CN115577077A - Criminal case evidence guiding method based on NLP technology - Google Patents
Criminal case evidence guiding method based on NLP technology Download PDFInfo
- Publication number
- CN115577077A CN115577077A CN202211306617.7A CN202211306617A CN115577077A CN 115577077 A CN115577077 A CN 115577077A CN 202211306617 A CN202211306617 A CN 202211306617A CN 115577077 A CN115577077 A CN 115577077A
- Authority
- CN
- China
- Prior art keywords
- case
- evidence
- recommendation
- criminal
- corpus
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language 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
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/18—Legal services; Handling legal documents
Abstract
The invention discloses a criminal case evidence guiding method based on NLP technology, which comprises the steps of judging the type of a case; specifically, on the basis of improving the accuracy of brief case word segmentation, a text classification algorithm in an NLP technology is used for case category prediction so as to improve the accuracy of case category judgment; a recommendation engine collects case information; generating a recommendation result of case evidence collection content; and loading the recommendation result by the task flow to form the task to be handled. The invention provides a method for combining an artificial intelligence recommendation engine with a task flow, and combines a case evidence obtaining method and a evidence obtaining task flow by utilizing automatic recommendation, so as to solve the problems that a case is returned to a hospital for inspection to be supplemented and detected due to errors in the traditional method of collecting evidence of a criminal case by means of experience and manual record, or the evidence is insufficient and the criminal case is reduced even prevented from being taken out due to errors, and effectively improve the case handling efficiency and the final complaint rate of a public security organization in criminal case detection.
Description
Technical Field
The invention relates to the technical field of criminal case evidence collection, in particular to a criminal case evidence guiding method based on an NLP (non line segment) technology.
Background
In criminal case investigation, firstly, inaccurate conditions often occur in registered case types, and if the evidence collection function of various cases is realized by depending on hard coding alone, the condition that an evidence system is inconsistent with the case types is bound to occur.
Secondly, since various crimes have different crime constitutions and various cases have different case characteristics, various specific criminal cases inevitably have a lot of differences in the aspects of evidence forms, evidence systems, evidence obtaining ideas, evidence obtaining methods, evidence obtaining cautions and the like. Therefore, evidence collection of specific criminal cases follows general evidence standards, evidence collection ideas and evidence collection requirements of the criminal cases, and adheres to personalized and differentiated evidence standards, evidence collection ideas and evidence collection requirements by combining the constituent characteristics of various specific criminal cases and the case characteristics.
The personalized and differentiated evidence standards, evidence-taking ideas and evidence-taking requirements need guidance of considerable experience and strong theoretical knowledge of the police officers. In criminal case investigation, criminal investigation personnel especially need to strengthen the consciousness of important evidences, and need to carefully learn and master the legal provisions of the concepts and the types of evidences, legal procedures for collecting evidences, examination and verification of evidences and the like, and comprehensively, objectively and truly collect evidences according to law, so that the criminal action work can be smoothly carried out. Otherwise, the conditions that evidence collection is incomplete, evidence obtaining thinking is unclear and the like occur in actual work, so that the conditions that criminal facts are unclear and evidence is not adopted by an insufficient evidence inspection organization occur, cases cannot be made true, and the case detecting and handling efficiency and the case solving rate are influenced.
In criminal case investigation, evidence forms are divided into eight legal forms, (one) material evidence; (II) book and certificate; (III) witness testimony; (IV) the victim states; (V) providing and resolving criminal suspects and defendees; (VI) appraising the opinions; (VII) recording of experiments such as investigation, inspection, identification and investigation; (eight) audiovisual material and electronic data, but in the physical evidence classification, the content requirements of the physical evidence are greatly different according to the case types and even the specific conditions in the similar case handling process. For example, the rape case needs to extract fine spots, bloodstains, traces and the like, the intentional personnel killing case needs to extract crime instruments, bloodcoats, bloodstains and the like, and the telecom fraud case needs to extract case-involved bank cards, network equipment for criminal activities, mobile phone cards and the like, which are guided by the experience and theoretical knowledge of criminal polices.
