CN106780001A - A kind of invoice writes out falsely enterprise supervision recognition methods and system - Google Patents
A kind of invoice writes out falsely enterprise supervision recognition methods and system Download PDFInfo
- Publication number
- CN106780001A CN106780001A CN201611220015.4A CN201611220015A CN106780001A CN 106780001 A CN106780001 A CN 106780001A CN 201611220015 A CN201611220015 A CN 201611220015A CN 106780001 A CN106780001 A CN 106780001A
- Authority
- CN
- China
- Prior art keywords
- enterprise
- invoice
- business
- type
- month
- 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
Classifications
-
- 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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/12—Accounting
- G06Q40/123—Tax preparation or submission
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Engineering & Computer Science (AREA)
- Marketing (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Enterprise supervision recognition methods is write out falsely this application discloses a kind of invoice, including：The characteristic information of Target Enterprise is extracted, target signature information is obtained；Target signature information is input into the type of business forecast model being pre-created；Type of business forecast model is, using default machine learning algorithm and sample set, to carry out the discrimination model obtained after corresponding model training；Every part of sample data in sample set includes the type of business of the type of business characteristic information of clear and definite enterprise and the enterprise；Wherein, the type of business of enterprise is voiding invoice enterprise or legal enterprise of paying taxes；Obtain type of business forecast model to be predicted the outcome according to the type of business that target signature information export, whether the type of business for determining Target Enterprise to predict the outcome according to the type of business is voiding invoice enterprise.The application improves the monitoring recognition effect and recognition efficiency that enterprise is write out falsely to invoice.In addition, the application further correspondingly discloses a kind of invoice writes out falsely enterprise supervision identifying system.
Description
Technical field
The present invention relates to enterprise's regulation technique field, more particularly to a kind of invoice writes out falsely enterprise supervision recognition methods and is
System.
Background technology
As tax system examines and approves continuous development of the reform, while simplification does tax program, excites the market vitality,
The tax jurisdiction risk of portion link also be increased therewith, and particularly lawless person is write out falsely by setting up ghost commerce and trade enterprise
Invoice phenomenon come back in indivedual areas, very disruptive normal tax revenue and economic order.
It is still to rely on tax analysis in the real work of the existing tax risk prevention and control that enterprise is write out falsely for invoice
The business experience of personnel, cause to pay taxes defer to risk model subjective composition it is heavier, accuracy is not strong enough.And due to people's work point
Analysis it is less efficient the reason for, cause be difficult comprehensively identify in the presence of write out falsely invoice enterprise.
In sum as can be seen that how to be lifted to the monitoring recognition effect and recognition efficiency of invoice voiding enterprise is mesh
It is preceding to also have problem to be solved.
The content of the invention
In view of this, enterprise supervision recognition methods and system, lifting are write out falsely it is an object of the invention to provide a kind of invoice
The monitoring recognition effect and recognition efficiency of enterprise are write out falsely to invoice.Its concrete scheme is as follows：
A kind of invoice writes out falsely enterprise supervision recognition methods, including：
The characteristic information of Target Enterprise is extracted, target signature information is obtained；
The target signature information is input into the type of business forecast model being pre-created；The type of business predicts mould
Type is, using default machine learning algorithm and sample set, to carry out the discrimination model obtained after corresponding model training；The sample
Every part of sample data of this concentration includes the type of business of the type of business characteristic information of clear and definite enterprise and the enterprise；
Wherein, the type of business of enterprise is voiding invoice enterprise or legal enterprise of paying taxes；
The type of business forecast model is obtained to be predicted the outcome according to the type of business that the target signature information is exported, with
Predicted the outcome according to the type of business and determine whether the type of business of the Target Enterprise is to write out falsely invoice enterprise.
Optionally, the process of the characteristic information for extracting Target Enterprise, including：
The first kind characteristic information of the Target Enterprise is extracted, Enterprise Age, the enterprises registration of the Target Enterprise is obtained
Registered number of the enterprise, enterprises registration address and the Enterprise Law that registered number of the enterprise corresponding to address, business entity possess
The age of improper pay taxes status information and the business entity of otherness, business entity between the personal household register ground of people.
Optionally, the establishment process of the type of business forecast model, including：
First group of sample data and second group of sample data are obtained, the sample set is obtained；Wherein, first group of sample
Data include N1The corresponding first kind characteristic information of enterprise of family and the type of business, second group of sample data include M1
The corresponding first kind characteristic information of enterprise of family and the type of business, N1And M1It is positive integer, also, the N1Appoint in enterprise of family
The type of business of one enterprise is voiding invoice enterprise, the M1The type of business of any enterprise is legal paying taxes in enterprise of family
Enterprise；
According to default first division proportion, the sample set is divided into two parts, obtains corresponding first training set
With the first test set；
Using the machine learning algorithm, first training set and first test set, corresponding model instruction is carried out
Practice and model measurement, obtain the type of business forecast model.
Optionally, the process of the characteristic information for extracting Target Enterprise, including：
The Equations of The Second Kind characteristic information of the Target Enterprise is extracted, the Enterprise Age of the Target Enterprise is obtained, is continuously issued
The quantity information and time interval information of invoice, evening this month amount of money of making out an invoice account for this month to date and make out an invoice ratio, this month of the amount of money
Evening number of making out an invoice accounts for this month to date ratio of number, this month amount of money of making out an invoice of making out an invoice and is accounted for more than the invoice amount of preset cost threshold value
This month to date ratio of the amount of money, this month amount of money of making out an invoice of making out an invoice accounts for this month and tires out more than the invoice number of the default amount of money threshold value of making out an invoice
Meter make out an invoice the ratio of number, this month to the number that same purchaser issues invoice account for this month to date make out an invoice number ratio and
This month, accounts for this month to date and makes out an invoice the ratio of the amount of money to the invoice amount that same purchaser issues invoice.
