CN114611515B - Method and system for identifying enterprise actual control person based on enterprise public opinion information - Google Patents
Method and system for identifying enterprise actual control person based on enterprise public opinion information Download PDFInfo
- Publication number
- CN114611515B CN114611515B CN202210106055.5A CN202210106055A CN114611515B CN 114611515 B CN114611515 B CN 114611515B CN 202210106055 A CN202210106055 A CN 202210106055A CN 114611515 B CN114611515 B CN 114611515B
- Authority
- CN
- China
- Prior art keywords
- enterprise
- relation
- public opinion
- training
- relationship
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 17
- 238000012549 training Methods 0.000 claims abstract description 105
- 230000011218 segmentation Effects 0.000 claims abstract description 55
- 238000012795 verification Methods 0.000 claims description 47
- 238000010801 machine learning Methods 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 18
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Classifications
-
- 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
- G06F40/295—Named entity recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/335—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/105—Human resources
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Entrepreneurship & Innovation (AREA)
- Strategic Management (AREA)
- Artificial Intelligence (AREA)
- Operations Research (AREA)
- Evolutionary Computation (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Economics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Marketing (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method and a system for identifying an enterprise actual control person based on enterprise public opinion information, wherein the method comprises the following steps: training an enterprise name word segmentation model through an NLP word segmentation algorithm; querying enterprise public opinion information meeting the conditions through enterprise full names and word segmentation results; training a character name recognition model through NLP entity recognition; training a relationship name recognition model through NLP entity recognition; extracting a relationship identification model between the character name and the relationship name of the training enterprise through the NLP relationship; defining a rule base of actual control person relation words; calculating the relation weight of the actual control person; and calculating the actual control person of the enterprise through a rule algorithm. According to the invention, the NLP natural language recognition technology is utilized to mine the public opinion information related to enterprises from massive public opinion information, the association personnel and association relation related to the public opinion are analyzed, the suspected actual control person recognized by the public opinion information is analyzed, and the recognition accuracy is high.
Description
Technical Field
The invention belongs to the technical field of computer software, relates to a data processing technology, and particularly relates to a method and a system for identifying an enterprise actual control person based on enterprise public opinion information.
Background
The current difficulty of small and medium-sized micro enterprises in China can be summarized as 'two-high dilemma': high cost, high tax burden, difficult labor and difficult financing. The problem of difficult financing is particularly remarkable, the middle and small micro enterprises often face the problem of financing in the daily operation process, when financial institutions such as banks and the like provide loans for the middle and small micro enterprises, practical control staff of the enterprises are considered except for factors such as resistance risk capability, existence of collateral, operation conditions and the like of the small micro enterprises, the practical control staff of the enterprises play a decisive role in the operation direction and development of the enterprises in the middle and small micro enterprises, legal representatives registered by wagons of a plurality of enterprises are often not practical control staff of the enterprises, so if the practical control staff of the enterprises can be accurately identified, the financial institutions can eliminate the concern of the financial institutions after the practical control staff of the enterprises communicate and know, and the financial institutions and the middle and small micro enterprises are promoted to achieve the financing cooperation relationship.
At present, most scientific and technological companies in the domestic market recognize suspected actual control persons through acquiring enterprise business information and recognizing the information through the stock right structural relationship in the business, and the accuracy is not high although partial situations can be solved. The main reasons include the following two points: firstly, the acquired business information of the enterprises is not comprehensive enough, the map calculation capability is not enough, the uppermost stakeholder information cannot be traced, the calculation of the share right duty ratio is inaccurate, and secondly, the actual control person of many small, medium and micro enterprises is not reflected in the business relationship of the enterprises, and the actual control person cannot be identified through the share right relationship.
Disclosure of Invention
In order to solve the problems, the invention discloses a method and a system for identifying an enterprise actual control person based on enterprise public opinion information.