In the process of processing the evidence extraction by the public security organization, the condition that the evidence is insufficient and is returned by the inspected hospital for supplementary investigation often occurs. Criminal policemen often do not only deal with cases together at the same time in the case-dealing process, but often a plurality of cases are interlaced together in time, and errors are caused. The main reasons for returning to the public security organization to supplement and detect the physical causes are as follows:
1. the fact of the main crime is unclear, and contradictions exist among main evidences;
2. crime composition requirements are deficient;
3. the main plot is not verified;
4. omitting important crime facts and case-by-case crime suspects who should investigate criminal responsibilities;
5. other facts of great significance to conviction and sentencing have not been investigated.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a criminal case evidence guiding method based on NLP technology, which can guide the evidence collection in the criminal case, thereby avoiding the case returning to the public security organization for detecting the physical affairs and supplementing the affairs due to the unclear main crime facts, contradictions between the main evidences and the lack of crime constitution essentials, and further improving the efficiency and the accuracy of the public security organization in criminal case evidence search concentration.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows.
The criminal case evidence guiding method based on the NLP technology combines an artificial intelligence recommendation engine with task flow, realizes the combination of an evidence obtaining method and an evidence obtaining task flow in an automatically recommended case evidence obtaining process, and specifically comprises the following steps:
judging the case type; specifically, on the basis of improving the accuracy of brief case word segmentation, a text classification algorithm in an NLP technology is used for case category prediction so as to improve the accuracy of case category judgment;
a recommendation engine collects case information;
generating a recommendation result of case evidence collection content;
and loading the recommendation result by the task flow to form the task to be handled.
Preferably, the method for improving the word segmentation accuracy of the brief case comprises the steps of segmenting words of the case by using a conditional random field word segmentation algorithm and labeling a public security domain corpus.
Preferably, the public security domain corpus takes a national language commission modern balanced corpus as a training model, and a labeling specification (information processing modern Chinese word class label set specification) (GB/T20532-2006) consistent with the national language commission modern balanced corpus is selected for expanding the public security domain corpus;
the raw corpus of the public security field corpus is derived from basic case data in the public security industry; firstly, performing lexical analysis on a raw corpus by a model generated by a national language commission modern balanced corpus, and correcting labels in an output result so as to improve the labeling efficiency of the public security field corpus;
in the lexical analysis of the raw corpora, the lexical of new words which are not registered needs to be marked manually so as to convert the raw corpora into the cooked corpora; and (5) continuing to use the conditional random field as a training algorithm to finish the training of the public security field corpus.
Preferably, the text classification algorithm is a linear support vector machine classifier, and classification samples required by the linear support vector machine classifier and raw corpus required in the process of improving the accuracy of word segmentation of brief cases are managed by a sample training management module comprising sample preview and sample uploading functions.
Preferably, before collecting case information, the recommendation engine further comprises a recommendation subject determining unit, a recommendation object determining unit and a recommendation algorithm determining unit;
the recommended main body is a case to be dealt with; after the case type judgment is completed, generating a recommendation main body;
the recommended objects are various material evidence materials; and decomposing eight types of evidence forms in a common evidence system of criminal cases into operable minimum-granularity evidence elements as recommended objects.
Preferably, the recommendation algorithm is to classify the historical cases according to recommendation categories and count the number of various evidence materials used by the cases, and specifically includes the following steps:
firstly, preparing a case table, wherein the prepared case table comprises case numbers and case types;
next, an evidence element table is prepared, and the evidence element table includes the case number, the evidence type, and the evidence element.
Preferably, the recommendation engine collects case information by associating the case table with the evidence element table, and statistics is performed on evidence elements existing in history cases of the same type through a unitary algorithm according to the judged case type.
