CN111651997A - Method and system for recommending case - Google Patents
Method and system for recommending case Download PDFInfo
- Publication number
- CN111651997A CN111651997A CN202010312497.6A CN202010312497A CN111651997A CN 111651997 A CN111651997 A CN 111651997A CN 202010312497 A CN202010312497 A CN 202010312497A CN 111651997 A CN111651997 A CN 111651997A
- Authority
- CN
- China
- Prior art keywords
- initial
- case
- initial class
- class
- target
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 239000013598 vector Substances 0.000 claims abstract description 67
- 238000003062 neural network model Methods 0.000 claims abstract description 21
- 210000000056 organ Anatomy 0.000 claims description 17
- 238000012549 training Methods 0.000 claims description 17
- 238000004590 computer program Methods 0.000 claims description 8
- 238000005516 engineering process Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 4
- 238000004891 communication Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 206010063385 Intellectualisation Diseases 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- 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
- G06F16/355—Class or cluster creation or modification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/18—Legal services
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Tourism & Hospitality (AREA)
- Evolutionary Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Databases & Information Systems (AREA)
- Technology Law (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention provides a scheme recommendation method and a system, wherein the method comprises the following steps: acquiring the attribute characteristics of a target case; inputting the attribute characteristics into a neural network model, and acquiring a class case vector corresponding to a target case; according to the distance between the class vector corresponding to the target case and the class vector corresponding to each candidate case in the case library, acquiring a first initial class set, a second initial class set, a third initial class set, a fourth initial class set, a fifth initial class set, a sixth initial class set and a seventh initial class set corresponding to the target case from the case library; and acquiring a plurality of target classes of the target cases so as to judge the target cases according to the judgment result of each target class. The method and the system for recommending the class case can improve the accuracy of recommending the class case, provide more accurate and controllable class cases for reference according to requirements, improve the efficiency of recommending the class case and enable the recommended case to be more reasonable.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a scheme recommendation method and system.
Background
The classification recommendation is a method capable of effectively improving judgment efficiency, but the classification recommendation faces many challenges, the existing classification recommendation method is not efficient, accurate classification recommendation cannot be achieved, especially when cases with high plot similarity with target cases are in large numbers, law enforcement personnel can face mass selection, the efficiency is affected, misleading is possibly caused, and the method runs counter to the original purpose of classification recommendation.
Therefore, a method for recommending a case is needed.
Disclosure of Invention
In order to solve the above problems, embodiments of the present invention provide a method and a system for recommending a scenario.
In a first aspect, an embodiment of the present invention provides a method for recommending a class, including:
acquiring attribute characteristics of a target case, wherein the characteristic attributes comprise one or more of party information, administrative organ information, case program information, case routing information, illegal fact information, evidence information and law foundation information of the target case;
inputting the attribute characteristics into a neural network model, and acquiring class case vectors corresponding to the target case, wherein the neural network model is obtained by training a plurality of sample cases as training samples and class case vectors corresponding to the sample cases as labels;
according to the distance between the class vector corresponding to the target case and the class vector corresponding to each candidate case in the case library, acquiring a first initial class set, a second initial class set, a third initial class set, a fourth initial class set, a fifth initial class set, a sixth initial class set and a seventh initial class set corresponding to the target case from the case library;
acquiring a plurality of target classes of the target case according to the first initial class set, the second initial class set, the third initial class set, the fourth initial class set, the fifth initial class set, the sixth initial class set and the seventh initial class set, so as to judge the target case according to a judgment result of each target class;
wherein the first initial class set comprises a plurality of first initial classes, for any first initial class, the similarity between the party of the any first initial class and the party of the target case is greater than a first preset threshold, the second initial class set comprises a plurality of second initial classes, for any second initial class, the similarity between the administrative authority of the second initial class and the administrative authority of the target case is greater than a second preset threshold, the third initial class set comprises a plurality of third initial classes, for any third initial class, the similarity between the case program of the any third initial class and the case program of the target case is greater than a third preset threshold, the fourth initial class set comprises a plurality of fourth initial classes, and for any fourth initial class, the similarity between the case program of the any fourth initial class and the case program of the target case is greater than a fourth preset threshold The fifth initial class set comprises a plurality of fifth initial classes, for any fifth initial class, the similarity between the law violation fact of any fifth initial class and the law violation fact of the target case is greater than a fifth preset threshold, the sixth initial class set comprises a plurality of sixth initial classes, for any sixth initial class, the similarity between the evidence of any sixth initial class and the evidence of the target case is greater than a sixth preset threshold, the seventh initial class set comprises a plurality of seventh initial classes, and for any seventh initial class, the similarity between the law evidence of any seventh initial class and the law evidence of the target case is greater than a seventh preset threshold.