Optionally, the process of the characteristic information for extracting Target Enterprise, including：
The 3rd group of sample data and the 4th group of sample data are obtained, the sample set is obtained；Wherein, the 3rd group of sample
Data include N2The corresponding Equations of The Second Kind characteristic information of enterprise of family and the type of business, the 4th group of sample data include M2
The corresponding Equations of The Second Kind characteristic information of enterprise of family and the type of business, N2And M2It is positive integer, also, the N2Appoint in enterprise of family
The type of business of one enterprise is voiding invoice enterprise, the M2The type of business of any enterprise is legal paying taxes in enterprise of family
Enterprise；
According to default second division proportion, the sample set is divided into two parts, obtains corresponding second training set
With the second test set；
Using the machine learning algorithm, second training set and second test set, corresponding model instruction is carried out
Practice and model measurement, obtain the type of business forecast model.
Optionally, the machine learning algorithm is decision Tree algorithms.
Enterprise supervision identifying system is write out falsely the invention also discloses a kind of invoice, including：
Information extraction modules, the characteristic information for extracting Target Enterprise, obtain target signature information；
MIM message input module, for the target signature information to be input into the type of business forecast model being pre-created；
The type of business forecast model is, using default machine learning algorithm and sample set, to be obtained after carrying out corresponding model training
Discrimination model；Every part of sample data in the sample set include the type of business the characteristic information of clear and definite enterprise and
The type of business of the enterprise；Wherein, the type of business of enterprise is voiding invoice enterprise or legal enterprise of paying taxes；
As a result acquisition module, for obtaining the enterprise that the type of business forecast model is exported according to the target signature information
Industry type prediction result, whether the type of business that the Target Enterprise is determined to predict the outcome according to the type of business is voiding
Invoice enterprise.
Optionally, described information extraction module, including：
First kind feature extraction submodule, the first kind characteristic information for extracting the Target Enterprise, obtains the mesh
Mark the Enterprise Age of enterprise, the registered enterprise that the registered number of the enterprise corresponding to enterprises registration address, business entity possess
The improper state of paying taxes of otherness, business entity between the personal household register ground of quantity, enterprises registration address and business entity
Information and the age of business entity.
Optionally, described information extraction module, including：
Equations of The Second Kind feature extraction submodule, the Equations of The Second Kind characteristic information for extracting the Target Enterprise, obtains the mesh
Mark the Enterprise Age of enterprise, the continuous quantity information for issuing invoice and time interval information, evening this month amount of money of making out an invoice and account for this
The make out an invoice ratio of the amount of money, evening this month of month to date makes out an invoice that to account for this month to date ratio of number, this month amount of money of making out an invoice of making out an invoice big for number
This month to date ratio of the amount of money, this month amount of money of making out an invoice of making out an invoice is accounted in the invoice amount of preset cost threshold value default to be made out an invoice more than described
The invoice number of amount of money threshold value accounts for the make out an invoice ratio of number, this month of this month to date and accounts for this to the number that same purchaser issues invoice
Month to date makes out an invoice the ratio of number and this month accounts for this month to date and makes out an invoice the amount of money to the invoice amount that same purchaser issues invoice
Ratio.
Optionally, the machine learning algorithm is decision Tree algorithms.
In the present invention, invoice writes out falsely enterprise supervision recognition methods, including：The characteristic information of Target Enterprise is extracted, mesh is obtained
Mark characteristic information；Target signature information is input into the type of business forecast model being pre-created；Type of business forecast model is
Using default machine learning algorithm and sample set, the discrimination model obtained after corresponding model training is carried out；In sample set
Every part of sample data includes the type of business of the type of business characteristic information of clear and definite enterprise and the enterprise；Wherein, look forward to
The type of business of industry is voiding invoice enterprise or legal enterprise of paying taxes；Type of business forecast model is obtained according to target signature information
The type of business of output predicts the outcome, with according to the type of business predict the outcome determine Target Enterprise the type of business whether be voiding
Invoice enterprise.
It can be seen that, the present invention advances with machine learning algorithm and the sample comprising enterprise characteristic information and the type of business
Collection, creates the forecast model for being predicted to the type of business, when needing to be monitored identification to Target Enterprise, can be right
The Target Enterprise carries out corresponding feature extraction, and the target signature information for extracting then is delivered to above-mentioned forecast model, from
And the type of business of above-mentioned Target Enterprise can be obtained, actually or above-mentioned Target Enterprise voiding invoice enterprise is predicted with this
Legal enterprise of paying taxes.Therefore, the present invention is to be monitored identification to enterprise based on machine learning algorithm, to determine enterprise
Whether industry be to write out falsely invoice enterprise, it is to avoid the various defects that manual analysis is brought, and thus improves and writes out falsely enterprise to invoice
Monitoring recognition effect and recognition efficiency.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
The accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are only this
Inventive embodiment, for those of ordinary skill in the art, on the premise of not paying creative work, can also basis
The accompanying drawing of offer obtains other accompanying drawings.
Fig. 1 is that a kind of invoice writes out falsely enterprise supervision recognition methods flow chart disclosed in the embodiment of the present invention；
Fig. 2 is that a kind of specific invoice writes out falsely enterprise supervision recognition methods flow chart disclosed in the embodiment of the present invention；
Fig. 3 is that a kind of specific invoice writes out falsely enterprise supervision recognition methods flow chart disclosed in the embodiment of the present invention；
Fig. 4 is that a kind of invoice writes out falsely enterprise supervision identifying system structural representation disclosed in the embodiment of the present invention.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than whole embodiments.It is based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made
Embodiment, belongs to the scope of protection of the invention.
Enterprise supervision recognition methods, shown in Figure 1, the method bag are write out falsely the embodiment of the invention discloses a kind of invoice
Include：
Step S11：The characteristic information of Target Enterprise is extracted, target signature information is obtained.
It should be noted that in the present embodiment the characteristic information of enterprise be specifically including but not limited to enterprises registration information and/
Or enterprise's invoice issuing behavior characteristic information.