Preferably, the unary algorithm for counting the evidence elements is further improved to a binary and ternary algorithm, that is, the correlation degree between the probability of occurrence of one evidence element and the probability of occurrence of another evidence element in a case is further improved, and the evidence elements with high correlation degree are counted as the associated evidence elements.
Preferably, the recommendation result for generating case evidence gathering content is to calculate the proportion of all evidence elements in cases of respective case categories, and the proportion value is a recommendation threshold value range, and generate a recommendation result.
Preferably, the tasks to be handled are distributed to corresponding policemen; after the police officer finishes the task, the task flow automatically acquires an evidence collection and verification leader according to the organization structure of the police officer, the evidence collection and verification leader carries out verification by the leader, and after the leader completes the verification, the evidence collection and verification leader sends the evidence collection and verification leader to legal department personnel for rechecking according to relevant regulations.
Due to the adoption of the technical scheme, the technical progress of the invention is as follows.
The invention provides a case type judging method, which comprises the following steps: on the basis of improving the accuracy of word segmentation of the brief cases, the case category is predicted by using a text classification algorithm in the NLP technology, the accuracy of case category judgment is improved, and the problem that the subsequent evidence guidance is obstructed due to inaccurate or wrong case category labels of a part of cases in the manual case processing process is solved.
According to the automatic recommendation method, the evidence collection content of the case can be automatically recommended according to the case type and by combining with the historical cases of the same type, so that the criminal police can make decisions in the evidence collection process, collect what material evidence is collected, what investigation is performed, what identification is performed and the like, and how to perform the recommendation of the evidence collection mode is carried out, so that important facts in the evidence collection process are avoided being ignored due to personal errors or experiences of policemen.
According to the task flow method, after recommended contents of case evidence collected contents are generated, the generated recommended results are distributed to corresponding policemen to form tasks to be handled; after the task to be dealt with is completed, the complete evidence chain and the main evidence information are all completed, so that the problems that the main crime fact is unclear and the main evidence is contradictory due to personal errors and limited experience are fundamentally solved.
The invention integrally provides a method for combining an artificial intelligence recommendation engine with a task flow, and combines a case evidence obtaining method and a evidence obtaining task flow by utilizing automatic recommendation, so that the problems that a case is returned by a hospital for inspection to be supplemented and detected, or evidence is insufficient and a criminal is reduced or even a complaint is avoided due to errors caused by the fact that the conventional method relies on experience and manual record to collect criminal case evidence are solved, and the case handling efficiency and the final complaint rate of a public security organization in criminal case detection can be effectively improved.
Drawings
FIG. 1 is a block flow diagram of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
A criminal case evidence guiding method based on an NLP technology adopts a method of combining an artificial intelligent recommendation engine with a mission flow, utilizes automatic recommendation to combine a case evidence obtaining method and a evidence obtaining mission flow in a case evidence obtaining process, and is used for solving the problem that a traditional mode relies on experience, manually records errors existing in collecting criminal case evidence, causes a case to be returned to a hospital for supplementary investigation, or reduces criminal evidence or even avoids prosecution due to insufficient evidence caused by errors, thereby effectively improving case handling efficiency and final prosecution rate of public security institutions in criminal case investigation.
As shown in fig. 1, the method specifically includes the following steps:
s1: and judging the case type.
S2: and initiating a case type confirmation flow.
S3: the recommendation engine collects case information.
S4: a recommendation is generated.
S5: and loading the recommendation result by the task flow to form the task to be handled.
Each step is described in detail below:
s1: and judging the case type.
The method for judging the case category of the case on which the case is put specifically comprises the following steps of on the basis of improving the accuracy rate of brief case word segmentation, predicting the case category by using a text classification algorithm in an NLP technology to improve the accuracy rate of case category judgment, and specifically comprises the following steps:
s11: the word segmentation accuracy of the brief case is improved.