In order to achieve the above purpose, the technical scheme of the invention is as follows:
a method for identifying actual control persons of enterprises based on public opinion information of the enterprises comprises the following steps:
s10, training an enterprise name word segmentation model through an NLP word segmentation algorithm
Selecting an enterprise list from an enterprise public information base, randomly dividing a sample into two groups of a training set and a verification set, initializing and word segmentation by the training set sample through an NLP open source word segmentation interface, then manually checking and repairing a word segmentation result, training by a machine learning algorithm, and verifying model accuracy by the verification set after training is completed;
s20, inquiring enterprise public opinion information meeting conditions through enterprise full scale and word segmentation results
Inputting enterprise full names to be queried according to the enterprise name word segmentation model provided in the step S10, obtaining word segmentation results of the enterprise names, taking the word segmentation results as query key phrases, and entering the step S30 if any one or more key phrases exist in the title, the outline and the content matching public opinion information of the public opinion library, wherein the key phrases are candidate sets meeting public opinion conditions; if the public opinion information does not exist, the public opinion information does not meet the matching rule, and the actual control person of the enterprise cannot be identified;
s30, training a character name recognition model through NLP entity recognition
Randomly extracting a plurality of pieces of public opinion information from a public opinion library, marking character names appearing in the public opinion information through manual reading to obtain a standard sample library, randomly extracting part of samples from the sample library to serve as a training set, taking the rest samples as a verification set, training a character name recognition model of the training set sample through a machine learning algorithm, and verifying through the verification set after training is completed;
s31, training a relationship name recognition model through NLP entity recognition
Randomly extracting a plurality of pieces of public opinion information from a public opinion library, marking out relation words between people appearing in the public opinion information and enterprises through manual reading to obtain a standard relation name sample library, randomly extracting part of samples from the sample library to serve as a training set, and randomly extracting the rest samples from the sample library to serve as a verification set, performing relation name recognition model training on the training set samples through a machine learning algorithm, and performing verification through the verification set after training is completed;
s40, extracting a relationship identification model between the character name and the relationship name of the training enterprise through NLP relationship
Randomly extracting a plurality of pieces of public opinion information from a public opinion library, marking the relationship words between people and enterprises appearing in the public opinion information through manual reading, establishing an association relationship with the people and the enterprises to obtain a sample library, randomly extracting part of samples from the sample library to serve as a training set, and the rest of samples to serve as a verification set, performing enterprise name, enterprise personnel and relationship recognition model training between the people and the enterprises on the training set samples through a machine learning algorithm, and performing verification through the verification set after multiple rounds of training are completed;
s41, definition of actual control person relation word rule base
After unstructured public opinion information is identified through NLP, defining a set of actual controller strong matching recognition relation word stock according to expert rules, wherein the relation word stock comprises keywords related to the relationship of the actual controller, and the keywords comprise two types, namely a strong rule recognition keyword and a weak rule recognition keyword;
s50, calculating the relation weight of the actual control person
Performing relationship word name matching judgment through the relationship word set between the enterprise and the personnel identified in the step S40 and the relationship word library of the actual control person defined in the step S41; if the identified enterprise keywords in the step S40 hit any one of the defined strong rule keywords in the actual control person relationship word stock, the relationship coefficient between the corresponding enterprise and the person is increased by a strong relationship coefficient value, and if a weak rule is hit, the relationship coefficient between the corresponding enterprise and the person is increased by a weak relationship value;
s60, calculating the actual control person of the enterprise through a rule algorithm
After the coefficient calculation is completed for all enterprise relations through the step S50, the relation coefficient of the actual control person between the target enterprise and all persons with relations is counted, the relation person with the highest relation coefficient and the relation coefficient required to be larger than 1 is taken as the actual control person, and if the relation coefficient is smaller than 1, the actual control person relation which is not recognized by public opinion of the enterprise is indicated.
Further, in the step S10, the enterprise list is selected based on the following rules: randomly lottery enterprises with enterprise name length meeting the requirement, and extracting a plurality of enterprises with each length.
Further, the steps S31 and S40 follow the public opinion information in the step S30.
Further, in the step S41, part of the keywords are selected from the keyword library identified in the step S40 or defined according to expert service experience.