Preferably, the obtaining a plurality of target cases of the target case according to the first initial class set, the second initial class set, the third initial class set, the fourth initial class set, the fifth initial class set, the sixth initial class set, and the seventh initial class set specifically includes:
according to a first preset weight of the first initial class set, a second preset weight of the second initial class set, a third preset weight of the third initial class set, a fourth preset weight of the fourth initial class set, a fifth preset weight of the fifth initial class set, a sixth preset weight of the sixth initial class set and a seventh preset weight of the seventh initial class set, carrying out weighted average on each initial class to obtain the similarity of each initial class;
and taking the initial class with the similarity larger than a preset threshold value as the target class of the target case.
Preferably, the attribute features of the target case are obtained by:
performing word segmentation processing on the target case, and converting the segmented target case into an initial vector through a text embedding technology;
extracting information from the initial vector to obtain text characteristics, wherein the text attribute characteristics comprise one or more of a party, an administrative organ, a case program, a case cause, illegal facts, evidence and law bases;
and carrying out semantic embedding on the text features to obtain the attribute features.
Preferably, the segmented target case is converted into an initial vector by a text embedding technology, specifically as follows:
and converting the participled target case into an initial vector by a Skip-gram method.
Preferably, the semantic embedding is performed on the text features to obtain the attribute features, which is as follows:
and performing semantic embedding on the text features through a PV-DM method to obtain the attribute features.
Preferably, the neural network model is a Transformer model.
In a second aspect, an embodiment of the present invention provides a class recommendation system, including:
the system comprises an acquisition module, a storage module and a processing module, wherein the acquisition module is used for acquiring the attribute characteristics of a target case, and the characteristic attributes comprise one or more of party information, administrative organ information, case program information, case routing information, illegal fact information, evidence information and law foundation information of the target case;
the class case module is used for inputting the attribute characteristics into a neural network model to obtain class case vectors corresponding to the target cases, and the neural network model is obtained by training with a plurality of sample cases as training samples and with the class case vectors corresponding to the sample cases as labels;
a computing module, configured to obtain, from a case library, a first initial class set, a second initial class set, a third initial class set, a fourth initial class set, a fifth initial class set, a sixth initial class set, and a seventh initial class set corresponding to the target case according to a distance between a class vector corresponding to the target case and a class vector corresponding to each candidate case in the case library, where the first initial class set includes a plurality of first initial classes, and for any first initial class, a similarity between a principal of the any first initial class and a principal of the target case is greater than a first preset threshold, and the second initial class set includes a plurality of second initial classes, and for any second initial class, a similarity between an administrative organ of the second initial class and an administrative organ of the target case is greater than a second preset threshold, the third initial class set comprises a plurality of third initial classes, for any third initial class, the similarity between the case program of any third initial class and the case program of the target case is greater than a third preset threshold, the fourth initial class set comprises a plurality of fourth initial classes, for any fourth initial class, the similarity between the case program of any fourth initial class and the case program of the target case is greater than a fourth preset threshold, the fifth initial class set comprises a plurality of fifth initial classes, for any fifth initial class, the similarity between the violation fact of any fifth initial class and the violation fact of the target case is greater than a fifth preset threshold, the sixth initial class set comprises a plurality of sixth initial classes, for any sixth initial class, the similarity between the evidence of any sixth initial class and the case of the target case is greater than a sixth preset threshold, the seventh initial class set comprises a plurality of seventh initial classes, and for any seventh initial class, the similarity between the legal basis of any seventh initial class and the legal basis of the target case is greater than a seventh preset threshold;
and the recommending module is used for acquiring a plurality of target cases of the target cases according to the first initial class set, the second initial class set, the third initial class set, the fourth initial class set, the fifth initial class set, the sixth initial class set and the seventh initial class set so as to judge the target cases according to the judgment result of each target case.
In a third aspect, an embodiment of the present invention provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor executes the computer program to implement the steps of the class recommendation method provided in the first aspect of the present invention.
In a fourth aspect, an embodiment of the present invention provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the case recommendation method provided in the first aspect of the present invention.
The method and the system for recommending the class case can improve the accuracy of recommending the class case, provide more accurate and controllable-quantity class cases for reference according to requirements, and provide data support for other intelligent applications as well as use the processed information in the system; the efficiency of recommending the class plan is improved, and the recommended case is more reasonable; the intellectualization is high, and the actual use requirement of a court is met.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a flowchart of a method for recommending a scenario according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a scenario recommendation system according to an embodiment of the present invention;
fig. 3 is a schematic physical structure diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, 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, but 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.