The word segmentation algorithm used in the invention is a conditional random field, and the word segmentation accuracy indexes of the conditional random field on MSR (Microsoft Asia institute named entity recognition corpus) respectively reach: the precision rate is 96.86%, the recall rate is 96.64%, the blending average value of the precision rate and the recall rate is 96.75%, the recall rate of new words is 71.54, and the recall rate of login words is 97.33, so that the requirements of industrial application can be completely met.
In order to more accurately identify the special words in the case, a public security domain corpus is selected and labeled, so that the case word segmentation accuracy is improved, and the case classification accuracy is further improved. Considering the scale and quality of the corpus, the national language commission modern balanced corpus is used as a training model, and a labeling standard consistent with the national language commission modern balanced corpus, namely a modern Chinese word class label set standard for information processing (GB/T20532-2006) is selected for expanding the public security domain corpus.
The raw corpus of the public security field corpus is derived from basic case data in the public security industry, the raw corpus is firstly analyzed according to the morphology through a model generated by the national language commission modern balanced corpus, and labels are corrected in output results so as to improve the labeling efficiency of the field corpus.
In the lexical analysis of the raw corpus, the morphology of new unregistered words needs to be marked manually. If the "packet loss fraud" appearing in the output result is divided into "packet loss": dynamic nouns, "fraud": nouns, it needs to be defined as a "packet loss fraud" in the output result of the raw corpus: a noun is a term. All such domain named entities are identified, and the raw corpus is finally converted into a corpus.
And after the labeling is finished, the raw corpus is converted into the cooked corpus, and the conditional random field is continuously used as a training algorithm to finish the training of the field model.
S12: the case classification accuracy is improved on the basis of improving the word segmentation accuracy.
After the word segmentation accuracy rate of a brief case is improved, a text classification algorithm in an NLP technology is used for predicting the category of the case.
In order to improve the accuracy of classification prediction as much as possible, a text classification algorithm in the NLP technology selects and uses a linear support vector machine classifier, the algorithm is a general algorithm, has corresponding thesis support, needs code realization, and can also call a corresponding open source software package.
In order to improve the operation efficiency of the whole NLP processing process and reduce the operation difficulty, a model management module needs to be realized, and the specific contents are as follows: and establishing a sample training management module which comprises a sample preview function and a sample uploading function and is used for managing the raw linguistic data required in the process of improving the word segmentation accuracy of the brief case and the classification samples required by the classifier of the branching support vector machine.
S2: and initiating a case type confirmation flow.
And after the case type is judged, the case type is distributed to corresponding policemen, and the judged case type is confirmed.
After case category judgment and confirmation are completed, in order to realize the combination of the artificial intelligence recommendation engine and the task flow through the steps S3, S4 and S5 and the combination of the evidence obtaining method of the case evidence obtaining process of automatic recommendation and the evidence obtaining task flow, a recommendation subject, a recommendation object and a recommendation algorithm need to be determined, and the following concrete steps are as follows:
1. determining a recommended subject and a recommended object:
recommending that the subject is a case to be dealt with; after the case type judgment is completed, a recommended subject is generated.
Recommending objects as various material evidence materials; eight types of evidence forms in a criminal case common evidence system are decomposed into operable minimum-granularity evidence elements, which are referred to as evidence elements for short hereinafter and serve as recommended objects.
The first type: material evidence
(1) Tools for crime, such as knifes for killing people and iron bars for prying and pressing safety boxes during theft.
(2) Physical objects directly attacked by criminals, such as stolen color TV, robbed cash, dirty dirt, bribed cars.
(3) Articles which represent the consequences of social hazards from criminal activities, such as damaged machinery, burned buildings, stolen telecommunication lines.
(4) Illegal articles generated by criminal acts, such as drugs, counterfeit money, manufactured shotguns, ammunition, and the like.
(5) The trace of criminal behavior, such as the trace left by prying a door.
(6) In the criminal process, articles or traces which can reflect the perpetrators are left at the crime scene, such as clothes worn during crime, hair of criminal suspects torn off at the scene by victims at rape time, fingerprints and foot prints left at the scene by the perpetrators, and the like.
