method, total segmentation path selection Method, Association-Backtracking Method (AB Method for short), and the like.

The word segmentation method based on statistics mainly comprises the following steps: N-Gram Model, Hidden Markov Model (HMM) sequence notation, Maximum Entropy Model (MEM) sequence notation, Maximum Entropy Model (MEMM) sequence notation, and Conditional Random Field (CRF) sequence notation, and the like.

In artificial intelligence, word vector representation refers primarily to a formal or mathematical description of a language, in order to represent the language in a computer and to enable automatic processing by a computer program. The word vector in this embodiment is expressed in the form of a vector to represent the basic participle.

Specifically, each basic participle is mapped into vectors according to a preset corpus, the vectors are connected together to form a word vector space, each vector is equivalent to a point in the space, and each vector is used as a space word vector.

For example, two to-be-matched participles such as a bmw and a gallop are provided in a certain product name, and all possible classifications of the two to-be-matched participles are obtained according to a preset corpus: "car", "luxury", "animal", "action", and "food". Therefore, a vector representation is introduced for the two to-be-matched participles:

< cars, luxuries, animals, actions, food >

Calculating the probability of the two to-be-matched participles belonging to each classification according to a statistical learning method, wherein the probability learned by a computer is as follows:

bma ═ 0.5,0.2,0.2,0.0,0.1>

Gallop ═ 0.7,0.2,0.0,0.1,0.0>

It will be appreciated that the values of each dimension of the space word vector represent a feature that has some semantic and grammatical interpretation.

Further, the spatial word vectors corresponding to the same basic statement are clustered and analyzed in a clustering mode to obtain a clustering result corresponding to the basic statement, and preferably, the proposal uses a K-Means aggregation algorithm to perform clustering analysis on the spatial word vectors.

The K-means algorithm is a distance-based clustering algorithm, and the distance is used as an evaluation index of similarity, that is, the closer the distance between two objects is, the greater the similarity of the two objects is. The algorithm considers clusters to be composed of closely spaced objects, and therefore targets the resulting compact and independent clusters as final targets.

In this embodiment, each preset case type corresponds to one word vector, the euclidean distance between the clustering analysis result and the preset case type is calculated, and the preset case type having the smallest euclidean distance value from the clustering analysis result is used as the target case type.

In this embodiment, the target case type corresponding to the set text information is determined by performing word segmentation and clustering on the set text information, which is beneficial to subsequently evaluating the difficulty level according to the target case type.

In some optional implementation manners of this embodiment, before step S203, the case classification method further includes:

acquiring historical case information corresponding to a preset case type, wherein the historical case information comprises historical case setting materials and historical case judgment results;

acquiring reference knowledge points from historical case judgment results in a regular matching mode, wherein the reference knowledge points comprise evidence elements, associated parties, appeal elements and case trial results;

acquiring basic knowledge points corresponding to the reference knowledge points from historical case setting materials;

acquiring contents corresponding to the reference knowledge points and contents corresponding to the basic knowledge points, and storing the contents into a relational database;

establishing a relation link between a reference knowledge point and a basic knowledge point through a Resource Description Framework (RDF), and storing the reference knowledge point, the basic knowledge point and the relation link into a legal knowledge map database;

constructing a mapping relation between a relational database and a knowledge map database;

and generating a legal knowledge map system corresponding to the basic field information based on the relational database, the knowledge map database and the mapping relation.

Relational Database (Relational Database) is a Database built on the basis of a Relational Database model, data in the Database are processed by means of concepts and methods such as set algebra and the like, and the Relational Database (Relational Database) is also organized into a group of tables with formal descriptive properties, wherein the table function of the form is essentially a special collection body loaded with data items, and data in the tables can be accessed or recalled in many different ways without reorganizing the Database tables. The definition of a relational database results in a table of metadata or in a formal description of tables, columns, ranges, and constraints. Each table (sometimes referred to as a relationship) contains one or more data categories represented by columns. Each row contains a unique data entity, which is a category defined by the columns.

Resource Description Framework (RDF) is a data model (DataModel) expressed by XM L syntax and used for describing the characteristics of resources and the relationship between the resources, and is used for the situation that information needs to be processed by application programs rather than just displayed for people to watch.

The knowledge map database is a database for storing the basic knowledge points, the core knowledge points and the relationship links, and may be a database, such as Neo4j, Mongo DB, or a logic database.