Fig. 1 is a flowchart of a category recommendation method according to an embodiment of the present invention, as shown in fig. 1, the method includes:
s1, acquiring attribute characteristics of the target case, wherein the characteristic attributes comprise one or more of party information, administrative organ information, case program information, case routing information, illegal fact information, evidence information and law foundation information of the target case;
s2, inputting the attribute characteristics into a neural network model, and obtaining class case vectors corresponding to the target case, wherein the neural network model is obtained by training with a plurality of sample cases as training samples and class case vectors corresponding to the sample cases as labels;
s3, acquiring a first initial class set, a second initial class set, a third initial class set, a fourth initial class set, a fifth initial class set, a sixth initial class set and a seventh initial class set corresponding to the target case from the case library according to the distance between the class vector corresponding to the target case and the class vector corresponding to each candidate case in the case library;
s4, obtaining a plurality of target cases of the target cases according to the first initial class set, the second initial class set, the third initial class set, the fourth initial class set, the fifth initial class set, the sixth initial class set and the seventh initial class set, so as to judge the target cases according to the judgment result of each target case;
wherein the first initial class set comprises a plurality of first initial classes, for any first initial class, the similarity between the party of the any first initial class and the party of the target case is greater than a first preset threshold, the second initial class set comprises a plurality of second initial classes, for any second initial class, the similarity between the administrative authority of the second initial class and the administrative authority of the target case is greater than a second preset threshold, the third initial class set comprises a plurality of third initial classes, for any third initial class, the similarity between the case program of the any third initial class and the case program of the target case is greater than a third preset threshold, the fourth initial class set comprises a plurality of fourth initial classes, and for any fourth initial class, the similarity between the case program of the any fourth initial class and the case program of the target case is greater than a fourth preset threshold The fifth initial class set comprises a plurality of fifth initial classes, for any fifth initial class, the similarity between the law violation fact of any fifth initial class and the law violation fact of the target case is greater than a fifth preset threshold, the sixth initial class set comprises a plurality of sixth initial classes, for any sixth initial class, the similarity between the evidence of any sixth initial class and the evidence of the target case is greater than a sixth preset threshold, the seventh initial class set comprises a plurality of seventh initial classes, and for any seventh initial class, the similarity between the law evidence of any seventh initial class and the law evidence of the target case is greater than a seventh preset threshold.
In the embodiment of the invention, the target case is a case which needs to be subjected to class recommendation, and the attribute characteristics of the target case are extracted according to the description of the target case, wherein the attribute characteristics comprise one or more of the following seven aspects, namely a party of the target case, an administrative department of the target case, a case cause of the target case, a violation fact of the target case, evidence of the target case and a legal basis of the target case.
The information contained in the attribute features of the target case can be selected according to actual needs, for example, if the evidence of the target case is temporarily unavailable, the attribute features of the target case are composed of other six aspects, in a specific programming implementation, the feature attributes can be represented as a vector, each element in the vector corresponds to seven aspects of the feature attributes, and when a certain aspect is not contained in the feature attributes, the element in the corresponding position is set to 0.
And then inputting the attribute characteristics of the target case into a neural network model to obtain a class case vector corresponding to the target case.
And then calculating the distance between the class vector corresponding to the target case and the class vector corresponding to each candidate case in the case library, specifically, the case library comprises a plurality of candidate cases which are cases processed before and judged, collecting and sorting the cases, and establishing the case library. Moreover, all candidate cases in the case library have the same party classified as one type, have the same administrative department classified as one type, have the same case program classified as one type, have the same case organization classified as one type, have the same law violation fact classified as one type, have the same evidence classified as one type, and have the same legal basis classified as one type.
Calculating the distance between the class vector corresponding to the target case and the class vector corresponding to each candidate case, specifically, calculating the distance between the principal information in all candidate cases and the principal information of the target case for the principal information in the class vector corresponding to the target case, and adding all candidate cases with the distance greater than a first preset threshold value into a first initial class set.
Similarly, for the administrative authorities in the class vector corresponding to the target case, the distances between the administrative authorities in all the candidate cases and the administrative authorities of the target case are calculated, and all the candidate cases with the distances larger than a second preset threshold are added into a second initial class set.
Similarly, for the case programs in the class vector corresponding to the target case, the distances between the case programs in all candidate cases and the case program of the target case are calculated, and all candidate cases with the distances larger than a third preset threshold are added into a third initial class set.
Similarly, for the case pairs in the class vector corresponding to the target case, the distances between the case pairs in all candidate cases and the case pair of the target case are calculated, and all candidate cases with the distances larger than a fourth preset threshold are added into a fourth initial class set.