(7) Criminal suspects fight against various goods and substance traces counterfeited by detection, such as after a certain first crime, clothes of a certain second crime are left on the scene, a second killer scene counterfeited for transferring sight, and the like.
(8) Articles or traces which can indicate that a criminal suspect and an defendee are innocent, such as resident identification cards which are registered and used in hotels confirm evidence that the criminal suspect is lost before.
(9) Other articles or traces for ascertaining the true status of the case.
The second type: book certificate
(1) Contract book
(2) Various public and private documents
(3) Lease contract
(4) Marriage certificate
(5) House property card
(6) Trade mark
(7) Letter
(8) Telegraph
(9) Trade mark
(10) Vehicle and ship ticket
(11) Various transportation documents
(12) Traffic accident responsibility subscription book
And the like, and the addition is carried out according to the case requirements.
In the third category: witness testimony (all cases are consistent without decomposition)
The fourth type: statement of the victim (all cases are consistent without decomposition)
The fifth type: debate and explanation of criminal suspect and defender (all cases are consistent without decomposition)
The sixth type: opinion identification
(1) Forensic identification: forensic pathological identification, forensic clinical identification, forensic psychosis identification, forensic physical evidence identification, forensic toxicant identification
(2) Material evidence identification: document identification, trace identification and trace identification
The seventh type: notes for investigation, examination, identification and investigation experiment
(1) Survey and test writing record
(2) Examination record
(3) Recognition pen record
(4) Investigation experiment record
Eighth type: audiovisual data, electronic data
(1) Electronic mail
(2) Electronic data exchange
(3) Online chat logging
(4) Blog
(5) Micro-blogs
(6) Short message of mobile phone
(7) Electronic signature
(8) Domain name
2. Determining recommendation algorithms
And after the recommended subject and the object are determined, collecting corresponding evidence material contents of the historical case, and carrying out evidence material statistical recommendation according to the predicted case category.
Recommendation priority calculation
The basic principle is as follows: the evidence material of the same case is similar in a certain time.
And classifying the historical cases according to recommended categories, and counting the number of various evidence materials used by the cases.
First, an case table tb _ entry _ creatnalcase is prepared, which contains the case number (case _ id) and the case type (case _ category _ id).
Next, an evidence element table tb _ entry _ evidence is prepared, which contains the case number (case _ id), the evidence type (evidence _ category _ id), and the evidence element (evidence _ element _ id).
And associating the case table with the evidence element table, counting the evidence elements existing in the history cases of the same type according to the judged case type through a unitary algorithm, and calculating the proportion of all the evidence elements in the cases of the respective case type, wherein the proportion value is a recommended threshold value range.
In order to improve the accuracy of the recommended structure, the unitary algorithm of the evidence elements can be further improved to binary and ternary algorithms, namely the correlation degree of the probability of one element appearing and the probability of the other element appearing in a case is improved, and the evidence elements with high correlation degree are recommended as the related elements.
And setting a recommendation threshold value to generate recommendation information.
S3: the recommendation engine collects case information.
And S3, associating the case table with the evidence element table according to the recommendation subject, the recommendation object and the determination rule of the recommendation algorithm, counting the evidence elements in the historical cases of the same type according to the judged case type, and finishing the collection of case information.
S4: a recommendation is generated.
Step S4, according to the determination rule of the recommendation algorithm, calculating the proportion of all evidence elements counted in the step S3 in the cases of the respective case types, wherein the proportion value is a recommendation threshold value range, and generating a recommendation result, such as: for cases of the type of case-id-case-90% containing evidence of "material evidence" - "knife" open to the sun, please note the collection.
S5: and loading the recommendation result by the task flow to form the task to be handled.
After the recommendation result of the case evidence collection content is generated, the recommendation result is loaded by the task flow, the sponsoring and cooperative investigators of the case are obtained according to the case information, the generated recommendation result is distributed to the corresponding policemen, and the method can be realized by forming the tasks to be handled of the sponsoring and cooperative investigators in a main informatization system used by criminal polices.