Further, in the step S50, when the same relation word hits multiple times, the relation coefficient is increased only once.
Further, in the step S60, when there are a plurality of relationship coefficients exceeding 1, the enterprise corresponding to the highest value relationship is selected.
A system for identifying actual control persons of an enterprise based on public opinion information of the enterprise, comprising: the system comprises an enterprise name word segmentation model training module, an enterprise public opinion information query module, a character name recognition model training module, a relationship recognition model training module, an actual controller relationship word rule base definition module, an actual controller relationship weight calculation module and an enterprise actual controller calculation module;
the enterprise name word segmentation model training module carries out initialization word segmentation on a training set sample through an NLP open source word segmentation interface, carries out artificial check and repair on a word segmentation result, carries out training through a machine learning algorithm, and carries out model accuracy verification through a verification set after the training is completed;
the enterprise public opinion information query module is used for obtaining word segmentation results according to enterprise names based on the model obtained by the enterprise name word segmentation model training module, and querying in a public opinion library according to the word segmentation results;
the character name recognition model training module trains a character name recognition model through a machine learning algorithm based on a training set and a verification set selected by a character name sample library;
the relation name recognition model training module trains a relation name recognition model through a machine learning algorithm based on a training set and a verification set selected by a relation name sample library between personnel and enterprises;
the relation recognition model training module trains a relation recognition model through a machine learning algorithm based on a relation sample library between character names and relation names and selected training sets and verification sets;
the actual controller relation word rule base definition module defines a set of actual controller strong matching recognition relation word base, and actual controller relation keywords comprise strong rule recognition keywords and weak rule recognition keywords;
the actual controller relation weight calculation module is used for carrying out relation matching judgment on the relation word set between the enterprise and the person identified by the relation identification model and the relation word library of the actual controller defined by the actual controller relation word rule library definition module to obtain the relation coefficient between the enterprise and the person;
the enterprise actual control person calculation module counts actual control person relation coefficients between the inquired target enterprise and all persons with relations, and takes the relation person with the highest relation coefficient and the relation coefficient which is required to be larger than 1 as the actual control person.
The beneficial effects of the invention are as follows:
according to the invention, the NLP natural language recognition technology is utilized to mine the public opinion information related to enterprises from massive public opinion information, the association personnel and association relation related to the public opinion are analyzed, the suspected actual control person recognized by the public opinion information is analyzed, and the recognition accuracy is high.
Drawings
Fig. 1 is a flowchart of a method for identifying an actual control person of an enterprise based on public opinion information of the enterprise.
Detailed Description
The technical scheme provided by the present invention will be described in detail with reference to the following specific examples, and it should be understood that the following specific examples are only for illustrating the present invention and are not intended to limit the scope of the present invention. Additionally, the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer executable instructions, and although a logical order is illustrated in the flowcharts, in some cases the steps illustrated or described may be performed in an order other than that herein.
The invention provides a method for identifying an enterprise actual control person based on enterprise public opinion information, which is shown in a figure 1 and comprises the following steps:
s10, training an enterprise name word segmentation model through an NLP word segmentation algorithm
And selecting an enterprise list from an enterprise public information base (a selection rule is that enterprises with enterprise name lengths ranging from 10 bits to 20 bits are randomly selected, 1000 enterprises with each length are extracted, and the total number is 11000 enterprises), randomly dividing a sample into two groups of a training set (7000) and a verification set (4000), carrying out initialization word segmentation on the training set sample through an NLP open source word segmentation interface, then carrying out artificial check and repair on a word segmentation result, carrying out training through a machine learning algorithm, and carrying out model accuracy verification through the verification set after training is finished, wherein the accuracy of the word segmentation result reaches more than 80%.
S20, inquiring enterprise public opinion information meeting conditions through enterprise full scale and word segmentation results
And (3) inputting enterprise full names to be queried according to the enterprise name word segmentation model provided in the step S10, obtaining word segmentation results of the enterprise names, taking the word segmentation results as query key phrases, and matching whether any one or more key phrases exist in the key phrases in the public opinion information through the title, the outline and the content of the public opinion in a public opinion library, if so, the key phrases are not exist, and if so, the key phrases do not meet the matching rule.