Similarly, for the illegal facts in the class vector corresponding to the target case, the distances between the illegal facts in all the candidate cases and the illegal facts of the target case are calculated, and all the candidate cases with the distances larger than a fifth preset threshold are added into a fifth initial class set.
Similarly, for the evidences in the class vector corresponding to the target case, the distances between the evidences in all the candidate cases and the evidences of the target case are calculated, and all the candidate cases with the distances larger than a sixth preset threshold value are added into a sixth initial class set.
Similarly, for the legal basis in the class vector corresponding to the target case, the distance between the legal basis in all candidate cases and the legal basis of the target case is calculated, and all candidate cases with the distance greater than a seventh preset threshold are added into a seventh initial class set.
Then, according to the first initial class set, the second initial class set, the third initial class set, the fourth initial class set, the fifth initial class set, the sixth initial class set and the seventh initial class set, the class with the highest similarity is selected as the target class of the target case, and recommended to relevant law enforcement personnel, so that the relevant personnel can judge the target case by referring to the judgment result of the target class.
Specifically, according to a first initial class set, a second initial class set, a third initial class set, a fourth initial class set, a fifth initial class set, a sixth initial class set, and a seventh initial class set, a class with the highest similarity is selected as a target class of the target case, and the following methods may be used:
for any initial class, judging whether the initial class is used as a target class according to the times of the initial class appearing in the first initial class set, the second initial class set, the third initial class set, the fourth initial class set, the fifth initial class set, the sixth initial class set and the seventh initial class set. For example, the seven initial class sets are added to include the initial class, which means that the similarity between the initial class and the target class is high, so that the initial class can be used as the target class.
Or carrying out weighted average on the similarity of the parties, the similarity of administrative organs, the similarity of case programs, the similarity of case law, the similarity of illegal facts, the similarity of evidences and the preset weight of law according to the similarity, and taking the initial class with the value larger than the preset condition as the target class.
The class case recommendation method provided by the embodiment of the invention can improve the precision rate of class case recommendation, provide more accurate class cases with controllable quantity for reference according to requirements, and provide data support for other intelligent applications as well as use the processed information in the system; the efficiency of recommending the class plan is improved, and the recommended case is more reasonable; the intellectualization is high, and the actual use requirement of a court is met.
On the basis of the foregoing embodiment, preferably, the acquiring a plurality of target classes of the target case according to the first initial class set, the second initial class set, the third initial class set, the fourth initial class set, the fifth initial class set, the sixth initial class set, and the seventh initial class set specifically includes:
according to a first preset weight of the first initial class set, a second preset weight of the second initial class set, a third preset weight of the third initial class set, a fourth preset weight of the fourth initial class set, a fifth preset weight of the fifth initial class set, a sixth preset weight of the sixth initial class set and a seventh preset weight of the seventh initial class set, carrying out weighted average on each initial class to obtain the similarity of each initial class;
and taking the initial class with the similarity larger than a preset threshold value as the target class of the target case.
Specifically, for any initial class, if the initial class appears in the first initial class set, the third initial class set and the sixth initial class set, the similarity of the initial class is the sum of the first preset weight, the third preset weight and the sixth preset weight, according to the method, the similarity of each initial class is calculated, and the initial class with the similarity greater than a preset threshold is taken as the target class.
On the basis of the above embodiment, preferably, the attribute features of the target case are obtained by:
performing word segmentation processing on the target case, and converting the segmented target case into an initial vector through a text embedding technology;
extracting information from the initial vector to obtain text characteristics, wherein the text attribute characteristics comprise one or more of a party, an administrative organ, a case program, a case cause, illegal facts, evidence and law bases;
and carrying out semantic embedding on the text features to obtain the attribute features.
Generally, the target case is described by text at the time of initial description, the target case is subjected to word segmentation, and the input text is converted into an initial vector through a text embedding technology.
And extracting information from the initial vector to obtain text features, wherein the text features comprise one or more of the party, the administrative agency, the case program, the case law, the law violation fact, the evidence and the law foundation, specifically, the content contained in the text features and the content contained in the attribute features are in one-to-one correspondence, and if the attribute features comprise only the party, the administrative agency and the case program, the text features comprise only the party, the administrative agency and the case program.
And then, carrying out semantic embedding on the text features to obtain attribute features.
On the basis of the foregoing embodiment, preferably, the segmented target case is converted into an initial vector by a text embedding technique, which is specifically as follows:
and converting the participled target case into an initial vector by a Skip-gram method.