Specifically, historical case planning materials and historical case judgment results in historical case information are analyzed to determine reference knowledge points and basic knowledge points, content corresponding to the reference knowledge points and content corresponding to the basic knowledge points, relational links are constructed through a resource description framework RDF to obtain a legal knowledge map database, and a legal knowledge map system is generated together with the relational database, so that the reference knowledge points contained in cases can be extracted according to the legal knowledge map system when cases are classified subsequently, and the type and weight information of target cases are determined.

In the embodiment, a legal knowledge graph system is constructed by analyzing the historical case filing materials and the historical case judgment results, so that each element of a case can be extracted and evaluated by using the legal knowledge graph system in the following process, and the accuracy of case classification can be improved.

In some optional implementation manners of this embodiment, in step S206, the weight information includes a weight value, and determining the target score of the case corresponding to the case material image based on the evidence element, the relevant party, the appeal element, and the corresponding weight information includes:

calculating the target score of the case corresponding to the case material image by the following formula:

wherein S is a target score, N₁Number of evidence elements, N₂Number of parties to be associated, N₃Number of appeal elements, Q₁Weight values, Q, corresponding to evidence elements₂Weight value, Q, corresponding to the associated party₃The weight value corresponding to the appeal element.

In the embodiment, the target score is calculated through the formula, so that the case difficulty degree is conveniently quantified, and the accuracy of subsequent difficulty classification is improved.

It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

Fig. 3 is a schematic block diagram of a case classification device corresponding to the case classification method according to the above-described embodiment. As shown in fig. 3, the case classifying apparatus includes an image processing module 31, a type determining module 32, a map obtaining module 33, an element extracting module 34, a weight generating module 35, a scoring module 36, and a classifying module 37. The functional modules are explained in detail as follows:

the image processing module 31 is configured to perform image preprocessing and gray level clustering on the record material image to obtain a basic image, and perform optical identification on the basic image to obtain record text information corresponding to the basic image;

the type determining module 32 is configured to perform text semantic clustering on the set case text information, and determine a target case type corresponding to the set case text information;

the map acquisition module 33 is configured to acquire a legal knowledge map corresponding to the type of the target case from a preset legal knowledge map system, and use the legal knowledge map as a target knowledge map;

the element extraction module 34 is used for extracting the evidence elements, the associated parties and the appeal elements from the case setting text information according to the target knowledge graph;

the weight generation module 35 is configured to generate weight information corresponding to the evidence element, the related party and the appeal element respectively according to the historical filing material image corresponding to the target case type and the scoring result corresponding to the historical filing material image;

the scoring module 36 is used for determining a target scoring score of a case corresponding to the case material image based on the evidence element, the associated party, the appeal element and the corresponding weight information;

and the classification module 37 is configured to determine a difficulty category of the case corresponding to the case material image according to a preset classification score interval and a target score.

Optionally, the image processing module 31 comprises:

Optionally, the type determining module 32 includes:

the text word segmentation unit is used for carrying out Chinese automatic word segmentation on the file text information in a preset word segmentation mode to obtain basic word segmentation;

and the type determining unit is used for calculating the Euclidean distance between the clustering analysis result and the preset case type, and taking the preset case type with the minimum Euclidean distance value with the clustering analysis result as the target case type.

Optionally, the case classification method further includes:

the system comprises a historical information acquisition module, a history information processing module and a history information processing module, wherein the historical information acquisition module is used for acquiring historical case information corresponding to a preset case type, and the historical case information comprises historical case setting materials and historical case judgment results;

the reference knowledge point acquisition module is used for acquiring reference knowledge points from the historical case judgment results in a regular matching mode, wherein the reference knowledge points comprise evidence elements, associated parties, appeal elements and case judgment results;

the basic knowledge point acquisition module is used for acquiring basic knowledge points corresponding to the reference knowledge points from historical case planning materials;

the data storage module is used for acquiring the content corresponding to the reference knowledge point and the content corresponding to the basic knowledge point and storing the content into the relational database;

and the system generation module is used for generating a legal knowledge map system corresponding to the basic field information based on the relational database, the knowledge map database and the mapping relation.