And counting all public opinion information of the enterprise to be queried, if the public opinion information exists, entering into the step S30, and if the public opinion information does not exist in the enterprise, indicating that the public opinion information cannot identify the actual control person of the enterprise.
S30, training a character name recognition model through NLP entity recognition
10000 pieces of public opinion information are randomly extracted from a public opinion library, character names appearing in the public opinion information are marked through manual reading, a standard sample library is obtained, 60% of the sample library is randomly extracted to serve as a training set, 40% of the sample library serves as a verification set, character name recognition model training is conducted on 60% of training set samples through machine learning algorithms such as Lattice LSTM, and after training is completed, verification is conducted through 40% of the verification set, so that character name recognition accuracy reaches 90%.
S31, training a relationship name recognition model through NLP entity recognition
10000 public opinion information (in order to ensure that enterprises can have association relations in a selected set, the public opinion information in the step S30 is used here) is randomly extracted from the public opinion library, the relation words between people appearing in the public opinion information and the enterprises are marked through manual reading, a standard relation name sample library is obtained, 60% of the standard relation name sample library is randomly extracted from the sample library to serve as a training set, 40% of the standard relation name sample library serves as a verification set, 60% of training set samples are subjected to relation name recognition model training through machine learning algorithms such as Lattice LSTM, and after training is completed, verification is performed through 40% of the verification set, and the relation name recognition accuracy is about 70%.
S40, extracting a relationship identification model between the character name and the relationship name of the training enterprise through NLP relationship
10000 public opinion information (in order to ensure that enterprises can have association relations in a selected set, public opinion information in the steps S30 and S31 is adopted here) is randomly extracted from a public opinion library, relational words between people and enterprises which appear in the public opinion information are marked through manual reading (for example, zhao XX is used as CEO of Jiangsu XXXX company, then CEO is extracted as relational words, the relational parties are Zhao XX and Jiangsu XX company), association relations between people and enterprises are established, a sample library is obtained, 60% of the sample library is randomly extracted as a training set, 40% of the sample library is used as a verification set, and the relationship recognition model training among enterprise names, enterprise personnel and people and enterprises is carried out through machine learning algorithms such as Lattice LSTM, and after the multi-round training is completed, verification is carried out through 40% of the verification set, so that the accuracy of relationship name recognition is about 60%.
S41, definition of actual control person relation word rule base
Considering that the accuracy of the relationship identified by the model is not necessarily accurate, after unstructured public opinion information is identified by NLP, defining a set of actual controller strong matching identification relationship word stock according to expert rules, wherein the relationship word stock definitely gives out some keywords with relatively definite relationships of the actual controller, the keyword stock is defined by expert rules, the keywords are classified into two types, namely, strong rule identification keywords and weak rule identification keywords, part of the keywords can be selected from the keyword stock identified in the step S40 or defined according to expert business experience, and part of the keywords are listed as references, as shown in Table 1:
sequence number | Relational word name | Degree of relationship | Guan Jici weight |
1 | Actually controlling person | Strong strength | 1.0 |
2 | Controlling stockholder | Strong strength | 1.0 |
3 | Overrule right for ticket | Strong strength | 1.0 |
… | … | … | … |
N | Representative of | Weak and weak | 0.2 |
TABLE 1
S50, calculating the relation weight of the actual control person
And (3) carrying out matching judgment on the names of the relational words through the relational word library of the actual control people defined in the step S41 and the relational word set between the enterprises and the personnel identified in the step S40, if the enterprise keywords identified in the step S40 hit strong rules in the relational word library of the actual control people defined, determining keywords, if any strong rule keyword is hit, then the relational coefficient between the corresponding enterprises and personnel is +1, if a weak rule is hit, then the relational coefficient between the corresponding enterprises and personnel is +0.2, and the occurrence times of the same relational words are not included in the calculation logic range.