Step S12：Target signature information is input into the type of business forecast model being pre-created；Wherein, the type of business
Forecast model is, using default machine learning algorithm and sample set, to carry out the discrimination model obtained after corresponding model training；
Every part of sample data in above-mentioned sample set includes the enterprise of the type of business characteristic information of clear and definite enterprise and the enterprise
Industry type；In the present embodiment, the type of business of enterprise is voiding invoice enterprise or legal enterprise of paying taxes.
In the present embodiment, above-mentioned machine learning algorithm can be disclosed various machine in normal service study in the prior art
Algorithm.Wherein, the present embodiment is preferentially using decision Tree algorithms as above-mentioned machine learning algorithm.
Step S13：Type of business forecast model is obtained to be predicted the outcome according to the type of business that target signature information is exported, with
Predicted the outcome according to the type of business and determine whether the type of business of Target Enterprise is to write out falsely invoice enterprise.
It can be seen that, the present invention advances with machine learning algorithm and the sample comprising enterprise characteristic information and the type of business
Collection, creates the forecast model for being predicted to the type of business, when needing to be monitored identification to Target Enterprise, can be right
The Target Enterprise carries out corresponding feature extraction, and the target signature information for extracting then is delivered to above-mentioned forecast model, from
And the type of business of above-mentioned Target Enterprise can be obtained, actually or above-mentioned Target Enterprise voiding invoice enterprise is predicted with this
Legal enterprise of paying taxes.Therefore, the present invention is to be monitored identification to enterprise based on machine learning algorithm, to determine enterprise
Whether industry be to write out falsely invoice enterprise, it is to avoid the various defects that manual analysis is brought, and thus improves and writes out falsely enterprise to invoice
Monitoring recognition effect and recognition efficiency.
It is shown in Figure 2, enterprise supervision recognition methods is write out falsely the embodiment of the invention discloses a kind of specific invoice, including
Following steps：
Step S21：The first kind characteristic information of Target Enterprise is extracted, Enterprise Age, the enterprises registration of Target Enterprise is obtained
Registered number of the enterprise, enterprises registration address and the Enterprise Law that registered number of the enterprise corresponding to address, business entity possess
The age of improper pay taxes status information and the business entity of otherness, business entity between the personal household register ground of people.
That is, in the present embodiment, the first kind characteristic information of Target Enterprise can be based on, whether the Target Enterprise is judged
To write out falsely invoice enterprise.Wherein, the Enterprise Age of the first kind characteristic information of enterprise including enterprise, corresponding to enterprises registration address
Registered number of the enterprise, the personal family of registered number of the enterprise, enterprises registration address and business entity that possesses of business entity
The age of improper pay taxes status information and the business entity of otherness, business entity between nationality ground.
Invoice behavior is write out falsely due to existing now with many newly-set-up commerce and trade enterprises, so the embodiment of the present invention can be preferential
Identification is monitored to newly-set-up commerce and trade enterprise.So-called newly-set-up commerce and trade enterprise refers to wholesale less than or equal to 2 years Enterprise Age
Retailer.
In the present embodiment, between the Enterprise Age of above-mentioned enterprise refers to the time between current time and enterprises registration date of registration
Every, wherein, if the Enterprise Age of Target Enterprise is less than or equal to 2 years, mean that the Target Enterprise possesses voiding higher
The suspicion of invoice.Registered number of the enterprise corresponding to above-mentioned enterprises registration address refers on the enterprises registration address where enterprise
The quantity of the registered enterprise for possessing.If on the enterprises registration address that Target Enterprise is currently located, while possessing two families or two
Taxpayer more than family, then the Target Enterprise possess the suspicion of voiding invoice higher.It is registered that above-mentioned business entity possesses
Number of the enterprise refers to the quantity of registered enterprise that possesses of same juristic person of enterprise.If the same Enterprise Law person of Target Enterprise
To that should have taxpayers more than two families or two families, then the Target Enterprise possesses the suspicion of voiding invoice higher to part card number.It is above-mentioned
Otherness between enterprises registration address and the personal household register ground of business entity refers to enterprises registration address and the Enterprise Law of enterprise
Difference between the personal household register ground of people.If the individual of the business entity of the enterprises registration address of Target Enterprise and Target Enterprise
Differ, then the Target Enterprise possesses the suspicion of voiding invoice higher household register.The improper shape of paying taxes of above-mentioned business entity
State information refers to whether the business entity of enterprise once had improper family criminal record.If the business entity of Target Enterprise was once
In the presence of excessively improper family criminal record, then the Target Enterprise possesses the suspicion of voiding invoice higher.The age of above-mentioned business entity is
Refer to the actual age of the business entity of enterprise.If the age of the business entity of Target Enterprise is less than 25 years old or more than 60 years old,
Then the Target Enterprise possesses the suspicion of voiding invoice higher.
Step S22：The first kind characteristic information of Target Enterprise is input into the type of business forecast model being pre-created.
Specifically, the establishment process of above-mentioned type of business forecast model, including below step S201 to step S203：
Step S201：First group of sample data and second group of sample data are obtained, sample set is obtained；Wherein, first group of sample
Notebook data includes N1The corresponding first kind characteristic information of enterprise of family and the type of business, second group of sample data include M1Family
The corresponding first kind characteristic information of enterprise and the type of business, N1And M1It is positive integer, also, N1Any enterprise in enterprise of family
The type of business be voiding invoice enterprise, M1The type of business of any enterprise is legal enterprise of paying taxes in enterprise of family.
In the present embodiment, in order to ensure the uniformity of sample, above-mentioned N1Value and M1Value is preferentially taken as identical numerical value.In addition,
In the present embodiment, above-mentioned N1Value and M1Value sum is preferentially more than or equal to 10000.
Step S202：According to default first division proportion, sample set is divided into two parts, obtains corresponding first instruction
Practice collection and the first test set.