The Skip-gram method belongs to a method in word2vec, and the Skip-gram method is adopted in the embodiment of the invention. Compared with the traditional one-hot vector, the Skip-gram method can enable the generated vector to have semantic information, so that the accuracy of the method is greatly improved.
On the basis of the foregoing embodiment, preferably, the semantic embedding is performed on the text feature to obtain the attribute feature, specifically as follows:
and performing semantic embedding on the text features through a PV-DM method to obtain the attribute features.
PV-DM: paragraph level semantic embedding method. On the basis of having the word vector, the generated vector can be relatively accurately made to obtain semantic information.
On the basis of the above embodiment, preferably, the neural network model is a Transformer model.
Transformer model-model in seq2seq form, used as classifier in the method. The Transformer model is used as a model proposed by Google, and has the characteristics of fast training and good processing capability on long texts. In the method, high efficiency can be kept when large-scale texts are processed.
In summary, by means of the technical scheme of the embodiment of the invention, the efficiency of recommending the class plan is improved, and the processed information is not only used in the system, but also provides data support for other intelligent applications; the accuracy of the recommendation of the class plan is improved, and the mediation scheme is more transparent, standardized and rationalized.
Fig. 2 is a schematic structural diagram of a scenario recommendation system according to an embodiment of the present invention, as shown in fig. 2, the system includes: an obtaining module 201, a classification module 202, a calculating module 203 and a recommending module 204, wherein:
the acquisition module 201 is configured to acquire attribute characteristics of a target case, where the characteristic attributes include one or more of party information, administrative information, case program information, case routing information, illegal fact information, evidence information, and law compliance information of the target case;
the class model module 202 is configured to input the attribute features into a neural network model, and obtain class vectors corresponding to the target cases, where the neural network model is obtained by training using a plurality of sample cases as training samples and using the class vectors corresponding to the sample cases as labels;
the computing module 203 is configured to obtain, from the case library, a first initial class set, a second initial class set, a third initial class set, a fourth initial class set, a fifth initial class set, a sixth initial class set, and a seventh initial class set corresponding to the target case according to a distance between a class vector corresponding to the target case and a class vector corresponding to each candidate case in the case library, where the first initial class set includes a plurality of first initial classes, and for any first initial class, a similarity between a principal of the any first initial class and a principal of the target case is greater than a first preset threshold, and the second initial class set includes a plurality of second initial classes, and for any second initial class, a similarity between an administrative organ of the second initial class and an administrative organ of the target case is greater than a second preset threshold, the third initial class set comprises a plurality of third initial classes, for any third initial class, the similarity between the case program of any third initial class and the case program of the target case is greater than a third preset threshold, the fourth initial class set comprises a plurality of fourth initial classes, for any fourth initial class, the similarity between the case program of any fourth initial class and the case program of the target case is greater than a fourth preset threshold, the fifth initial class set comprises a plurality of fifth initial classes, for any fifth initial class, the similarity between the violation fact of any fifth initial class and the violation fact of the target case is greater than a fifth preset threshold, the sixth initial class set comprises a plurality of sixth initial classes, for any sixth initial class, the similarity between the evidence of any sixth initial class and the case of the target case is greater than a sixth preset threshold, the seventh initial class set comprises a plurality of seventh initial classes, and for any seventh initial class, the similarity between the legal basis of any seventh initial class and the legal basis of the target case is greater than a seventh preset threshold;
the recommending module 204 is configured to obtain a plurality of target classes of the target case according to the first initial class set, the second initial class set, the third initial class set, the fourth initial class set, the fifth initial class set, the sixth initial class set, and the seventh initial class set, so that the target case is determined according to a determination result of each target class.
The system embodiment provided in the embodiments of the present invention is for implementing the above method embodiments, and for details of the process and the details, reference is made to the above method embodiments, which are not described herein again.