After the police finish the task, the task flow automatically acquires evidence collection and audit leaders according to organization structures of units where the investigators are located, the leaders conduct audit, and after the leaders complete audit, the leaders send the evidence collection and audit leaders to legal department personnel for rechecking according to relevant regulations.
When the invention is used, after all tasks to be done are completed, the complete evidence chain and the main evidence information are all completed, thereby fundamentally solving the problems of unclear main crime facts, contradictions between main evidences and deficiency of crime forming essential elements caused by personal errors and experience limitation, and improving the case detecting efficiency and the success rate of prosecution.
The criminal case evidence guiding method based on the NLP technology is only one of the general methods in the application of the present invention, and in addition, the processing ideas and principles similar to the principles of the present invention are also within the protection scope of the present invention.
Claims (10)
1. A criminal case evidence guiding method based on NLP technology is characterized in that: the method combines an artificial intelligence recommendation engine with a task flow to realize the combination of a forensics method and a forensics task flow in the case forensics process of automatic recommendation, and specifically comprises the following steps:
judging the case type; specifically, on the basis of improving the accuracy of brief case word segmentation, a text classification algorithm in an NLP technology is used for case category prediction so as to improve the accuracy of case category judgment;
the recommendation engine collects case information;
generating a recommendation result of case evidence collection content;
and loading the recommendation result by the task flow to form the task to be handled.
2. The criminal case evidence guiding method based on NLP technology as claimed in claim 1, wherein: the method for improving the accuracy of word segmentation of the brief case comprises the steps of performing word segmentation on the case by using a conditional random field word segmentation algorithm and labeling a public security field corpus.
3. The criminal case evidence guiding method based on NLP technology according to claim 2, characterized in that: the public security domain corpus takes a national language commission modern balanced corpus as a training model, and a labeling specification (information processing modern Chinese word class label set specification) (GB/T20532-2006) consistent with the national language commission modern balanced corpus is selected for expanding the public security domain corpus;
the raw corpus of the public security field corpus is derived from basic case data in the public security industry; firstly, performing lexical analysis on a raw corpus through a model generated by a national language commission modern balanced corpus, and correcting a label in an output result so as to improve the labeling efficiency of the public security field corpus;
in the lexical analysis of the raw corpus, the new words which are not logged in need to be manually marked with the lexical method so as to convert the raw corpus into the cooked corpus; and (5) continuing to use the conditional random field as a training algorithm to finish the training of the public security field corpus.
4. The criminal case evidence guiding method based on NLP technology as claimed in claim 3, wherein: the text classification algorithm is a linear support vector machine classifier, and classification samples required by the linear support vector machine classifier and raw corpora required in the process of improving the accuracy rate of brief case situation word segmentation are managed by a sample training management module comprising sample preview and sample uploading functions.
5. The criminal case evidence guiding method based on NLP technology as claimed in claim 1, wherein: before collecting case information, the recommendation engine also comprises a recommendation subject, a recommendation object and a recommendation algorithm;
the recommended main body is a case to be dealt with; after the case type judgment is completed, generating a recommendation main body;
the recommended objects are various material evidence materials; and decomposing eight types of evidence forms in a common evidence system of criminal cases into operable minimum-granularity evidence elements as recommended objects.
6. The criminal case evidence guiding method based on NLP technology according to claim 5, characterized in that: the recommendation algorithm is used for classifying historical cases according to recommended categories and counting the number of various evidence materials used by the cases, and specifically comprises the following steps:
firstly, preparing a case table, wherein the prepared case table comprises case numbers and case types;
next, an evidence element table is prepared, and the evidence element table includes the case number, the evidence type, and the evidence element.
7. The criminal case evidence guiding method based on NLP technology as claimed in claim 6, wherein: and the recommendation engine collects case information, namely, the case table is associated with the evidence element table, and evidence elements existing in the history cases of the same type are counted through a unitary algorithm according to the judged case type.