Wherein, above-mentioned first division proportion can need specifically to be set according to practical application, for example, can be set to 7:3
Ratio, the ratio between the sample size in sample size and the first test set in the first training set for so finally giving is 7:3.
Step S203：Using machine learning algorithm, the first training set and the first test set, carry out corresponding model training with
And model measurement, obtain above-mentioned type of business forecast model.
Wherein, the machine learning algorithm in above-mentioned steps S203 is specially decision Tree algorithms.In addition, in the present embodiment, base
The model depth of the type of business forecast model obtained in decision Tree algorithms training is preferentially set to 5.
Step S23：Obtain the enterprise that above-mentioned type of business forecast model is exported according to the first kind characteristic information of Target Enterprise
Industry type prediction result, with according to the type of business predict the outcome determine Target Enterprise the type of business whether be write out falsely invoice look forward to
Industry.
Technical scheme disclosed in the present embodiment is applied to invoice and leads the purchase stage, in the invoice neck purchase Structure Stage of enterprise,
By the technical scheme disclosed in the embodiment of the present invention, identification can be monitored to the potential invoice enterprise that writes out falsely, so that real
Present invoice voiding event finds potential voiding invoice enterprise in time before occurring.
It is shown in Figure 3, enterprise supervision recognition methods is write out falsely the embodiment of the invention discloses a kind of specific invoice, including
Following steps：
Step S31：The Equations of The Second Kind characteristic information of Target Enterprise is extracted, the Enterprise Age of Target Enterprise is obtained, is continuously issued
The quantity information and time interval information of invoice, evening this month amount of money of making out an invoice account for this month to date and make out an invoice ratio, this month of the amount of money
Evening number of making out an invoice accounts for this month to date ratio of number, this month amount of money of making out an invoice of making out an invoice and is accounted for more than the invoice amount of preset cost threshold value
This month to date ratio of the amount of money, this month amount of money of making out an invoice of making out an invoice accounts for this month to date and opens more than the invoice number of default amount of money threshold value of making out an invoice
The ratio of ticket number, this month to the number that same purchaser issues invoice account for this month to date make out an invoice number ratio and this month
The invoice amount issued invoice to same purchaser accounts for this month to date and makes out an invoice the ratio of the amount of money.
That is, in the present embodiment, the Equations of The Second Kind characteristic information of Target Enterprise can be based on, whether the Target Enterprise is judged
To write out falsely invoice enterprise.Wherein, the Equations of The Second Kind characteristic information of enterprise includes the Enterprise Age of enterprise, the continuous quantity for issuing invoice
Information and time interval information, evening this month amount of money of making out an invoice account for the make out an invoice ratio of the amount of money, evening this month of this month to date and make out an invoice part
Number accounts for this month to date ratio of number, this month amount of money of making out an invoice of making out an invoice and accounts for this month to date more than the invoice amount of preset cost threshold value and open
The ratio of the ticket amount of money, this month amount of money of making out an invoice account for this month to date and make out an invoice the ratio of number more than the invoice number of default amount of money threshold value of making out an invoice
Example, this month to the number that same purchaser issues invoice account for this month to date make out an invoice number ratio and this month to same purchase
The invoice amount that issues invoice of side accounts for this month to date and makes out an invoice the ratio of the amount of money.
Specifically, in the present embodiment, if the Enterprise Age of Target Enterprise is less than or equal to 2 years, meaning the target
Enterprise possesses the suspicion of voiding invoice higher.If Target Enterprise issued a large amount of invoices in 30 minutes, and issued invoice
Time interval be no more than 3 minutes, then mean that the Target Enterprise possesses the suspicion of voiding invoice higher.If this month target
Enterprise issue at night special invoice without tax volume and this month to date issue special invoice without the ratio between tax volume
Example is more than 50%, then mean that the Target Enterprise possesses the suspicion of voiding invoice higher.If this month Target Enterprise is at night
The ratio between the special invoice number that special invoice number and this month to date are issued is issued more than 20%, then means the target
Enterprise possesses the suspicion of voiding invoice higher.If all top special invoices issued of lattice of this month Target Enterprise without tax
Ratio between the amount of money that volume and this month to date issue invoice is more than 20%, then mean that the Target Enterprise possesses voiding higher
The suspicion of invoice.Wherein, the special invoice that above-mentioned top lattice are issued refer to make out an invoice the amount of money more than maximum issue limit 99% it is special
Use invoice.The ratio if quantity and this month to date of the special invoice that all top lattice of this month Target Enterprise are issued are made out an invoice between number
Example is more than 20%, then mean that the Target Enterprise possesses the suspicion of voiding invoice higher.If this month Target Enterprise is outwards
The invoice number issued of same purchaser and the ratio made out an invoice between number of this month to date more than 80%, then mean the target
Enterprise possesses the suspicion of voiding invoice higher.If all invoices that this month Target Enterprise same purchaser outwards issues
Invoice amount and the amount of money that issues invoice of this month to date between ratio more than 20%, then mean the Target Enterprise possess compared with
The suspicion of voiding invoice high.
Step S32：The Equations of The Second Kind characteristic information of Target Enterprise is input into the type of business forecast model being pre-created.
Specifically, the establishment process of above-mentioned type of business forecast model, including below step S301 to step S303：
Step S301：The 3rd group of sample data and the 4th group of sample data are obtained, sample set is obtained；Wherein, the 3rd group of sample
Notebook data includes N2The corresponding Equations of The Second Kind characteristic information of enterprise of family and the type of business, the 4th group of sample data include M2Family
The corresponding Equations of The Second Kind characteristic information of enterprise and the type of business, N2And M2It is positive integer, also, N2Any enterprise in enterprise of family
The type of business be voiding invoice enterprise, M2The type of business of any enterprise is legal enterprise of paying taxes in enterprise of family.
In the present embodiment, in order to ensure the uniformity of sample, above-mentioned N2Value and M2Value is preferentially taken as identical numerical value.In addition,
In the present embodiment, above-mentioned N2Value and M2Value sum is preferentially more than or equal to 10000.