Fig. 3 is a schematic entity structure diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 3, the electronic device may include: a processor (processor)301, a communication Interface (communication Interface)302, a memory (memory)303 and a bus 304, wherein the processor 301, the communication Interface 302 and the memory 303 complete communication with each other through the bus 304. The communication interface 302 may be used for information transfer of an electronic device. Processor 301 may call logic instructions in memory 303 to perform a method comprising:
acquiring attribute characteristics of a target case, wherein the characteristic attributes comprise one or more of party information, administrative organ information, case program information, case routing information, illegal fact information, evidence information and law foundation information of the target case;
inputting the attribute characteristics into a neural network model, and acquiring class case vectors corresponding to the target case, wherein the neural network model is obtained by training a plurality of sample cases as training samples and class case vectors corresponding to the sample cases as labels;
according to the distance between the class vector corresponding to the target case and the class vector corresponding to each candidate case in the case library, acquiring a first initial class set, a second initial class set, a third initial class set, a fourth initial class set, a fifth initial class set, a sixth initial class set and a seventh initial class set corresponding to the target case from the case library;
acquiring a plurality of target classes of the target case according to the first initial class set, the second initial class set, the third initial class set, the fourth initial class set, the fifth initial class set, the sixth initial class set and the seventh initial class set, so as to judge the target case according to a judgment result of each target class;
wherein the first initial class set comprises a plurality of first initial classes, for any first initial class, the similarity between the party of the any first initial class and the party of the target case is greater than a first preset threshold, the second initial class set comprises a plurality of second initial classes, for any second initial class, the similarity between the administrative authority of the second initial class and the administrative authority of the target case is greater than a second preset threshold, the third initial class set comprises a plurality of third initial classes, for any third initial class, the similarity between the case program of the any third initial class and the case program of the target case is greater than a third preset threshold, the fourth initial class set comprises a plurality of fourth initial classes, and for any fourth initial class, the similarity between the case program of the any fourth initial class and the case program of the target case is greater than a fourth preset threshold The fifth initial class set comprises a plurality of fifth initial classes, for any fifth initial class, the similarity between the law violation fact of any fifth initial class and the law violation fact of the target case is greater than a fifth preset threshold, the sixth initial class set comprises a plurality of sixth initial classes, for any sixth initial class, the similarity between the evidence of any sixth initial class and the evidence of the target case is greater than a sixth preset threshold, the seventh initial class set comprises a plurality of seventh initial classes, and for any seventh initial class, the similarity between the law evidence of any seventh initial class and the law evidence of the target case is greater than a seventh preset threshold.
In addition, the logic instructions in the memory 303 may be implemented in the form of software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the above-described method embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented to perform the transmission method provided in the foregoing embodiments when executed by a processor, and for example, the method includes:
acquiring attribute characteristics of a target case, wherein the characteristic attributes comprise one or more of party information, administrative organ information, case program information, case routing information, illegal fact information, evidence information and law foundation information of the target case;
inputting the attribute characteristics into a neural network model, and acquiring class case vectors corresponding to the target case, wherein the neural network model is obtained by training a plurality of sample cases as training samples and class case vectors corresponding to the sample cases as labels;
according to the distance between the class vector corresponding to the target case and the class vector corresponding to each candidate case in the case library, acquiring a first initial class set, a second initial class set, a third initial class set, a fourth initial class set, a fifth initial class set, a sixth initial class set and a seventh initial class set corresponding to the target case from the case library;
acquiring a plurality of target classes of the target case according to the first initial class set, the second initial class set, the third initial class set, the fourth initial class set, the fifth initial class set, the sixth initial class set and the seventh initial class set, so as to judge the target case according to a judgment result of each target class;
wherein the first initial class set comprises a plurality of first initial classes, for any first initial class, the similarity between the party of the any first initial class and the party of the target case is greater than a first preset threshold, the second initial class set comprises a plurality of second initial classes, for any second initial class, the similarity between the administrative authority of the second initial class and the administrative authority of the target case is greater than a second preset threshold, the third initial class set comprises a plurality of third initial classes, for any third initial class, the similarity between the case program of the any third initial class and the case program of the target case is greater than a third preset threshold, the fourth initial class set comprises a plurality of fourth initial classes, and for any fourth initial class, the similarity between the case program of the any fourth initial class and the case program of the target case is greater than a fourth preset threshold The fifth initial class set comprises a plurality of fifth initial classes, for any fifth initial class, the similarity between the law violation fact of any fifth initial class and the law violation fact of the target case is greater than a fifth preset threshold, the sixth initial class set comprises a plurality of sixth initial classes, for any sixth initial class, the similarity between the evidence of any sixth initial class and the evidence of the target case is greater than a sixth preset threshold, the seventh initial class set comprises a plurality of seventh initial classes, and for any seventh initial class, the similarity between the law evidence of any seventh initial class and the law evidence of the target case is greater than a seventh preset threshold.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (9)
1. A method for recommending a scenario, comprising:
acquiring attribute characteristics of a target case, wherein the characteristic attributes comprise one or more of party information, administrative organ information, case program information, case routing information, illegal fact information, evidence information and law foundation information of the target case;
inputting the attribute characteristics into a neural network model, and acquiring class case vectors corresponding to the target case, wherein the neural network model is obtained by training a plurality of sample cases as training samples and class case vectors corresponding to the sample cases as labels;