S60, calculating the actual control person of the enterprise through a rule algorithm
After the calculation of the coefficients is completed for all the enterprise relationships, the actual control person relationship coefficients between the target enterprise and all the persons with the relationships are counted, the relationship person with the highest relationship coefficient and the relationship coefficient greater than 1 is taken as the actual control person (the enterprise corresponding to the highest relationship if a plurality of relationship coefficients exceed 1 is taken), and if the relationship coefficient is less than 1, the relationship is represented that the enterprise has no public opinion.
In order to realize the method for identifying the actual enterprise controllers based on the enterprise public opinion information, the invention also provides a system for identifying the actual enterprise controllers based on the enterprise public opinion information, which comprises the following steps: the system comprises an enterprise name word segmentation model training module, an enterprise public opinion information query module, a character name recognition model training module, a relationship recognition model training module, an actual controller relationship word rule base definition module, an actual controller relationship weight calculation module and an enterprise actual controller calculation module.
The enterprise name word segmentation model training module carries out initialization word segmentation on a training set sample through an NLP open source word segmentation interface, carries out artificial check and repair on a word segmentation result, carries out training through a machine learning algorithm, carries out model accuracy verification through a verification set after the training is finished, and specifically realizes the content of the step S10; the enterprise public opinion information query module obtains word segmentation results according to enterprise full names based on the model obtained by the enterprise name word segmentation model training module, queries in a public opinion library according to the word segmentation results, and particularly realizes the content of the step S20; the character name recognition model training module trains a character name recognition model through a machine learning algorithm based on a training set and a verification set selected by a character name sample library, and the content of the step S30 is specifically realized; the relation name recognition model training module trains a relation name recognition model through a machine learning algorithm based on a training set and a verification set selected by a relation name sample library between personnel and enterprises, and the content of the step S31 is specifically realized; the relation recognition model training module trains a relation recognition model through a machine learning algorithm based on a relation sample library between character names and relation names and a selected training set and a verification set, and specifically realizes the content of the step S40; the actual controller relation word rule base definition module defines a set of actual controller strong matching recognition relation word base, and actual controller relation keywords comprise strong rule recognition keywords and weak rule recognition keywords, and the content of the step S41 is specifically realized; the actual controller relation weight calculation module is used for carrying out relation matching judgment on the relation word set between the enterprise and the person identified by the relation identification model and the relation word library of the actual controller defined by the actual controller relation word rule library definition module to obtain the relation coefficient between the enterprise and the person, and the content of the step S50 is realized specifically; the enterprise actual control person calculation module calculates actual control person relation coefficients between the inquired target enterprise and all persons with relations, and takes the relation person with the highest relation coefficient and the relation coefficient more than 1 as the actual control person, thereby specifically realizing the content of the step S60.
It should be noted that the foregoing merely illustrates the technical idea of the present invention and is not intended to limit the scope of the present invention, and that a person skilled in the art may make several improvements and modifications without departing from the principles of the present invention, which fall within the scope of the claims of the present invention.