Step S302：According to default second division proportion, sample set is divided into two parts, obtains corresponding second instruction
Practice collection and the second test set.
Wherein, above-mentioned second division proportion can need specifically to be set according to practical application, for example, can be set to 7:3
Ratio, the ratio between the sample size in sample size and the second test set in the second training set for so finally giving is 7:3.
Step S303：Using machine learning algorithm, the second training set and the second test set, carry out corresponding model training with
And model measurement, obtain type of business forecast model.
Wherein, the machine learning algorithm in above-mentioned steps S303 is specially decision Tree algorithms.In addition, in the present embodiment, base
The model depth of the type of business forecast model obtained in decision Tree algorithms training is preferentially set to 5.
Step S33：Obtain the enterprise that above-mentioned type of business forecast model is exported according to the Equations of The Second Kind characteristic information of Target Enterprise
Industry type prediction result, with according to the type of business predict the outcome determine Target Enterprise the type of business whether be write out falsely invoice look forward to
Industry.
Technical scheme disclosed in the present embodiment is applied to invoice issuing monitor stages, and rank is monitored in the invoice issuing of enterprise
Section, by the technical scheme disclosed in the embodiment of the present invention, can be to invoice that may be present during the invoice issuing of enterprise
Voiding behavior, so as to identify whether the enterprise is to write out falsely invoice enterprise.
Accordingly, enterprise supervision identifying system is write out falsely the embodiment of the invention also discloses a kind of invoice, it is shown in Figure 4,
The system includes：
Information extraction modules 11, the characteristic information for extracting Target Enterprise, obtain target signature information；
MIM message input module 12, for target signature information to be input into the type of business forecast model being pre-created；Enterprise
Industry type prediction model is, using default machine learning algorithm and sample set, to carry out the differentiation obtained after corresponding model training
Model；Every part of sample data in sample set includes the enterprise of the type of business characteristic information of clear and definite enterprise and the enterprise
Industry type；Wherein, the type of business of enterprise is voiding invoice enterprise or legal enterprise of paying taxes；
As a result acquisition module 13, for obtaining the type of business that type of business forecast model is exported according to target signature information
Predict the outcome, whether the type of business that Target Enterprise is determined to predict the outcome according to the type of business is to write out falsely invoice enterprise.
It can be seen that, the present invention advances with machine learning algorithm and the sample comprising enterprise characteristic information and the type of business
Collection, creates the forecast model for being predicted to the type of business, when needing to be monitored identification to Target Enterprise, can be right
The Target Enterprise carries out corresponding feature extraction, and the target signature information for extracting then is delivered to above-mentioned forecast model, from
And the type of business of above-mentioned Target Enterprise can be obtained, actually or above-mentioned Target Enterprise voiding invoice enterprise is predicted with this
Legal enterprise of paying taxes.Therefore, the present invention is to be monitored identification to enterprise based on machine learning algorithm, to determine enterprise
Whether industry be to write out falsely invoice enterprise, it is to avoid the various defects that manual analysis is brought, and thus improves and writes out falsely enterprise to invoice
Monitoring recognition effect and recognition efficiency.
Specifically, above- mentioned information extraction module, can include first kind feature extraction submodule, for extracting Target Enterprise
First kind characteristic information, obtain the Enterprise Age of Target Enterprise, registered number of the enterprise, enterprise corresponding to enterprises registration address
Otherness, enterprise between the personal household register ground of registered number of the enterprise, enterprises registration address and business entity that industry legal person possesses
The age of improper pay taxes status information and the business entity of industry legal person.
Corresponding with above-mentioned first kind feature extraction submodule, the type of business forecast model in the present embodiment is specifically profit
The sample set of the first kind characteristic information with default machine learning algorithm and comprising enterprise and the type of business carries out model instruction
The model obtained after white silk.
In addition, the information extraction modules in the present embodiment, it is also possible to including：
Equations of The Second Kind feature extraction submodule, the Equations of The Second Kind characteristic information for extracting Target Enterprise, obtains Target Enterprise
Enterprise Age, the continuous quantity information for issuing invoice and time interval information, evening this month amount of money of making out an invoice account for this month to date and open
The ratio of the ticket amount of money, evening this month number of making out an invoice account for the make out an invoice ratio of number, this month of this month to date and make out an invoice the amount of money more than default gold
The invoice amount of volume threshold value accounts for the make out an invoice ratio of the amount of money, this month of this month to date and makes out an invoice hair of the amount of money more than default amount of money threshold value of making out an invoice
Ticket number accounts for the make out an invoice ratio of number, this month of this month to date and accounts for this month to date to the number that same purchaser issues invoice and make out an invoice part
Several ratio and this month, account for this month to date and make out an invoice the ratio of the amount of money to the invoice amount that same purchaser issues invoice.
Corresponding with above-mentioned Equations of The Second Kind feature extraction submodule, the type of business forecast model in the present embodiment is specifically profit
The sample set of the Equations of The Second Kind characteristic information with default machine learning algorithm and comprising enterprise and the type of business carries out model instruction
The model obtained after white silk.
Specifically, above-mentioned default machine learning algorithm is preferably decision Tree algorithms.
Finally, in addition it is also necessary to explanation, herein, such as first and second or the like relational terms be used merely to by
One entity or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or operation
Between there is any this actual relation or order.And, term " including ", "comprising" or its any other variant meaning
Covering including for nonexcludability, so that process, method, article or equipment including a series of key elements not only include that
A little key elements, but also other key elements including being not expressly set out, or also include for this process, method, article or
The intrinsic key element of equipment.In the absence of more restrictions, the key element limited by sentence "including a ...", does not arrange
Except also there is other identical element in the process including the key element, method, article or equipment.