according to the distance between the class vector corresponding to the target case and the class vector corresponding to each candidate case in the case library, acquiring a first initial class set, a second initial class set, a third initial class set, a fourth initial class set, a fifth initial class set, a sixth initial class set and a seventh initial class set corresponding to the target case from the case library;
acquiring a plurality of target classes of the target case according to the first initial class set, the second initial class set, the third initial class set, the fourth initial class set, the fifth initial class set, the sixth initial class set and the seventh initial class set, so as to judge the target case according to a judgment result of each target class;
wherein the first initial class set comprises a plurality of first initial classes, for any first initial class, the similarity between the party of the any first initial class and the party of the target case is greater than a first preset threshold, the second initial class set comprises a plurality of second initial classes, for any second initial class, the similarity between the administrative authority of the second initial class and the administrative authority of the target case is greater than a second preset threshold, the third initial class set comprises a plurality of third initial classes, for any third initial class, the similarity between the case program of the any third initial class and the case program of the target case is greater than a third preset threshold, the fourth initial class set comprises a plurality of fourth initial classes, and for any fourth initial class, the similarity between the case program of the any fourth initial class and the case program of the target case is greater than a fourth preset threshold The fifth initial class set comprises a plurality of fifth initial classes, for any fifth initial class, the similarity between the law violation fact of any fifth initial class and the law violation fact of the target case is greater than a fifth preset threshold, the sixth initial class set comprises a plurality of sixth initial classes, for any sixth initial class, the similarity between the evidence of any sixth initial class and the evidence of the target case is greater than a sixth preset threshold, the seventh initial class set comprises a plurality of seventh initial classes, and for any seventh initial class, the similarity between the law evidence of any seventh initial class and the law evidence of the target case is greater than a seventh preset threshold.
2. The method of claim 1, wherein the obtaining a plurality of target cases of the target case according to the first initial set of cases, the second initial set of cases, the third initial set of cases, the fourth initial set of cases, the fifth initial set of cases, the sixth initial set of cases, and the seventh initial set of cases specifically comprises:
according to a first preset weight of the first initial class set, a second preset weight of the second initial class set, a third preset weight of the third initial class set, a fourth preset weight of the fourth initial class set, a fifth preset weight of the fifth initial class set, a sixth preset weight of the sixth initial class set and a seventh preset weight of the seventh initial class set, carrying out weighted average on each initial class to obtain the similarity of each initial class;
and taking the initial class with the similarity larger than a preset threshold value as the target class of the target case.
3. The case recommendation method according to claim 1, wherein the attribute features of the target case are obtained by:
performing word segmentation processing on the target case, and converting the segmented target case into an initial vector through a text embedding technology;
extracting information from the initial vector to obtain text characteristics, wherein the text attribute characteristics comprise one or more of a party, an administrative organ, a case program, a case cause, illegal facts, evidence and law bases;
and carrying out semantic embedding on the text features to obtain the attribute features.
4. The case recommendation method according to claim 3, wherein the segmented target cases are converted into initial vectors by text embedding technology, specifically as follows:
and converting the participled target case into an initial vector by a Skip-gram method.
5. The method of claim 3, wherein the semantic embedding is performed on the text features to obtain the attribute features, and the method specifically includes:
and performing semantic embedding on the text features through a PV-DM method to obtain the attribute features.
6. The case recommendation method of claim 1, wherein the neural network model is a Transformer model.
7. A scenario recommendation system, comprising:
the system comprises an acquisition module, a storage module and a processing module, wherein the acquisition module is used for acquiring the attribute characteristics of a target case, and the characteristic attributes comprise one or more of party information, administrative organ information, case program information, case routing information, illegal fact information, evidence information and law foundation information of the target case;
the class case module is used for inputting the attribute characteristics into a neural network model to obtain class case vectors corresponding to the target cases, and the neural network model is obtained by training with a plurality of sample cases as training samples and with the class case vectors corresponding to the sample cases as labels;
a computing module, configured to obtain, from a case library, a first initial class set, a second initial class set, a third initial class set, a fourth initial class set, a fifth initial class set, a sixth initial class set, and a seventh initial class set corresponding to the target case according to a distance between a class vector corresponding to the target case and a class vector corresponding to each candidate case in the case library, where the first initial class set includes a plurality of first initial classes, and for any first initial class, a similarity between a principal of the any first initial class and a principal of the target case is greater than a first preset threshold, and the second initial class set includes a plurality of second initial classes, and for any second initial class, a similarity between an administrative organ of the second initial class and an administrative organ of the target case is greater than a second preset threshold, the third initial class set comprises a plurality of third initial classes, for any third initial class, the similarity between the case program of any third initial class and the case program of the target case is greater than a third preset threshold, the fourth initial class set comprises a plurality of fourth initial classes, for any fourth initial class, the similarity between the case program of any fourth initial class and the case program of the target case is greater than a fourth preset threshold, the fifth initial class set comprises a plurality of fifth initial classes, for any fifth initial class, the similarity between the violation fact of any fifth initial class and the violation fact of the target case is greater than a fifth preset threshold, the sixth initial class set comprises a plurality of sixth initial classes, for any sixth initial class, the similarity between the evidence of any sixth initial class and the case of the target case is greater than a sixth preset threshold, the seventh initial class set comprises a plurality of seventh initial classes, and for any seventh initial class, the similarity between the legal basis of any seventh initial class and the legal basis of the target case is greater than a seventh preset threshold;
and the recommending module is used for acquiring a plurality of target cases of the target cases according to the first initial class set, the second initial class set, the third initial class set, the fourth initial class set, the fifth initial class set, the sixth initial class set and the seventh initial class set so as to judge the target cases according to the judgment result of each target case.