Optionally, the scoring module 36 includes:

the score calculating unit is used for calculating the target score of the case corresponding to the case material image through the following formula:

For the specific definition of the case classification device, reference may be made to the above definition of the case classification method, which is not described herein again. The modules in the case sorting device can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In order to solve the technical problem, an embodiment of the present application further provides a computer device. Referring to fig. 4, fig. 4 is a block diagram of a basic structure of a computer device according to the present embodiment.

The computer device 4 comprises a memory 41, a processor 42, a network interface 43 communicatively connected to each other via a system bus. It is noted that only the computer device 4 having the components connection memory 41, processor 42, network interface 43 is shown, but it is understood that not all of the shown components are required to be implemented, and that more or fewer components may be implemented instead. As will be understood by those skilled in the art, the computer device is a device capable of automatically performing numerical calculation and/or information processing according to a preset or stored instruction, and the hardware includes, but is not limited to, a microprocessor, an Application Specific Integrated Circuit (ASIC), a Programmable Gate Array (FPGA), a Digital Signal Processor (DSP), an embedded device, and the like.

The computer device can be a desktop computer, a notebook, a palm computer, a cloud server and other computing devices. The computer equipment can carry out man-machine interaction with a user through a keyboard, a mouse, a remote controller, a touch panel or voice control equipment and the like.

The memory 41 includes at least one type of readable storage medium including a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or D interface display memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Read Only Memory (ROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a Programmable Read Only Memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory 41 may be an internal storage unit of the computer device 4, such as a hard disk or a memory of the computer device 4. In other embodiments, the memory 41 may also be an external storage device of the computer device 4, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the computer device 4. Of course, the memory 41 may also include both internal and external storage devices of the computer device 4. In this embodiment, the memory 41 is generally used for storing an operating system installed in the computer device 4 and various types of application software, such as program codes for controlling electronic files. Further, the memory 41 may also be used to temporarily store various types of data that have been output or are to be output.

The processor 42 may be a Central Processing Unit (CPU), controller, microcontroller, microprocessor, or other data Processing chip in some embodiments. The processor 42 is typically used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is configured to execute the program code stored in the memory 41 or process data, such as program code for executing control of an electronic file.

The network interface 43 may comprise a wireless network interface or a wired network interface, and the network interface 43 is generally used for establishing communication connection between the computer device 4 and other electronic devices.

The present application further provides another embodiment, which is to provide a computer readable storage medium, wherein the computer readable storage medium stores an interface display program, and the interface display program can be executed by at least one processor, so as to cause the at least one processor to execute the steps of the case classification method.

Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present application may be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present application.

It is to be understood that the above-described embodiments are merely illustrative of some, but not restrictive, of the broad invention, and that the appended drawings illustrate preferred embodiments of the invention and do not limit the scope of the invention. This application is capable of embodiments in many different forms and is provided for the purpose of enabling a thorough understanding of the disclosure of the application. Although the present application has been described in detail with reference to the foregoing embodiments, it will be apparent to one skilled in the art that the present application may be practiced without modification or with equivalents of some of the features described in the foregoing embodiments. All equivalent structures made by using the contents of the specification and the drawings of the present application are directly or indirectly applied to other related technical fields and are within the protection scope of the present application.

Claims

1. A case classification method is characterized by comprising the following steps:

2. The case classification method according to claim 1, wherein the optically recognizing the basic image to obtain the case text information corresponding to the basic image comprises:

3. The case classification method according to claim 1, characterized in that said performing semantic clustering of the text of the proposal text information and determining the target case type corresponding to the proposal text information comprises:

4. The case classification method according to claim 3, characterized in that before said obtaining the legal knowledge graph corresponding to said target case type from a preset legal knowledge graph system as a target knowledge graph, said case classification method further comprises:

5. The case classification method according to any one of claims 1 to 4, wherein the weight information includes a weight value, and the determining a target score of the case corresponding to the case material image based on the evidence element, the interested party, the appeal element, and the corresponding weight information includes:

6. A case sorting apparatus, comprising:

7. Case classification apparatus according to claim 6, characterised in that said image processing module comprises:

8. The case sorting apparatus according to claim 6, further comprising:

9. A computer arrangement comprising a memory, a processor and a computer program stored in said memory and executable on said processor, characterized in that said processor implements the case classification method according to any of claims 1 to 5 when executing said computer program.

10. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the case classification method according to any one of claims 1 to 5.