Enterprise supervision recognition methods is write out falsely to a kind of invoice provided by the present invention above and system is described in detail,
Specific case used herein is set forth to principle of the invention and implementation method, and the explanation of above example is use
Understand the method for the present invention and its core concept in help；Simultaneously for those of ordinary skill in the art, according to of the invention
Thought, will change in specific embodiments and applications, and in sum, this specification content should not be construed as
Limitation of the present invention.
Claims (10)
1. a kind of invoice writes out falsely enterprise supervision recognition methods, it is characterised in that including：
The characteristic information of Target Enterprise is extracted, target signature information is obtained；
The target signature information is input into the type of business forecast model being pre-created；The type of business forecast model is
Using default machine learning algorithm and sample set, the discrimination model obtained after corresponding model training is carried out；The sample set
In every part of sample data include the type of business of the type of business characteristic information of clear and definite enterprise and the enterprise；Its
In, the type of business of enterprise is voiding invoice enterprise or legal enterprise of paying taxes；
Obtain the type of business forecast model to be predicted the outcome according to the type of business that the target signature information is exported, with basis
The type of business predicts the outcome and determines whether the type of business of the Target Enterprise is to write out falsely invoice enterprise.
2. invoice according to claim 1 writes out falsely enterprise supervision recognition methods, it is characterised in that the extraction Target Enterprise
Characteristic information process, including：
The first kind characteristic information of the Target Enterprise is extracted, Enterprise Age, the enterprises registration address of the Target Enterprise is obtained
Registered number of the enterprise, enterprises registration address and the business entity's that corresponding registered number of the enterprise, business entity possess
The age of improper pay taxes status information and the business entity of otherness, business entity between personal household register ground.
3. invoice according to claim 2 writes out falsely enterprise supervision recognition methods, it is characterised in that the type of business prediction
The establishment process of model, including：
First group of sample data and second group of sample data are obtained, the sample set is obtained；Wherein, first group of sample data
Include N1The corresponding first kind characteristic information of enterprise of family and the type of business, second group of sample data include M1Enterprise of family
The corresponding first kind characteristic information of industry and the type of business, N1And M1It is positive integer, also, the N1Any enterprise in enterprise of family
The type of business of industry is voiding invoice enterprise, the M1The type of business of any enterprise is legal enterprise of paying taxes in enterprise of family
Industry；
According to default first division proportion, the sample set is divided into two parts, obtains corresponding first training set and
One test set；
Using the machine learning algorithm, first training set and first test set, carry out corresponding model training with
And model measurement, obtain the type of business forecast model.
4. invoice according to claim 1 writes out falsely enterprise supervision recognition methods, it is characterised in that the extraction Target Enterprise
Characteristic information process, including：
The Equations of The Second Kind characteristic information of the Target Enterprise is extracted, the Enterprise Age of the Target Enterprise is obtained, is continuously issued invoice
Quantity information and time interval information, evening this month amount of money of making out an invoice account for this month to date and make out an invoice ratio, evening this month of the amount of money
Number of making out an invoice accounts for this month to date ratio of number, this month amount of money of making out an invoice of making out an invoice and accounts for this month more than the invoice amount of preset cost threshold value
The accumulative ratio of the amount of money of making out an invoice, this month amount of money of making out an invoice account for this month to date and open more than the invoice number of the default amount of money threshold value of making out an invoice
The ratio of ticket number, this month to the number that same purchaser issues invoice account for this month to date make out an invoice number ratio and this month
The invoice amount issued invoice to same purchaser accounts for this month to date and makes out an invoice the ratio of the amount of money.
5. invoice according to claim 4 writes out falsely enterprise supervision recognition methods, it is characterised in that the extraction Target Enterprise
Characteristic information process, including：
The 3rd group of sample data and the 4th group of sample data are obtained, the sample set is obtained；Wherein, the 3rd group of sample data
Include N2The corresponding Equations of The Second Kind characteristic information of enterprise of family and the type of business, the 4th group of sample data include M2Enterprise of family
The corresponding Equations of The Second Kind characteristic information of industry and the type of business, N2And M2It is positive integer, also, the N2Any enterprise in enterprise of family
The type of business of industry is voiding invoice enterprise, the M2The type of business of any enterprise is legal enterprise of paying taxes in enterprise of family
Industry；
According to default second division proportion, the sample set is divided into two parts, obtains corresponding second training set and
Two test sets；
Using the machine learning algorithm, second training set and second test set, carry out corresponding model training with
And model measurement, obtain the type of business forecast model.
6. the invoice according to any one of claim 1 to 5 writes out falsely enterprise supervision recognition methods, it is characterised in that the machine
Device learning algorithm is decision Tree algorithms.
7. a kind of invoice writes out falsely enterprise supervision identifying system, it is characterised in that including：
Information extraction modules, the characteristic information for extracting Target Enterprise, obtain target signature information；
MIM message input module, for the target signature information to be input into the type of business forecast model being pre-created；It is described
Type of business forecast model is that, using default machine learning algorithm and sample set, what is obtain after corresponding model training sentences
Other model；Every part of sample data in the sample set includes the type of business characteristic information of clear and definite enterprise and the enterprise
The type of business of industry；Wherein, the type of business of enterprise is voiding invoice enterprise or legal enterprise of paying taxes；
As a result acquisition module, for obtaining the enterprise-class that the type of business forecast model is exported according to the target signature information
Type predicts the outcome, and whether the type of business that the Target Enterprise is determined to predict the outcome according to the type of business is to write out falsely invoice
Enterprise.
8. invoice according to claim 7 writes out falsely enterprise supervision identifying system, it is characterised in that described information extracts mould
Block, including：
First kind feature extraction submodule, the first kind characteristic information for extracting the Target Enterprise obtains the target enterprise
Registered number of the enterprise that registered number of the enterprise, business entity corresponding to the Enterprise Age of industry, enterprises registration address possess,
The improper status information of paying taxes of otherness, business entity between enterprises registration address and the personal household register ground of business entity,
And the age of business entity.