8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the class recommendation method according to any one of claims 1 to 6 when executing the program.
9. A non-transitory computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the class recommendation method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010312497.6A CN111651997A (en) | 2020-04-20 | 2020-04-20 | Method and system for recommending case |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010312497.6A CN111651997A (en) | 2020-04-20 | 2020-04-20 | Method and system for recommending case |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111651997A true CN111651997A (en) | 2020-09-11 |
Family
ID=72346495
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010312497.6A Pending CN111651997A (en) | 2020-04-20 | 2020-04-20 | Method and system for recommending case |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111651997A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180373782A1 (en) * | 2017-06-27 | 2018-12-27 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for recommending answer to question based on artificial intelligence |
| CN110442684A (en) * | 2019-08-14 | 2019-11-12 | 山东大学 | A kind of class case recommended method based on content of text |
| CN110647631A (en) * | 2018-06-25 | 2020-01-03 | 阿里巴巴集团控股有限公司 | Case recommendation method and device, storage medium and processor |
-
2020
- 2020-04-20 CN CN202010312497.6A patent/CN111651997A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180373782A1 (en) * | 2017-06-27 | 2018-12-27 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for recommending answer to question based on artificial intelligence |
| CN110647631A (en) * | 2018-06-25 | 2020-01-03 | 阿里巴巴集团控股有限公司 | Case recommendation method and device, storage medium and processor |
| CN110442684A (en) * | 2019-08-14 | 2019-11-12 | 山东大学 | A kind of class case recommended method based on content of text |
Non-Patent Citations (1)
| Title |
|---|
| 吕宾等: "基于主题模型的法院文本典型案例推荐", 《微电子学与计算机》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112613501A (en) | Information auditing classification model construction method and information auditing method | |
| CN109872162B (en) | Wind control classification and identification method and system for processing user complaint information | |
| CN106651057B (en) | Mobile terminal user age prediction method based on installation package sequence list | |
| CN111371767B (en) | Malicious account identification method, malicious account identification device, medium and electronic device | |
| CN109299258A (en) | A kind of public sentiment event detecting method, device and equipment | |
| CN107545038B (en) | Text classification method and equipment | |
| CN111931809A (en) | Data processing method and device, storage medium and electronic equipment | |
| CN108228622A (en) | The sorting technique and device of traffic issues | |
| CN112818995B (en) | Image classification method, device, electronic equipment and storage medium | |
| CN112988963A (en) | User intention prediction method, device, equipment and medium based on multi-process node | |
| CN106778851A (en) | Social networks forecasting system and its method based on Mobile Phone Forensics data | |
| CN111782793A (en) | Intelligent customer service processing method, system and equipment | |
| CN114663002A (en) | Method and equipment for automatically matching performance assessment indexes | |
| CN109960719A (en) | A kind of document handling method and relevant apparatus | |
| CN104966109A (en) | Medical laboratory report image classification method and apparatus | |
| CN111723182B (en) | Key information extraction method and device for vulnerability text | |
| CN116402166B (en) | Training method and device of prediction model, electronic equipment and storage medium | |
| CN111651997A (en) | Method and system for recommending case | |
| CN113688232A (en) | Method and device for classifying bidding texts, storage medium and terminal | |
| CN113010664A (en) | Data processing method and device and computer equipment | |
| CN112632229A (en) | Text clustering method and device | |
| CN117391076B (en) | Acquisition method and device of identification model of sensitive data, electronic equipment and medium | |
| CN111680513B (en) | Feature information identification method and device and computer readable storage medium | |
| CN118152560A (en) | Extraction and classification integrated method and system based on entity feature fusion | |
| CN114722175A (en) | Automatic construction method and device for intelligent customer service prompt problem and terminal equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200911 |