Claims (7)
1. The method for identifying the actual control person of the enterprise based on the public opinion information of the enterprise is characterized by comprising the following steps:
s10, training an enterprise name word segmentation model through an NLP word segmentation algorithm
Selecting an enterprise list from an enterprise public information base, randomly dividing a sample into two groups of a training set and a verification set, initializing and word segmentation by the training set sample through an NLP open source word segmentation interface, then manually checking and repairing a word segmentation result, training by a machine learning algorithm, and verifying model accuracy by the verification set after training is completed;
s20, inquiring enterprise public opinion information meeting conditions through enterprise full scale and word segmentation results
Inputting enterprise full names to be queried according to the enterprise name word segmentation model provided in the step S10, obtaining word segmentation results of the enterprise names, taking the word segmentation results as query key phrases, and entering the step S30 if any one or more key phrases exist in the title, the outline and the content matching public opinion information of the public opinion library, wherein the key phrases are candidate sets meeting public opinion conditions; if the public opinion information does not exist, the public opinion information does not meet the matching rule, and the actual control person of the enterprise cannot be identified;
s30, training a character name recognition model through NLP entity recognition
Randomly extracting a plurality of pieces of public opinion information from a public opinion library, marking character names appearing in the public opinion information through manual reading to obtain a standard sample library, randomly extracting part of samples from the sample library to serve as a training set, taking the rest samples as a verification set, training a character name recognition model of the training set sample through a machine learning algorithm, and verifying through the verification set after training is completed;
s31, training a relationship name recognition model through NLP entity recognition
Randomly extracting a plurality of pieces of public opinion information from a public opinion library, marking out relation words between people appearing in the public opinion information and enterprises through manual reading to obtain a standard relation name sample library, randomly extracting part of samples from the sample library to serve as a training set, and randomly extracting the rest samples from the sample library to serve as a verification set, performing relation name recognition model training on the training set samples through a machine learning algorithm, and performing verification through the verification set after training is completed;
s40, extracting a relationship identification model between the character name and the relationship name of the training enterprise through NLP relationship
Randomly extracting a plurality of pieces of public opinion information from a public opinion library, marking the relationship words between people and enterprises appearing in the public opinion information through manual reading, establishing an association relationship with the people and the enterprises to obtain a sample library, randomly extracting part of samples from the sample library to serve as a training set, and the rest of samples to serve as a verification set, performing enterprise name, enterprise personnel and relationship recognition model training between the people and the enterprises on the training set samples through a machine learning algorithm, and performing verification through the verification set after multiple rounds of training are completed;
s41, definition of actual control person relation word rule base
After unstructured public opinion information is identified through NLP, defining a set of actual controller strong matching recognition relation word stock according to expert rules, wherein the relation word stock comprises keywords related to the relationship of the actual controller, and the keywords comprise two types, namely a strong rule recognition keyword and a weak rule recognition keyword;
s50, calculating the relation weight of the actual control person
Performing relationship word name matching judgment through the relationship word set between the enterprise and the personnel identified in the step S40 and the relationship word library of the actual control person defined in the step S41; if the identified relation word in the step S40 hits any one of the defined strong rule keywords in the relation word library of the actual control person, the relation coefficient between the corresponding enterprise and the person is increased by a strong relation coefficient value, and if a weak rule is hit, the relation coefficient between the corresponding enterprise and the person is increased by a weak relation value;
s60, calculating the actual control person of the enterprise through a rule algorithm
After the coefficient calculation is completed for all enterprise relations through the step S50, the relation coefficient of the actual control person between the target enterprise and all persons with relations is counted, the relation person with the highest relation coefficient and the relation coefficient required to be larger than 1 is taken as the actual control person, and if the relation coefficient is smaller than 1, the actual control person relation which is not recognized by public opinion of the enterprise is indicated.
2. The method for identifying actual control persons of enterprises based on the public opinion information of enterprises according to claim 1, wherein the step S10 selects the enterprise list based on the following rules: randomly lottery enterprises with enterprise name length meeting the requirement, and extracting a plurality of enterprises with each length.
3. The method for identifying actual control persons of an enterprise based on public opinion information of an enterprise according to claim 1, wherein the step S31 and the step S40 follow the public opinion information in the step S30.
4. The method for identifying actual control persons of an enterprise based on public opinion information of the enterprise according to claim 1, wherein the partial keywords are selected from the keyword library identified in step S40 or defined according to expert service experience in step S41.
5. The method for identifying actual control persons of an enterprise based on public opinion information of the enterprise according to claim 1, wherein the relationship coefficient is increased only once when the same relationship word hits a plurality of times in step S50.
6. The method for identifying actual control persons of enterprises based on public opinion information of enterprises according to claim 1, wherein in step S60, when there are a plurality of relationship coefficients exceeding 1, the enterprise corresponding to the highest value relationship is selected.