9. invoice according to claim 7 writes out falsely enterprise supervision identifying system, it is characterised in that described information extracts mould
Block, including：
Equations of The Second Kind feature extraction submodule, the Equations of The Second Kind characteristic information for extracting the Target Enterprise obtains the target enterprise
The Enterprise Age of industry, the continuous quantity information for issuing invoice and time interval information, evening this month amount of money of making out an invoice account for this month and tire out
The meter ratio of the amount of money, evening this month number of making out an invoice of making out an invoice accounts for the make out an invoice ratio of number, this month of this month to date and makes out an invoice the amount of money more than pre-
If the invoice amount of amount of money threshold value accounts for this month to date, the make out an invoice ratio of the amount of money, this month, makes out an invoice the amount of money more than the default amount of money of making out an invoice
The invoice number of threshold value accounts for the make out an invoice ratio of number, this month of this month to date and accounts for this month to the number that same purchaser issues invoice and tire out
Meter makes out an invoice the ratio of number and this month accounts for this month to date and makes out an invoice the ratio of the amount of money to the invoice amount that same purchaser issues invoice
Example.
10. the invoice according to any one of claim 7 to 9 writes out falsely enterprise supervision identifying system, it is characterised in that the machine
Device learning algorithm is decision Tree algorithms.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611220015.4A CN106780001A (en) | 2016-12-26 | 2016-12-26 | A kind of invoice writes out falsely enterprise supervision recognition methods and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611220015.4A CN106780001A (en) | 2016-12-26 | 2016-12-26 | A kind of invoice writes out falsely enterprise supervision recognition methods and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106780001A true CN106780001A (en) | 2017-05-31 |
Family
ID=58926284
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611220015.4A Pending CN106780001A (en) | 2016-12-26 | 2016-12-26 | A kind of invoice writes out falsely enterprise supervision recognition methods and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106780001A (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108038490A (en) * | 2017-10-30 | 2018-05-15 | 上海思贤信息技术股份有限公司 | A kind of P2P enterprises automatic identifying method and system based on internet data |
| CN108595621A (en) * | 2018-04-23 | 2018-09-28 | 泰华智慧产业集团股份有限公司 | A kind of early warning analysis method and system write false value added tax invoice |
| CN109409969A (en) * | 2018-10-24 | 2019-03-01 | 仲恺农业工程学院 | A kind of business taxation fraud detection method, electronic equipment and storage medium |
| CN109583978A (en) * | 2018-11-30 | 2019-04-05 | 税友软件集团股份有限公司 | The method, device and equipment of invoice enterprise is write out falsely in a kind of identification |
| CN109636036A (en) * | 2018-12-12 | 2019-04-16 | 税友软件集团股份有限公司 | A kind of method, system and the equipment of the prediction of enterprise's invoiced amount |
| CN109657887A (en) * | 2017-10-10 | 2019-04-19 | 爱信诺征信有限公司 | A kind of enterprise operation health Evaluation method and system |
| CN109727111A (en) * | 2018-12-29 | 2019-05-07 | 航天信息股份有限公司 | Method, system and the computer readable storage medium of invoice behavior are write out falsely in monitoring |
| CN109993641A (en) * | 2017-12-28 | 2019-07-09 | 航天信息股份有限公司 | A kind of invoice writes out falsely method for early warning and system |
| CN110532542A (en) * | 2019-07-15 | 2019-12-03 | 西安交通大学 | It is a kind of that recognition methods and system are write out falsely with the invoice for not marking study based on positive example |
| CN110852856A (en) * | 2019-11-04 | 2020-02-28 | 西安交通大学 | Invoice false invoice identification method based on dynamic network representation |
| CN110866536A (en) * | 2019-09-25 | 2020-03-06 | 西安交通大学 | Cross-regional enterprise tax evasion identification method based on PU learning |
| CN111178219A (en) * | 2019-12-24 | 2020-05-19 | 泰康保险集团股份有限公司 | Bill identification management method and device, storage medium and electronic equipment |
| CN111192121A (en) * | 2019-12-17 | 2020-05-22 | 航天信息股份有限公司 | ANN-based automatic risk taxpayer early warning method and system |
| CN111242422A (en) * | 2019-12-31 | 2020-06-05 | 航天信息股份有限公司 | Method and device for identifying false invoice of oil product |
| CN111311340A (en) * | 2018-12-10 | 2020-06-19 | 航天信息股份有限公司 | Method and device for identifying false invoice issuing behaviors |
| CN113159874A (en) * | 2021-05-25 | 2021-07-23 | 北京中科闻歌科技股份有限公司 | Method and device for detecting value-added tax invoice and readable storage medium |
| CN113869383A (en) * | 2021-09-17 | 2021-12-31 | 北京中科闻歌科技股份有限公司 | Ticket-changing entity detection method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103646335A (en) * | 2013-12-27 | 2014-03-19 | 税友软件集团股份有限公司 | Electronic invoice detecting method and device |
| CN104700304A (en) * | 2013-12-05 | 2015-06-10 | 航天信息股份有限公司 | Method and system for supervising enterprise tax evasion through values of plain invoices of value-added tax |
-
2016
- 2016-12-26 CN CN201611220015.4A patent/CN106780001A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104700304A (en) * | 2013-12-05 | 2015-06-10 | 航天信息股份有限公司 | Method and system for supervising enterprise tax evasion through values of plain invoices of value-added tax |
| CN103646335A (en) * | 2013-12-27 | 2014-03-19 | 税友软件集团股份有限公司 | Electronic invoice detecting method and device |
Non-Patent Citations (1)
| Title |
|---|
| 王宏争 等: "决策树生成算法在税务稽查选案中的应用", 《科技情报开发与经济》 * |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109657887A (en) * | 2017-10-10 | 2019-04-19 | 爱信诺征信有限公司 | A kind of enterprise operation health Evaluation method and system |