8. The criminal case evidence guiding method based on NLP technology as claimed in claim 7, wherein: and further improving the unary algorithm for counting the evidence elements to a binary algorithm and a ternary algorithm, namely, the correlation degree of the occurrence probability of one evidence element and the occurrence probability of the other evidence element in a case, and taking the evidence elements with high correlation degree as the associated evidence elements for counting.
9. The criminal case evidence guiding method based on NLP technology according to claim 7, characterized in that: and the recommendation result for generating case evidence collection content is to calculate the proportion of all evidence elements in cases of respective case types, wherein the proportion value is a recommendation threshold value range, and generate a recommendation result.
10. The criminal case evidence guiding method based on NLP technology as claimed in claim 1, wherein: the tasks to be handled are distributed to corresponding policemen; after the police finish the task, the task flow automatically acquires evidence collection and audit leaders according to organization structures of the units where the police are located, the leaders conduct audit, and after the leaders complete audit, the leaders send the evidence collection and audit leaders to legal department personnel for rechecking according to relevant regulations.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211306617.7A CN115577077A (en) | 2022-10-25 | 2022-10-25 | Criminal case evidence guiding method based on NLP technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211306617.7A CN115577077A (en) | 2022-10-25 | 2022-10-25 | Criminal case evidence guiding method based on NLP technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115577077A true CN115577077A (en) | 2023-01-06 |
Family
ID=84586498
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211306617.7A Pending CN115577077A (en) | 2022-10-25 | 2022-10-25 | Criminal case evidence guiding method based on NLP technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115577077A (en) |
-
2022
- 2022-10-25 CN CN202211306617.7A patent/CN115577077A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | Applying data mining in investigating money laundering crimes | |
| CN112053221A (en) | Knowledge graph-based internet financial group fraud detection method | |
| CN106874951B (en) | Passenger attention rating method and device | |
| CN108053111B (en) | Elevator shunting method and system | |
| CN111915468B (en) | Network anti-fraud active inspection and early warning system | |
| Ozgul et al. | Incorporating data sources and methodologies for crime data mining | |
| CN112053222A (en) | Knowledge graph-based internet financial group fraud detection method | |
| Baechler | Document fraud: Will your identity be secure in the twenty-first century? | |
| WO2021229568A1 (en) | System, method and computer program product for mitigating customer onboarding risk | |
| Jakubik et al. | Online emotions during the storming of the US Capitol: evidence from the social media network Parler | |
| Chen et al. | Identifying tampering operations in image operator chains based on decision fusion | |
| Bacalu | Digital policing tools as social control technologies: data-driven predictive algorithms, automated facial recognition surveillance, and law enforcement biometrics | |
| Adensamer et al. | “Part Man, Part Machine, All Cop”: Automation in Policing | |
| CN111353716A (en) | Illegal fundamentation detection method, system and computer readable storage medium | |
| Hamdy et al. | Criminal act detection and identification model | |
| US20090060285A1 (en) | Rating individuals on a voluntary basis using legal non-discriminatory criteria | |
| CN115577077A (en) | Criminal case evidence guiding method based on NLP technology | |
| Krysovatyy et al. | Classification Method of Fictitious Enterprises Based on Gaussian Naive Bayes | |
| Khan | How Data Mining Can Help Curb City Crime | |
| CN107909092A (en) | Being accused of based on body identity characteristic fears personnel cruelly and studies and judges recognition methods and system | |
| Henman | E-government, targeting and data profiling: Policy and ethical issues of differential treatment | |
| Afridi | The Current Status of Forensic Science and its Impact on Administration of Criminal Justice System in Pakistan: An Analytical Study | |
| Custers | Data Mining and Profiling in Big Data | |
| Kavoliūnaitė-Ragauskienė | Right to Privacy and Data Protection Concerns Raised by the Development and Usage of Face Recognition Technologies in the European Union | |
| KR102524828B1 (en) | Detective agency brokerage system based on big data |
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 |