7. A system for identifying actual control persons of an enterprise based on public opinion information of the enterprise, comprising: the system comprises an enterprise name word segmentation model training module, an enterprise public opinion information query module, a character name recognition model training module, a relationship recognition model training module, an actual controller relationship word rule base definition module, an actual controller relationship weight calculation module and an enterprise actual controller calculation module;
the enterprise name word segmentation model training module carries out initialization word segmentation on a training set sample through an NLP open source word segmentation interface, carries out artificial check and repair on a word segmentation result, carries out training through a machine learning algorithm, and carries out model accuracy verification through a verification set after the training is completed;
the enterprise public opinion information query module is used for obtaining word segmentation results according to enterprise names based on the model obtained by the enterprise name word segmentation model training module, and querying in a public opinion library according to the word segmentation results;
the character name recognition model training module trains a character name recognition model through a machine learning algorithm based on a training set and a verification set selected by a character name sample library;
the relation name recognition model training module trains a relation name recognition model through a machine learning algorithm based on a training set and a verification set selected by a relation name sample library between personnel and enterprises;
the relation recognition model training module trains a relation recognition model through a machine learning algorithm based on a relation sample library between character names and relation names and selected training sets and verification sets;
the actual controller relation word rule base definition module defines a set of actual controller strong matching recognition relation word base, and actual controller relation keywords comprise strong rule recognition keywords and weak rule recognition keywords;
the actual controller relation weight calculation module is used for carrying out relation matching judgment on the relation word set between the enterprise and the person identified by the relation identification model and the relation word library of the actual controller defined by the actual controller relation word rule library definition module to obtain the relation coefficient between the enterprise and the person;
the enterprise actual control person calculation module counts actual control person relation coefficients between the inquired target enterprise and all persons with relations, and takes the relation person with the highest relation coefficient and the relation coefficient which is required to be larger than 1 as the actual control person.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210106055.5A CN114611515B (en) | 2022-01-28 | 2022-01-28 | Method and system for identifying enterprise actual control person based on enterprise public opinion information |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210106055.5A CN114611515B (en) | 2022-01-28 | 2022-01-28 | Method and system for identifying enterprise actual control person based on enterprise public opinion information |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114611515A CN114611515A (en) | 2022-06-10 |
CN114611515B true CN114611515B (en) | 2023-12-12 |
Family
ID=81859766
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210106055.5A Active CN114611515B (en) | 2022-01-28 | 2022-01-28 | Method and system for identifying enterprise actual control person based on enterprise public opinion information |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114611515B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116227496B (en) * | 2023-05-06 | 2023-07-14 | 国网智能电网研究院有限公司 | Deep learning-based electric public opinion entity relation extraction method and system |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108763507A (en) * | 2018-05-30 | 2018-11-06 | 北京百度网讯科技有限公司 | Enterprise's incidence relation method for digging and device |
CN110597870A (en) * | 2019-08-05 | 2019-12-20 | 长春市万易科技有限公司 | Enterprise relation mining method |
CN110807697A (en) * | 2019-10-31 | 2020-02-18 | 支付宝(杭州)信息技术有限公司 | Method, device and equipment for determining actual controller of stock company |
CN110909986A (en) * | 2019-11-04 | 2020-03-24 | 苏宁金融科技(南京)有限公司 | Suspected actual controller risk identification method and system based on knowledge graph |
CN110990525A (en) * | 2019-11-15 | 2020-04-10 | 华融融通(北京)科技有限公司 | Natural language processing-based public opinion information extraction and knowledge base generation method |