| CN108038490A (en) * | 2017-10-30 | 2018-05-15 | 上海思贤信息技术股份有限公司 | A kind of P2P enterprises automatic identifying method and system based on internet data |
| CN109993641A (en) * | 2017-12-28 | 2019-07-09 | 航天信息股份有限公司 | A kind of invoice writes out falsely method for early warning and system |
| CN108595621B (en) * | 2018-04-23 | 2020-10-30 | 泰华智慧产业集团股份有限公司 | Early warning analysis method and system for false value-added tax invoice |
| CN108595621A (en) * | 2018-04-23 | 2018-09-28 | 泰华智慧产业集团股份有限公司 | A kind of early warning analysis method and system write false value added tax invoice |
| CN109409969A (en) * | 2018-10-24 | 2019-03-01 | 仲恺农业工程学院 | A kind of business taxation fraud detection method, electronic equipment and storage medium |
| CN109583978A (en) * | 2018-11-30 | 2019-04-05 | 税友软件集团股份有限公司 | The method, device and equipment of invoice enterprise is write out falsely in a kind of identification |
| CN111311340A (en) * | 2018-12-10 | 2020-06-19 | 航天信息股份有限公司 | Method and device for identifying false invoice issuing behaviors |
| CN111311340B (en) * | 2018-12-10 | 2024-02-13 | 航天信息股份有限公司 | Method and device for identifying virtual invoice behavior |
| CN109636036A (en) * | 2018-12-12 | 2019-04-16 | 税友软件集团股份有限公司 | A kind of method, system and the equipment of the prediction of enterprise's invoiced amount |
| CN109636036B (en) * | 2018-12-12 | 2021-03-26 | 亿企赢网络科技有限公司 | Method, system and equipment for enterprise invoice quantity prediction |
| CN109727111A (en) * | 2018-12-29 | 2019-05-07 | 航天信息股份有限公司 | Method, system and the computer readable storage medium of invoice behavior are write out falsely in monitoring |
| CN110532542A (en) * | 2019-07-15 | 2019-12-03 | 西安交通大学 | It is a kind of that recognition methods and system are write out falsely with the invoice for not marking study based on positive example |
| CN110532542B (en) * | 2019-07-15 | 2021-07-13 | 西安交通大学 | Invoice false invoice identification method and system based on positive case and unmarked learning |
| CN110866536A (en) * | 2019-09-25 | 2020-03-06 | 西安交通大学 | Cross-regional enterprise tax evasion identification method based on PU learning |
| WO2021057427A1 (en) * | 2019-09-25 | 2021-04-01 | 西安交通大学 | Pu learning based cross-regional enterprise tax evasion recognition method and system |
| CN110866536B (en) * | 2019-09-25 | 2022-06-07 | 西安交通大学 | Cross-regional enterprise tax evasion identification method based on PU learning |
| CN110852856A (en) * | 2019-11-04 | 2020-02-28 | 西安交通大学 | Invoice false invoice identification method based on dynamic network representation |
| CN111192121A (en) * | 2019-12-17 | 2020-05-22 | 航天信息股份有限公司 | ANN-based automatic risk taxpayer early warning method and system |
| CN111178219A (en) * | 2019-12-24 | 2020-05-19 | 泰康保险集团股份有限公司 | Bill identification management method and device, storage medium and electronic equipment |
| CN111242422A (en) * | 2019-12-31 | 2020-06-05 | 航天信息股份有限公司 | Method and device for identifying false invoice of oil product |
| CN113159874A (en) * | 2021-05-25 | 2021-07-23 | 北京中科闻歌科技股份有限公司 | Method and device for detecting value-added tax invoice and readable storage medium |
| CN113869383A (en) * | 2021-09-17 | 2021-12-31 | 北京中科闻歌科技股份有限公司 | Ticket-changing entity detection method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106780001A (en) | A kind of invoice writes out falsely enterprise supervision recognition methods and system | |
| CN111352971A (en) | Bank system monitoring data anomaly detection method and system | |
| CN108053310A (en) | Credit scoring method, apparatus, computer equipment and storage medium | |
| CN106557882A (en) | Power consumer screening technique and system based on various dimensions Risk Evaluation Factors | |
| CN106067088A (en) | E-bank accesses detection method and the device of behavior | |
| CN110895758B (en) | Screening method, device and system for credit card account with cheating transaction | |
| CN107507038A (en) | A kind of electricity charge sensitive users analysis method based on stacking and bagging algorithms | |
| CN109102396A (en) | A kind of user credit ranking method, computer equipment and readable medium | |
| CN112613977A (en) | Personal credit loan admission credit granting method and system based on government affair data | |
| CN106570753A (en) | Method and server for controlling default risk | |
| CN104036367A (en) | Intelligent customer credit optimizing processing method based on ERP system | |
| CN112907356A (en) | Overdue collection method, device and system and computer readable storage medium | |
| CN109242165A (en) | A kind of model training and prediction technique and device based on model training | |
| CN110097443A (en) | A kind of accounting processing method and its system based on economic affairs classification | |
| CN107194536A (en) | The power supply reliability negative benefit appraisal procedure and device of a kind of project investment | |
| CN107679862B (en) | Method and device for determining characteristic value of fraud transaction model | |
| CN107291922B (en) | A kind of data distributing method and device | |
| Banihashemi | Is horizontal curvature a significant factor of safety in rural multilane highways? | |
| CN106372969A (en) | Power user feature identification method and system | |
| CN114757596B (en) | Power business monitoring and early warning method and system and readable storage medium | |
| CN114998022A (en) | Compliance wind control method and system | |
| CN106022915A (en) | Enterprise credit risk assessment method and apparatus | |
| CN109272412A (en) | A kind of vehicle insurance Claims Resolution air control maintenance system | |
| CN108009271A (en) | With the screening method and device of the car fee evasion of falling card vehicle | |
| CN106960019A (en) | A kind of data Growth Rate Calculation method and apparatus |
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 |
Application publication date: 20170531 |
|
| RJ01 | Rejection of invention patent application after publication |