CN111325355A (en) * | 2020-03-19 | 2020-06-23 | 中国建设银行股份有限公司 | Method and device for determining actual control persons of enterprises, computer equipment and medium |
CN112199588A (en) * | 2020-09-30 | 2021-01-08 | 深圳壹账通智能科技有限公司 | Public opinion text screening method and device |
CN112364178A (en) * | 2020-11-08 | 2021-02-12 | 杭州有数金融信息服务有限公司 | Method for identifying invisible real control people of enterprise based on enterprise associated knowledge graph |
WO2021217843A1 (en) * | 2020-04-29 | 2021-11-04 | 平安科技(深圳)有限公司 | Enterprise public opinion analysis method and apparatus, and electronic device and medium |
-
2022
- 2022-01-28 CN CN202210106055.5A patent/CN114611515B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108763507A (en) * | 2018-05-30 | 2018-11-06 | 北京百度网讯科技有限公司 | Enterprise's incidence relation method for digging and device |
CN110597870A (en) * | 2019-08-05 | 2019-12-20 | 长春市万易科技有限公司 | Enterprise relation mining method |
CN110807697A (en) * | 2019-10-31 | 2020-02-18 | 支付宝(杭州)信息技术有限公司 | Method, device and equipment for determining actual controller of stock company |
CN110909986A (en) * | 2019-11-04 | 2020-03-24 | 苏宁金融科技(南京)有限公司 | Suspected actual controller risk identification method and system based on knowledge graph |
CN110990525A (en) * | 2019-11-15 | 2020-04-10 | 华融融通(北京)科技有限公司 | Natural language processing-based public opinion information extraction and knowledge base generation method |
CN111325355A (en) * | 2020-03-19 | 2020-06-23 | 中国建设银行股份有限公司 | Method and device for determining actual control persons of enterprises, computer equipment and medium |
WO2021217843A1 (en) * | 2020-04-29 | 2021-11-04 | 平安科技(深圳)有限公司 | Enterprise public opinion analysis method and apparatus, and electronic device and medium |
CN112199588A (en) * | 2020-09-30 | 2021-01-08 | 深圳壹账通智能科技有限公司 | Public opinion text screening method and device |
CN112364178A (en) * | 2020-11-08 | 2021-02-12 | 杭州有数金融信息服务有限公司 | Method for identifying invisible real control people of enterprise based on enterprise associated knowledge graph |
Also Published As
Publication number | Publication date |
---|---|
CN114611515A (en) | 2022-06-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110826320B (en) | Sensitive data discovery method and system based on text recognition | |
CN106021545B (en) | Method for vehicle remote diagnosis and spare part retrieval | |
CN111191051B (en) | Method and system for constructing emergency knowledge map based on Chinese word segmentation technology | |
CN114741482B (en) | Bid-winning information matching method, system, equipment and medium based on bid-winning information | |
CN109033087B (en) | Method for calculating text semantic distance, deduplication method, clustering method and device | |
CN114611515B (en) | Method and system for identifying enterprise actual control person based on enterprise public opinion information | |
CN116226103A (en) | Method for detecting government data quality based on FPGrow algorithm | |
CN109617864B (en) | Website identification method and website identification system | |
CN112926577B (en) | Medical bill image structuring method and device and computer readable medium | |
CN107480126B (en) | Intelligent identification method for engineering material category | |
CN113569048A (en) | Method and system for automatically dividing affiliated industries based on enterprise operation range | |
CN117648581A (en) | Enterprise similarity evaluation method, device, terminal and medium | |
CN110083815B (en) | Synonymous variable identification method and system | |
CN112084773A (en) | Power grid power failure address matching method based on word bank bidirectional maximum matching method | |
CN113569005B (en) | Large-scale data characteristic intelligent extraction method based on data content | |
CN103984756B (en) | Semi-supervised probabilistic latent semantic analysis based software change log classification method | |
CN109460720A (en) | Ballot paper recognition methods based on convolutional neural networks | |
CN115186138A (en) | Comparison method and terminal for power distribution network data | |
CN114547294A (en) | Rumor detection method and system based on comprehensive information of propagation process | |
CN110414819B (en) | Work order scoring method | |
CN109299456B (en) | Geographical name recognition method | |
CN112966901A (en) | Lineage data quality analysis and verification method for inspection business collaborative flow | |
CN112102069A (en) | Personal property mortgage loan information input analysis system | |
CN110598973A (en) | IAP-based risk evaluation method for authentication process of green furniture product | |
CN105138544B (en) | A kind of searching method for remolding logical deduction chain |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |